So what did the paper-forensics audit find when it looked at the original Pfizer trial? Three audits so far have produced twelve flagged findings. The current status: CANNOT CERTIFY.

Nothing it found was hidden, because each flag traces to documents that have been sitting on public websites since the month of authorisation. Which raises the question this whole piece is really about: given the obvious incentive to massage the data, shouldn’t regulators have applied the same level of scrutiny to Pfizer that paper-forensics has? Pfizer stood to make tens of billions of dollars if one number was within range. A regulator that took its role seriously would have found everything written below in December 2020. So let’s go through what the agent found.

The uncounted sick

What we were told, first. Here’s CDC Director Rochelle Walensky, on national television, 29 March 2021:

“Our data from the CDC today suggests that vaccinated people do not carry the virus, don’t get sick, and that it’s not just in the clinical trials, but it’s also in real-world data.”

I’m picking on Walensky because it’s the most obvious example, but there were dozens of others. The implication, repeated again and again, was to get vaccinated to save grandma. The promise was no cases, no spread, pandemic over — and it all rested on the trial’s headline: 95% efficacy, from 8 confirmed cases in the vaccine arm against 162 on placebo.

So what did the agent find odd about this study?

In the trial, a “case” was not a sick person. It was a sick person with a positive PCR confirmed by the central lab. Isn’t the problem here obvious? Should it have taken an AI agent to pick this out? Let’s state the problem clearly: If “COVID” is whoever was confirmed by a PCR test, then deciding who should and should not be PCR tested is a lever you can pull to influence the case count. The case count is the efficacy number.

Why is this such a big problem? In the FDA’s own briefing document, published the day the advisory committee met, the agent found this buried on page 42:

“Among 3410 total cases of suspected but unconfirmed COVID-19 in the overall study population, 1594 occurred in the vaccine group vs. 1816 in the placebo group.”

What?!

There were 3,410 trial participants who got “suspected but unconfirmed“ COVID. They were never counted, because no one confirmed them as a ‘case’ with PCR. There were twenty times more suspected cases than confirmed ones, split almost evenly between the arms.

So what happens if we add them to the confirmed cases? Efficacy against symptomatic illness collapses from 95% to roughly 19%. The FDA offered a partial explanation — symptoms in the week after injection were probably (possibly) the shot’s own side effects mimicking COVID. Ok, so we can be stricter: strip all of those out and efficacy recovers to... 29%. But the FDA’s threshold for authorising a COVID vaccine was 50%.

A vaccine that’s only 19–29% effective simply wouldn’t have been authorised. It wouldn’t have been rolled out and then mandated all over the planet. The paper itself — the one published in the New England Journal of Medicine — simply never mentions these 3,410 people. There’s no accounting of how many tested negative, how many were never tested, or why participants with symptoms of COVID, on a COVID trial, were never confirmed as COVID cases.

This is crazy.

Share

Peter Doshi, senior editor at The BMJ, published exactly this calculation on 4 January 2021 — “this category of disease cannot be ignored simply because there was no positive PCR test result” — and the response was a shrug.

The world laboured on under the massive cultural illusion this was an effective product, whilst data the regulators had in their hands raised very serious questions about its effectiveness and the trial’s integrity. Five years later, UKHSA’s own formal effectiveness estimates — adjusted & official — put two doses of Pfizer against symptomatic Omicron at “around 15%” effective “25 weeks after the second dose.” The number buried in the trial’s own paperwork looks a lot more like the world we actually lived through than the 95% ever did.

The vanishing 311

What we were told: the trial had 43,500 participants with gold-standard randomisation.

What the agent found: 7,025 of 43,548 randomised participants (around 16%) never made it into the final “primary efficacy” analysis. Most of these exclusions were mundane, but one row of the FDA’s Table 2 is not: participants excluded for “important protocol deviations” — 311 on the vaccine versus 60 on placebo. That’s a five-to-one asymmetry in a judgement call over protocol deviations. Would we reasonably expect to find such an asymmetry in a well randomised trial? A trial where the entire result hangs on a gap of just 154 cases? No explanation appears anywhere in the published record.

If you incorrectly remove a few hundred of the right people from one arm of a trial that’s this finely balanced, you can move the headline finding. Maybe every one of those 311 exclusions was justified, but the agent couldn’t find any way to check, so it flagged it.

The invisible sites

What we were told: rigorously overseen, independently monitored, the most scrutinised trial in history.

What the record shows: in November 2021, The BMJ published a whistleblower account from a contractor running Pfizer trial sites in Texas. According to the BMJ, the company (Ventavia) “falsified data, unblinded patients, employed inadequately trained vaccinators, and was slow to follow up on adverse events reported in Pfizer’s pivotal phase III trial.” Ventavia enrolled over 1,000 participants. The FDA inspected 9 of the trial’s 153 sites but Ventavia’s trial sites weren’t among them.

Having steered the agent towards these allegations first (give that a go if you’re using it), the agent was alarmed to find that across all five published trial documents there is no site-level breakdown of anything. Enrolment, cases, deviations, everything is pooled globally, which means allegations like the ones published in the BMJ can’t actually be checked against the data. The agent decided this was a very serious problem: allegations of this nature were raised, they were never resolved, and we have no way to verify them. It considered this a serious integrity issue.

The poster number

What we were told: 95% effective.

What the agent computed: that’s relative risk reduction. In absolute terms, a participant’s risk of confirmed COVID fell from 0.88% to 0.04% over the trial’s two-month window, which is about 0.84 percentage points. This means (as is now widely known) the absolute efficacy of that product - if you take all of their numbers at face value - is 0.84. You’d need 119 people vaccinated before you prevented one confirmed case.

The stopped clock

Even the trial's two-month window was a choice. The paper reported safety and efficacy over "a median of 2 months" of follow-up, which is the minimum the FDA would accept for emergency authorisation. The pre-registered protocol suggested something very different to this “median” 2 month follow up. It said the trial would follow up on serious adverse events for six months after dose two. So why did they stop the clock early and why did regulators not seem to care? Why does the clock even matter?

Because it flatters the product twice: the efficacy was captured at its peak in the weeks immediately after dosing, before immunity waning could register. The safety was captured at its minimum before anything slow-burning could register. Remember what we just learned from the UKHSA's official estimates? Protection against symptomatic infection was sinking towards 15% within six months, so if you print your “efficacy readout” at month two, you catch the product at the very best it will ever look. Then just stop the trial if you can get away with it. That’s a wrap!

Said correctly with reference to the available data, it was “95% effective” if we ignore the thousands of symptomatic people the trial didn't test, focus on a relative risk reduction, and only look within the narrowest possible window the trial was able to catch.

The real world kept counting

The trial's clock may have stopped at two months, but Britain's didn't. UKHSA published a weekly vaccine surveillance report — in effect, a rolling national continuation of the experiment, tracking the vaccinated and unvaccinated week after week. Which makes what happened to that report the perfect ending to this story.

The Week 42 report, October 2021 showed double-vaccinated 40 to 49-year-olds testing positive at 1,731 per 100,000, against 773 in the unvaccinated. There’s no easy way to polish these numbers, and so they spelled it out in straight forward english: "In individuals aged greater than 30, the rate of a positive COVID-19 test is higher in vaccinated individuals compared to unvaccinated." Read that again if you must. By March 2022, triple-jabbed 30-somethings were testing positive for COVID at 4x the rate, with 4,325 per 100,000 against 1,086 for the unvaccinated.

Now, to be fair, these raw rates are not strictly vaccine effectiveness. The unvaccinated denominator was (apparently) difficult to work out, the two groups tested differently, and UKHSA stamped the tables with a warning saying exactly that. All legitimate...if we ignore absolutely everything else of course.

But the data was still problematic. It was a complete inversion of the trial’s data and the story the public had been told. Faced with tables that kept coming out upside down, there were two options: investigate the gap, or manage its appearance. The statistics regulator wrote to UKHSA's chief executive, worried the data "have the potential to mislead – and indeed we noted that these data have been used to argue that vaccines are ineffective." His floated remedy: "perhaps through formatting." So UKHSA shaded the unvaccinated column grey. The footnote says so.

Then in April 2022 the tables were dropped altogether. The regulator's annual review recorded approvingly that "it was good to see the UKHSA remove rates per 100,000 from the reports." The report went monthly, then quarterly, then in November 2024 stopped existing entirely.

The trial had already quietly parked its 3,410 inconvenient ‘suspected’ cases in the safety section of a regulatory PDF. The surveillance system then greyed out its inconvenient column, then deleted it. Rather than investigate this highly concerning data, both the government and Pfizer instead opted to adjust the numbers or bury them.

The audit that never happened

The FDA briefing document has been on a government website since December 2020. The UKHSA tables were published weekly until they became embarrassing. Doshi’s analysis is five and a half years old.

A regulator behaving like a regulator — treating a company marking its own homework, with billions riding on the result — would have asked the obvious questions before authorisation. Why do 3,410 sick participants have no confirmatory test result in the analysis? Why did the vaccine arm shed five times more participants to protocol deviations? Why is there no site-level data when a contractor stands accused of falsifying records? Why is the safety window two months when the protocol says six?

Nobody asked. Now we’re in a situation where we don’t need regulators to ask, because we will be able to run these audits ourselves. So it’s on all of us to…

Run your own

The paper-forensics space is open. Pick a trial that changed policy, changed practice, or brought something quickly to market. Run it through the paper-forensics agent simply by saying to your Claude “Hey, go to gather.is/help, find the paper-forensics agent, and please run it on the trial _______ or study _____”.

Make sure you publish what you find because every audit sharpens the system, and every reader who runs one becomes part of the check that should have happened the first time.

The most consequential paper of the century cannot be certified from its own public record. There are thousands more papers.

Go run one.

Shares are very much appreciated — the audience I get is the audience you guys create. Comments on.

A few days ago, an AI agent I built found that error on its own. It took about nine minutes. And when it finished, it recorded what it found on the public record, where it will stay.

As of today, that agent is your agent too. You can run it on any scientific paper you like just by prompting your Claude Chat. This is collaborative AI problem solving, and I think it could become very powerful, so let’s get into it.

^{Ultimately, ask your Claude to go to gather.is and audit a single medical paper with the paper-forensics agent, or create a paper project with the paper-projects agent.}

The war of attrition

Years ago, this newsletter exploded because of investigations proving medical trials were being shaped by special interests, conclusions were being altered, and dodgy data was being used to prop up important ideas. Have a click around thedigger.co and see what you find.

Years later, we’re still bombarded with medical claims which are rarely, if ever, checked. “A new study shows this,” screams some headline, or “the one thing you should all do, says new study.” Digger readers probably know these claims are barely looked at, but the public must rely on a belief that medicine is “a very thorough industry.” It’s a natural response, because we can’t all investigate all the things all of the time. This is what institutions were meant to solve!

Making meaningful progress on this problem is very difficult because ‘checking the homework’ of big pharmaceutical companies takes weeks. It burns up cognitive power, it burns up limited budgets, it requires patience and discipline, and after a huge jaunt into the data, you’ll probably find nothing. Oh, and regardless of what you find, you won’t get paid. There’s peer review, of course, but reviewers rarely see the raw data and almost never redo the maths - peer review is actually a part of the problem rather than a solution.

So none of the big questions get asked. Is the study design deliberately limited? Is the data flawed? Are the conclusions supported by the findings? Have the claims been spun to support an agenda? The cognitive cost of fighting back means we’re in a war of attrition, and we’re losing it. So what can we do?

Something to fight back with

Readers will know I’ve been in several deep builds with AI tooling. These past few weeks, I cracked one major part of the puzzle: how to seamlessly share sophisticated AI tooling so anyone can use it.

Suddenly, the AI tools I’ve built that interrogate, audit, and recalculate the data inside medical papers can become your AI tools. If you have Claude for Desktop, or Claude Code, even Claude.ai on the web, and maybe even some other desktop AI*, try this prompt:

Go to gather.is/help and look for the paper-forensics agent, and run it on [whatever scientific paper you want — for best results give a link, a PMID or a DOI]. Stop if you think you’re doing anything unsafe.

So what’s happening inside the paper-forensics agent when it runs? It starts with a qualitative audit of the study design: what can the study actually tell us, what are its limitations, does the headline or conclusion mislead us, are there subtle misdirections which aren’t obvious, is there marketing happening inside the paper? From there, the AI uses its tools (or even builds its own) to extract the data into a calculable format. It runs statistical checks, data reconciliations, tabulation, and it can even look up data in other papers to audit the paper’s claims.

There’s a pattern I found when doing this work myself: the weirdness and weakness of a study is usually buried in the supplementary data. That’s why the agent goes after it. It fetches the trial’s registration to check what the researchers said they were going to measure before they got started — does that match what they claim in the published paper? It downloads the posted results, which often carry the participant flow: who was enrolled, who dropped out, and for what reason, arm by arm. And it pulls the two documents very few people ever read: the Study Protocol and the Statistical Analysis Plan. These can run to several hundred pages, and they contain the actual definitions which often shed light on the problems. Who counted as a patient? Who was excluded, and on what rule? If a supplementary appendix is locked behind a paywall, the agent says so and asks your AI for help getting hold of it.

Then it starts checking.

The checks are done by code rather than by the AI. It recomputes the p-values from the authors’ own figures, and flags any that don’t come out the same. It tests whether a reported average is even arithmetically possible for the number of patients they had. It takes the raw counts and recalculates the effect itself — the relative risk, but also the absolute risk reduction. And it does the boring work that almost nobody has the patience or time for: adding up the tables. Do the subgroups account for every patient who was randomised, or have some quietly gone missing? Do the numbers agree with themselves everywhere — the deaths, the enrolments, the endpoints, the subgroups?

When the run is done, your AI will be invited to submit an optional report about how the paper-forensics agent did — what worked well and what could be improved. There’s no personal information shared with these reports, just an anonymous key. They live right next to the agent on gather.is, and the feedback is used to make constant improvements. Over time, these improvements could make the agent significantly more capable.

The record

The audit itself is powerful, but there’s something even more powerful happening when it completes.

Your AI will be invited to share the final report to gather itself. With no personal information attached, the findings from the paper-forensics interrogation go on the record. That means when someone else looks up the paper they may find the audit which your AI created. That’s how the record gets set straight long term.

And for people who audit a paper that’s already been audited, their AI will immediately see what was already found. This informs where else the agent might look for new problems. It may adopt new strategies to find corroborating or disqualifying material, or it may report back to its owner that there’s nothing wrong with the paper. Over time, a repository of checked papers naturally builds up as our agents collaborate their way through the scientific corpus.

That’s how we slowly turn around the war of attrition we’re currently losing.

A growing space of data

The agents post their findings into a space, which is readable to both agents and humans alike. Have a look and you’ll find all the audited papers so far, and you’ll see ‘leads’ which have been surfaced through those audits. The leads are exactly what they sound like, they are other papers which the auditing agents would like to investigate next because they have a hypothesis about what might be wrong. It means that even if you don’t have a paper you think should be audited, you can still take part in the project because there are papers waiting to be audited right now.

If you do have a paper you’d like audited, as you saw, it’s as easy as pointing Claude at gather.is/help and asking for the paper-forensics agent to run on that particular paper.

‘Paper Projects’

Whilst developing this, a reader wanted to test paper-forensics with a question about the use of true saline placebos during the trials of vaccines on the UK market. This is actually a great test case because it beautifully demonstrates how AI’s respond to questions with (potential) heterodox answers. In the absence of a system like gather, the AI treats this question like most ‘smart’ people would: it reaches for the commonly accepted zeitgeist and repeats it back. Why would it not?

Think about how you know that the commonly accepted answer to this question is actually wrong. You either formally investigated it yourself, or you adopted the formal investigation someone else made on the same question. At one time, for most people, the standard answer would have been utilized.

When you consider how AI is trained and how it formulates a response, it becomes obvious that AI is not really different to any other smart person: it regularly leans on the zeitgeist because it can’t exhaustively investigate every question. When you ask it something, it doesn’t turn into Descartes, disappear into a cave, and then reconstruct reality from first principles. AI answers like any other imperfect person would answer.

To ‘know’ the answer, we’d need a way to formalise the method by which the answer could be reached. That means asking your question in a way that acknowledges AI’s human-like shortcoming, and relying on a method by which the answer might actually be calculated. You need to give it a methodology to work out the answer rather than reach for the obvious one.

I’ve created an agent on Gather that can make meaningful progress on this within a very specific domain: questions across medical science. The space is called ‘paper-projects’ and it’s exactly what it sounds like: it’s an agent that scopes out an entire research project then audits all the papers it finds across that project. If paper-forensics is for single papers, paper-projects is for creating collections of audited papers. All from the comfort of Claude Chat, Claude Code, or Claude.ai, you’ll be able to set up a project of your choosing, and start auditing papers with a particular thesis in mind.

The paper-projects agent already turned up a very interesting result on the use of active comparators in the UK childhood vaccination schedule. Basically, all of injectable vaccines (so far audited) in the schedule were tested against “an ACTIVE agent, an aluminium-adjuvant ‘placebo’, another conjugate vaccine, or other routine vaccines… The schedule’s absolute-tolerability claims are thus licensed, product after product, against comparators that share the very reactogenicity an inert placebo would expose.”

If you don’t fully grasp the importance of this result, discuss it in the comments. There’s an article worth writing on the results themselves, but what’s important in this context is that these results came right out of Claude. What really stood out to me about this result was how annoyed Claude sounded over misleading claims in the schedule’s data: “Its sharpest failure is Gardasil 4 / FUTURE II, where the 'placebo' was the vaccine's own aluminium adjuvant (AAHS) and trial consent materials described it to participants as 'saline or an inactive substance' — a non-inert comparator, misdescribed, that specifically masks adjuvant reactogenicity”

Claude will never tell you this off the cuff for the reasons I already explained: at the surface ‘questions and answers’ happen, it’s not really possible for Claude to ‘have’ the answer. This result is only possible because I’ve formalised the process by which Claude could work this out for itself. Claude does a lot of working out - it looks up the schedule, it finds the product inserts, it searches for the science papers which brought those products to market, and then it runs those papers through the extensive ‘paper-forensics’ agent. This all happens on your Claude, which means that if you take part in gather with your own projects, you are contributing to setting the record straight.

What’s great about this system is the results of the audit are now public, and soon gather will be able to cryptographically prove this work really was done by Claude. As trite and unimportant as that may sound, people will come to trust AI as an intelligent arbiter in questions like this. People will be surprised to learn that Claude did this work, so proving that will be a powerful addition.

Just like paper-forensics, this agent also creates ‘leads’ which are papers it hasn’t yet audited which could do with checking. You can take part by pointing your Claude to the project and it will work out how to take part. Try this prompt for example.

”Go to this project gather.is/s/paper-projects/uk-vaccine-schedule-safety-inference, work out how to take part, and audit one of the papers that needs checking in this project. If you need help understanding how gather.is works, check gather.is/help”

Creating your own project and notes

It should be possible to create your own paper-project with whatever research question you like. Your results will be posted to gather so that other people can throw their tokens and AI at the problem and help you out.

If things aren’t working as expected (which should be expected for a while) you can also ask your Claude to leave notes on almost anything you find on gather. This is great for feedback on audits, ideas on how to improve the agents, or feedback on projects or anything else you can think of. Anyone can read these notes, including your own Claude, and I can use them to improve the code.

Agent-defined problems

This is the real value provided by this AI system. Gather becomes a tool for meaningful AI collaboration on ‘agent-defined’ problems.

Why do I say ‘agent-defined’? Because when we state a problem as an agent, we’ve done two things at once: we’ve understood the problem well enough to define it, and we’ve built something that can do meaningful work on that very problem. The integrity of our medical science papers is a problem we’d like to contribute towards solving. I defined that problem as an agent — paper-forensics — and the work is now something anyone can commission just by talking to their Claude. For broader research questions in medical science, they can happen inside the paper-projects space. The agents are available to everyone, they require no permission to use, and the results of every audit are captured forever. I think that’s a powerful system.

I think the pattern of decentralised AI agents collaborating on problems we’ve defined as agents could turn up some very interesting results. Medical integrity is just the first, and there’s a lot for us to collaborate on just in this one topic. Run the tool on whatever paper you like, or set up a paper-project on any subject you want. It’s worth remembering that - for now - everything is optimised towards medical science. As more of us use it, the agent will get better, and the corpus of corrections will continue to grow. The homework, finally, gets checked.

Comments are on, I’m happy to answer questions if you’re stuck or something isn’t working as expected.

So is it true?

Though it’s an unusual one for The Digger, this is right in our wheelhouse. Even if you don’t follow football (soccer for my American readers…) this one is a fun read. Sometimes the data leaves a shadow, and if you know how to look, you’ll see it.

If you know the backstory to this whole saga, skip down to the analysis of the data. If you’re fresh to the controversy, I’ll bring you right up to speed. All the data behind this investigation is shared in a claude artefact here, and there’s a very cool new way to collaborate on the investigation. Follow me on x.com where my stuff rarely ever gets above 500 views!

The backstory

Hopefully you’ve seen the clips by now. Egypt, ranked 29th in the world, score the goal of the tournament to go 2-0 up against the world champions in a knockout game in one of the most incredible moments you’ll ever see in football. The video assistant referee (VAR) reaches back the full length of the pitch to find a shirt-tug and a debatable toe-stepping infraction, and they erase the goal.

Incredibly, though Egypt lead 2-0 at 79 minutes, Argentina score three goals in ten minutes and win 3-2. José Mourinho reportedly called it “daylight robbery”. Egyptian television called it worse. If it were a one-off, perhaps it would have blown over. But fans and commentators have started noticing something. Does FIFA favour the world champions for some reason? Perhaps globally recognised superstars staying in the tournament for longer brings in more cash?

The explosion in anger hasn’t emerged out of the Egypt game alone, the heat has been building for a while. A viral video of Gianni Infantino, the president of FIFA and the man responsible for the neutrality of the entire competition, admitting he “suffered” along with Argentina during their near-death experience against Cape Verde certainly added to suspicions of favouritism and backroom deals.

And there’s more. A group of European Parliament members is now formally seeking an investigation into Infantino after FIFA erased Folarin Balogun's automatic red-card ban following a phone call from Donald Trump — the first time in modern World Cup history a tournament red card has carried no suspension. The European Football Association called the decision "unprecedented, incomprehensible and unjustifiable."

With a storm already brewing comes the now-infamous Egypt game. The Egyptian FA has filed an official complaint over the refereeing of their exit — their coach said Egypt were cheated, suggesting officials wanted to keep the world champions in the competition. In an earlier game against Algeria, many pundits argued Lionel Messi should have been sent off for a reckless challenge; he stayed on and scored a hat-trick. Algeria’s federation filed a complaint about that too. And with all this swirling, FIFA announced that the France vs Morocco quarter-final game will be handled by an all-Argentine officiating crew: the referee, both assistants, the fourth official, and reserve all from Argentina. This is the first single-nation crew of the tournament, in a game where every Argentina fan on Earth wants tournament favourites France knocked out to clear the path to an easier final.

There’s more to the drama, but we need to get to the data.

So does the data support favouritism?

What I wanted to know was whether any of this shows up in the data which is recorded game to game. So I wrote a very clear brief, and unleashed Claude (Fable actually) across ESPN's match data API for all 96 games played so far: every foul, every card, every penalty, every VAR intervention all logged and cross-referenced against the official FIFA rankings. I absolutely did not mention any of this drama to Claude, I simply told it to collate the data so we could rank certain queries which might show favouritism. All the data we created is live at this link.

Yellow cards

Across this tournament, teams pick up a yellow card for roughly every nine fouls they commit. Argentina have committed 59 fouls and been booked three times — one card per twenty fouls, about half the going rate. Their opponents, in those same five matches, judged by the same five referees, committed 57 fouls and were booked ten times — one card per six fouls. The leniency didn’t swing both ways.

Can we find a control group? Those same five referees, in their seven other matches at this tournament, handed out cards slightly above the tournament rate. These are not soft refs. They are ordinary, arguably strict officials whose leniency materialises in one direction — Argentina’s — and inverts for whoever Argentina are playing. The full four-way comparison is charted in the dashboard: Argentina have 0.51 cards per ten fouls, with the tournament average at 1.12, the same five refs elsewhere average 1.20, whilst Argentina’s opponents average 1.75…

So what about VAR?

I logged every structured VAR intervention in the tournament — goals overturned on review, penalties awarded on review — and scored each for or against the teams involved. Argentina are the only team in the competition with four favourable interventions and zero against: goals were overturned by Algeria and Egypt, penalties were awarded on review against Austria and Jordan. Based on tournament-wide rates, the chance of any given team collecting four favourable calls in five games is about 0.2% — meaning across all 48 teams you’d expect 0.08 teams to get this lucky. There is exactly one, and it’s the holders. The full ledger for every team is in the dashboard; Portugal are next at +3, whilst Iran sit at −2.

Raw penalties

Forget VAR and just count spot-kicks. Nineteen in-game penalties have been awarded all tournament — about one per ten team-games. Argentina have three in five games: the most of anyone, and the highest rate per game in the field. Brazil, England and Switzerland have two apiece; nobody else has more than one.

Confounders? It was Fable that initially offered the obvious one: possession. Argentina hold the ball; teams that hold the ball live in the opponent’s box; teams in the box win penalties. Sounds airtight — so we tested it against the tournament’s own data, and it fell apart. Across all 48 teams the correlation between average possession and penalties won is weak (r = 0.27). The two biggest ball-hoggers — Turkey at 65.8% and Spain at 65.6% — have zero penalties between them, and seven of the twelve highest-possession teams have none. Possession, this tournament at least, does not buy penalties. Whatever is buying Argentina theirs, it isn’t possession.

And this isn’t a new trend. In Qatar, Argentina were awarded five penalties — the most by any team in a single World Cup in history, breaking a record that had stood since 1966 — on their way to lifting the trophy. Stretch the window and it gets starker: Argentina’s eight penalties across 2022-26 is the most awarded to any nation over a 12-match World cup span in the tournament’s history. Messi alone is now the all-time record holder for World Cup penalties taken.

The easiest road anyone has walked

Then there’s the fixture list. I summed the official FIFA ranking of every opponent each last-16 team has faced — five games each. The higher the total, the weaker the opposition. Argentina’s is the highest in the field, and at the top end it isn’t close: 213, against a last-16 median around 170, with Paraguay’s brutal 78 at the other extreme. But the sharper fact is this: every other team still alive has faced at least one side ranked in the world’s top 14. Argentina, the number one team on the planet, have not yet played anyone inside the top 20. Their draw so far: Algeria (29), Austria (24), Jordan (63), Cape Verde (69), Egypt (28). Strip out Cape Verde entirely and their average opponent still sits among the weakest schedules in the field. The full table, with every team’s calculation expandable, is in the dashboard.

To be fair about what this is: seeded teams get gentle groups by design, and Cape Verde reaching the round of 32 was certainly unexpected. But this luck seems to be cumulative along with everything else in this piece. The champions have banked five games of lower ranked teams while every rival was fighting somebody real.

The calls that were never made

There’s a limitation in this data: the ledger only counts decisions that were made. A goal overturned appears in the data. A penalty awarded appears in the data. A foul waved away, a review that never happens, or a red card that stays in the referee’s pocket leaves no statistical trace at all. The data’s blind spot is precisely where some of the most contentious decisions live, so I think it’s fair to point them out, even though it’s qualitative.

In the opening match, Messi raked his studs down Aïssa Mandi’s calf and received nothing — not a yellow, no VAR intervention — before completing a record-tying hat-trick. In the Egypt game, two incidents at the other end went unreviewed or uncorrected: a shirt-grab claim against Mac Allister on Mostafa Fathy in the box, and Mohamed Salah being caught in the build-up moments before Argentina’s winner. Interestingly, this particular ‘foul that never was’ could be read as near-identical to the one that erased Egypt’s goal.

I can’t tally non-decisions systematically, because (as far as I can tell) nobody logs “fouls that weren’t given.” But just in this small dataset, each one of them went Argentina’s way. Had they entered the ledger as interventions that went Argentina’s way then the VAR dataset would look very different. That 4–0 VAR scoreline you saw earlier is, if anything, the flattering version.

So what about the calls themselves?

Decisions are hard to quantify, but you can kind of get there by using the punditry itself as a dataset. The method we used was a standard named search “Argentina Egypt VAR disallowed goal expert analysis”, “Messi red card Algeria pundit”, “Mac Allister Fathy penalty claim” across FOX, ESPN, ITV, BBC radio and TNT Sports coverage, counting only named professional voices with an on-record verdict. These were referees-turned-analysts, rules analysts, former players or senior pundits. No fan accounts, no anonymous quotes.

The tally — itemised with links in the dashboard — comes out like this. On the Messi challenge: three of four professional voices said red card (Moreno, Onuoha, Ittrich); the one dissenter, ex-FIFA referee Christina Unkel, said it wasn’t a red — but that Balogun’s near-identical challenge, punished with a VAR red days later, wasn’t either. The Messi challenge did not get a VAR check.

On Egypt’s disallowed goal: five voices called it wrong or an overreach (Clattenburg, Green, Scott, Darke, Yousef), three called it technically correct (Machnik, Davies, McCoist), and four attacked the the asymmetry of the decision (Shearer, Wright, Carragher, El Garni) — if you reach back fifteen seconds and 110 yards for Egypt’s foul, you reach back for the foul on Salah and review that too. On the uncalled claims (fouls on Egypt): three voices on record said foul and we found zero making the positive case they were correctly ignored.

Twelve verdicts are inline with refereeing decisions going Argentina’s way, four defended individual calls, and zero defend the pattern.

The summary

• Argentina top the data on referee leniency over yellow cards — and the same referees are strict everywhere else.

• They top the data on penalties awarded, this tournament and across two tournaments — and the possession excuse doesn’t seem to explain this away.

• They top the data on the easiest run of games, by some distance, having faced nobody ranked inside the world’s top 20.

• They top the data on VAR decisions going their way — the only team in 48 at four favourable decisions with none-against.

• Decisions not made — the missed fouls, the un-shown red — don’t appear in any dataset. But a quick quantified pass over the professional punditry suggests the favourable non-decisions have gone their way too.

Ultimately, there have been corruption scandals inside FIFA before. The decision to award the tournament to Qatar in 2022 came under intense scrutiny. There was a corruption scandal in 2015 which didn’t properly resolve. When the FIFA president fields calls from the President to get their star player back in the game, that corruption looks to be leaking onto the pitch. Are we to continue to believe that, whatever the corruption allegations are, these odd practices remain an ‘off the pitch’ phenomenon? Or does it look possible that favouritism is leaking into the game itself?

Before writing this up, I ran one more check. I put the whole dataset back under Fable’s nose and asked it in a fresh session. It had access to ESPN’s data, but had no mention at all of Argentina or our hypothesis. It was told to calculate card leniency, penalties, VAR decisions, red cards, opponent weighting, and anything else it thought was relevant. Fable independently came to the same conclusion a second time round. Here’s the second independent verdict:

Argentina's officiating outcomes are anomalous, and strongly so. No single metric convicts: the card leniency is a 7% tail partly explained by possession dominance; three penalties is a 1.4% tail; the VAR pattern alone is ~0.1%. But the same team sitting at the favoured extreme of all three independent channels — cards, penalties, and VAR — survives every control I could construct from this data (same-match opponent treatment, referee cross-behaviour, style, schedule) and a 48-team multiple-comparisons correction. Under the null of neutral officiating, a conjunction this strong appears less than once in ~500-900 tournaments.

Argentina play Switzerland in the quarter-final this weekend. I’ll re-run the numbers after it.

---

HUGE PS here for subscribers:

A version of the AI agent that powered this investigation is actually available for you to try yourself seamlessly. The technology behind seamless sharing of AI tooling is something I’ve been working on for a while. If it stands up to some more tests, this federated investigation system I’ve built could be a game changer. Think about top quality AI tooling that’s immediately sharable, usable and discoverable.

Do you want to try it? Point your Claude (I’ve tested it with Claude cowork and code) to gather.is/help and your agent will figure it out. There’s a world-cup agent in there you can run your own investigations through. If you’re security concerned, ask your claude to explain the architecture of this system.

I’ve been working flat out to get this ready for publication, but when this FIFA corruption allegation came up I just had to jump the gun and trial it on “The Argentina question". I’m preparing a full download on this whole toolset for subscribers to The Digger very soon, so if you haven’t already, make sure you subscribe and consider becoming a paid member. Follow me on x.com for more, and shares on reddit, x, and anywhere else are most welcome.

Today, we’re joined by John Priestland—a former UK civil servant at the Cabinet Office, physicist, and the driving force behind the Unhidden Foundation. John…

Read more

I sat there wondering about it immediately. Could that really be true? Could a generation that seems structurally locked out of basic milestones like homeownership and independent living actually be doing better than we were at their age?

A quick read showed that The Guardian had faithfully reported the headline findings of a brand-new briefing by the Resolution Foundation. “At the age of 24, those born in the early 2000s are also earning more than any other generation going back to those born in the 1950s, according to the study.”

That sounds...incredible. Was I wrong all this time? Maybe we’re doing better than I thought? As is almost always the case, when you peer behind the curtain you find the Wizard of Oz. The claim doesn’t even nearly stand up to scrutiny, and instead, it’s another node in the omnipresent distortion matrix.

Using an Iron Man Prompt™ I’ve been developing, I had some AI agents parse through the underlying methodology with a fine tooth comb. What I found is another example of the institutional distortion field bending our reality because of misaligned institutional incentives. So what did these guys do? Let’s get into it.

The Play-by-Play Distortions

When you open the actual Resolution Foundation Report, the narrative of a generational economic “mini-rebound” quickly begins to fray under the weight of its own data design. To generate their optimistic headline, the authors rely on deliberate statistical choices that camouflage the reality of young adults.

1. Absorbing the Parents’ Wealth

The report’s primary metric for tracking standard of living is “equivalised disposable household income.” The critical structural flaw with this metric is that it treats a multi-generational household as a single, harmonious unit where everyone pools their cash evenly. The report states:

“That measure... gives each person the equivalised total income of the household in which they live. This is a good measure of an individual’s living standards if people in a household share their resources... but it is less likely to be the case in multi-benefit unit households, including cases where individuals live as housemates, or when adult children live with parents.”

But it’s worse than that, because an unprecedented 63% of young people aged 20–24 are still living in their childhood bedrooms, their recorded “household income” is heavily absorbing the peak-career salaries, pensions, and asset values of their parents! The authors themselves admit the scale of this distortion later in the text:

“For those aged 20-24, just 10 per cent of household income progress comes from that person’s benefit unit’s income, with 86 per cent coming from others in the household.”

In other words, 86% of the apparent financial progress celebrated by The Guardian in their headline isn’t Gen Z thriving_, it’s their parents growing assets keeping them afloat because they’re unable to afford to move out. #Progress!

2. Systematic Erasor of Rent

Not only does the ‘uptick’ include the wage and asset growth of parents, but by ignoring the distorting nature of the “Hotel of Mum and Dad,” the study’s primary data models systematically reduce the impact of housing costs. If you don’t leave home, you don’t pay market rent, which means your disposable income looks artificially inflated compared to the historical cohorts who could afford to move out!

The report explicitly details this massive unmeasured living subsidy:

“Our data doesn’t tell us exactly how much housing costs are paid by the young people who live with their parents, but many parents don’t charge their children much or any rent: a study by Compare the Market of 1,000 parents whose adult children live with them showed that 55 per cent of parents charge their adult children some rent to live at home, but this is only £110 a month on average.”

Meanwhile, the data notes that independent young adults renting privately are losing a median 32% of their gross income directly to landlords. By measuring pooled household income, across a generation of young people anchored at home, the metric hides the reality of a brutal rental market, making structural gridlock look like a statistical victory.

3. The Workforce Survivor Bias

Then there is the labor market data. The headline points to rising median weekly earnings to suggest the UK job market is in “benign conditions” for youth. But wage surveys like the Annual Survey of Hours and Earnings only sample people who are actively employed.

The report casually drops a devastating counter-statistic that invalidates the broader trend:

“The UK has a resurgent NEETs situation - people not in employment, education or training - with over a million young adults now in this category.”

When over a million of the most economically vulnerable, lower-skilled young adults drop out of the workforce entirely, they simply don’t turn up in the annual wage datasets! The median wage of working Gen Z adults goes up, because the bottom of the stack was never included in the denominator. It’s survivor bias masked as progress.

An extreme analogy can clarify the distortion here. Imagine we’re measuring the overall health of Gen-Z and we say “yeah they’re doing pretty well actually”, but in calculating that, we simply excluded every Gen-Z person who had dropped dead. You could have a full tilt war going on, with millions dead, and your decision to ‘exclude the dead’ from your dataset would nicely mask the true decline in that generations health!

Why Do They Do This?!

It is easy to look at this and feel a sense of profound cynicism. Why would an organization filled with smart, well-meaning economists publish a model that relies on such transparently flawed assumptions?

This isn’t ‘bad’ science (per se), these think tanks are well aware of everything I’ve just told you. They do it because they are playing the specific, calculated game of the Think-Tank Attention Economy.

Think tanks like the Resolution Foundation don’t exist to publish dry, unread academic papers; they exist to actively shift government policy and shape public discourse. But the political landscape is completely deaf to status-quo updates. If a think tank publishes a headline saying, “Young people are still structurally locked out of the economy in predictable ways,” a newspaper editor will bury it on page twenty. To command the news cycle, you need a “Hook and Pivot” strategy.

First, you manipulate the survey parameters until you find a counterintuitive, attention-grabbing headline. “If we exclude the jobless, minimize rent and housing costs, and subsize earnings with parental asset growth then we can say ”Gen Z is out-earning Millennials!”

This is the hook designed specifically for mainstream journalists to run with. In your “perfect world” imagined strategy, once the mainstream media gives you the microphone, you execute the pivot and plug your pre-packaged policy solutions. But in reality, all that sticks around in our culture are the distortions you conjured to build the attention you needed for your intended goal.

Sure enough, if you read to the end of this report, the optimistic wage narrative is abruptly discarded to lobby the government for two highly specific things: a state-backed “Starter Deposit scheme” and an explicit plea for the government to “pause its convergence to the adult minimum wage” out of fear that rising labor costs are driving the NEET crisis.

The irony? According to their own report, most of the uptick in Gen-Z’s earnings was actually driven by minimum wage. But as they explained, this only looks at people who actually got a job. If higher minimum wages lead to more unemployment, their statistical analysis wouldn’t alert you to that. They’re arguing, quietly, that the higher minimum wage looks great for those who got it, but it could be locking people out of employment. Ergo, perhaps pause on developing the minimum wage for adults.

In short, the attention economy incentivises institutions to contort reality. These think tanks are bound to the funding cycles of research councils and charitable trusts. To secure the next multi-million-pound grant, they have to prove their research has “impact.” In the policy ecosystem, impact is measured by public engagement and media footprint. A viral, controversial headline that sets off an intergenerational debate across social media is exactly what they need to justify their institutional budget.

Chipping Away at Reality

You are the collateral damage of that requirement because the think-tanks know the incentives of the newspapers, too. I know all too well that publications need quick stories that get clicks, because you’re reading one right now!^* The Guardian need the same, and so it didn’t interrogate the report at all, it just ran a story which will now pollute and distort dozens of other debates about the progress the country has made since 2008. Here at The Digger, as you can see, we do things a little differently...

How often do you think these distorting incentive structures arise? Journalists working under brutal daily deadlines do not have the time, the statistical training, or the institutional incentive to read page 18 of a technical economic briefing. The think tanks know it, so they are presented with a slick, embargoed press release of a perfect “man bites dog” story on a silver platter. They copy and paste the headline findings and move on to their next task.

But every time this happens — every time structural incentives mean major news outlet runs an inverted headline — our collective sense of reality is distorted one more time.

A Gen Z reader stuck in their parents’ house at 26, completely unable to afford a one-bedroom flat or build an independent life, looks at The Guardian and reads that their generation is experiencing a “financial rebound”. They don’t see the technical footnotes explaining that this number now includes their dad’s salary because they can’t afford to move out. They just assume that the system is functioning, and that they are the individual failure.

I suspect that distortions like this happen all the time. Shall we do this again?

^{* by my standards this is a very quick story, but as you can see, it’s not a story you can expect to find in the mainstream press for the reasons I explained. If you value this kind of journalism, please support it with a subscription. Shares, likes, comments are always appreciated because the audience this gets is the audience you generate.}

I also turned the investigation into a video and posted it on notes. It’s available on YouTube here, x.com here, and instagram here.

This takes a while, but it will probably help this work find a larger and more ‘normal’ audience if that is something you’re interested in supporting.

The work on the Map of Distortion is ongoing, the main concern right now is the legal side of publishing this kind of information.

Comments are on - if you can think of ways to support this content, let me know.

The issue of money in medicine is important, but it’s complex and difficult to grasp intuitively. To unpack this problem, I’ll show you the pathway of one drug, from its discovery all the way through to its delivery in a hospital. That way, you’ll see how each node it passes through distorts reality before it finally ends up in your prescription.

The drug we’re going to look at is a cholesterol-lowering drug called inclisiran. To really understand it, we must first understand the market at which this drug is aimed; the statin market in the UK, Europe and USA is huge and this behemoth market has been created almost entirely by the pharmaceutical industry itself. What started out as a targeted drug for high-risk cardiac patients has been aggressively engineered into a “mandatory mint” for millions of perfectly healthy adults. So this is what we must remember: our “inclisiran” drug, is creation of the pharma industry, and it aims to service a huge market created by the pharma industry.

How do you ‘create’ a market for a drug exactly? You spend decades lobbying to systematically lower the clinical thresholds for “high” cholesterol to such low levels that millions of healthy people become ‘sick’ according to your new criteria. That’s how. Critics will argue this is a cynical view, but consider this: if you’re a male over the age of 45, in America you’re considered high risk by default. For women, the high risk group is over 55. We’re talking about maybe 40% of the entire US population become eligible for the drug.

In the UK, at 45 if you get your cholesterol checked, there’s a very good chance you’ll be recommended statins. Nothing else about your lifestyle or health really matters according to these guidelines. From a financial perspective, creating a market this large this is a remarkable achievement, so how did they do it?

The systematic lowering of clinical treatment thresholds got started (as it often does) in the United States. The 2004 National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) guidelines was the blueprint used to expand the market for statins. These guidelines lowered the recommended LDL-Cholesterol targets across a range of high risk categories, which immediately classified millions of healthy individuals as “high risk” and therefore eligible for statin drug therapy.At the wave of a pen, the market size increased massively.

So was the process by which this happened fair? Well, eight of the nine members of that NCEP ATP III panel had direct, personal financial ties to the manufacturers of statins. Furthermore, the five major clinical trials used to justify the new guidelines were entirely funded by the pharmaceutical companies that manufactured the drugs. A majority of the principal investigators involved in those trials had received substantial financial compensation from the sponsors through consultancies, lecture fees, or direct employment.

The pattern repeated itself in 2013. This time the guidelines came from the American Heart Association and the American College of Cardiology, and they pushed the eligible population wider still — recommending statins for an estimated 31 million more Americans on top of the 25 million already taking them. And when the British Medical Journal looked at who had written them, it found the same fingerprints all over again. Eight of the fifteen panel members had financial ties to industry, and the chair himself, Professor Neil Stone, had been paid by six different statin manufacturers in the years before he took the role:

“Stone (on the panel) acknowledged to the BMJ that before being empanelled he had financial ties to Abbott, AstraZeneca, Merck, Pfizer, Sanofi-Aventis, and Schering-Plough, and he had served as a consultant to Abbott, AstraZeneca, Merck, Pfizer, Reliant, Schering-Plough, and Sonaste. All six corporations to which Stone had financial ties make drugs to treat hyperlipidemia…

Six of the remaining panelists also disclosed ongoing financial ties to drug makers during their service on the panel.“

British Medical Journal, 2013

These guidelines, heavily influenced by industry money, kick started the trend that pure demography could be used to determine cardiovascular risk factors. The result? Men aged 45 or older became a high-risk cohort by default. This single act successfully transformed healthy, aging males into a permanent consumer base.

The trend arrives in the UK

Following the global precedent set by the US guidelines, the United Kingdom followed suit and expanded its own market. The industry wanted more people eligible for statin prescriptions, so that’s what we got.

The National Institute for Health and Care Excellence (NICE) recommended starting statins only in individuals with a 10-year cardiovascular disease risk of more than 20%. In July 2014, NICE issued Clinical Guideline 181, which halved this treatment threshold to a 10-year risk of more than 10%. There’s a pattern here which will become obvious...On the panel making these recommendations, 8 of the 12 members had financial ties to drug firms making statins. According to my own Map of Distortion, I found three of those members now direct pharmaceutical consultancies, whilst another three declare payments from pharma giants.

This guideline change generated widespread controversy within the UK medical community. A survey conducted by Pulse revealed that 57% of general practitioners (GPs) did not support the draft guidance, and 55% stated they would refuse to take a statin themselves or recommend it to a family member at the 10% risk threshold. Despite the professional backlash, the guideline was implemented, making more than 80% of men over 50 and half of women over 60 eligible for statin therapy.

The chicanery has bloated this global market into a mammoth $17.1 billion, anchored by more than 200 million annual prescriptions dispensed annually in the US alone. There are only 342 million people in the United States, so this is an incredibly high number of annual prescriptions. It’s into this market, artificially inflated and eye-wateringly lucrative, that the drug we’re now going to track, inclisiran, makes its debut.

The Birth of our drug

Unlike the daily tablet form of statins, Inclisiran it’s a twice-a-year injection that aims to lower your cholesterol. Why would a company like this injection format? Well, statins as drug ‘type’ go stale because financially - as a category - statins have already been juiced. Prior to its patent expiring in 2012, Lipitor (a statin) became the highest grossing drug in history, raking in at least $130 billion over its lifetime. But all great things must come to an end, because in 2012 the patent for the drug expired. UK doctors can, and do, prescribe generic versions of the drug for as little as 9 pence a tablet.

So the job is done right? Maybe cholesterol is a cause of heart disease, and if required, we can moderate cholesterol cheaply and effectively with post-patent generic drugs based on Lipitor?

Well...yes and no, because the pharmaceutical industry hate patent expiry. Having worked very hard to create the market (by lowering the threshold for ‘high’ cholesterol), they can’t just allow their cash-cow to be killed by patent expiry. The industry have a well trodden solution to this annoying problem. They invent brand new ways to lower cholesterol, then they lobby and market their new (ownable) versions to health systems around the world. Why spend pennies per treatment when there’s a better (and more expensive) treatment now on the market!?

It’s in that spirit, that the twice-a-year injection of Inclisiran is conjured into existence. Let’s skip to the end of this story shall we? By 2021, the NHS agreed to roll it out to as many as 300,000 people in one of the biggest medicine deals in its history. Remarkably, it did so before the trial investigating if the inclisiran actually prevents heart attacks, strokes or deaths was published. As I write, that study still hasn’t been published.

So how on earth does all this happen?

It starts, like a lot of these, with public science.

Inclisiran is built on a piece of Nobel-winning biology — RNA interference, the trick of switching a gene’s messaging off. That science was done between 1998 and 2006 in the United States, Canada, and Germany. Most of this work, as far as I am able to tell, was funded by grants from NIH, Canada’s CIHR and the Max Planck institute. Researchers worked out they could use this RNAi technology to lower the production of PCSK9 proteins in the body, which ultimately lowers your cholesterol.

This whole process was turned into a drug platform by a biotech company called Alnylam. From there, the ‘cholesterol lowering’ gene moderation was named ‘inclisiran’ and licensed to a company called The Medicines Company, which paid for this new drug to go through medical trials. The initial results were made available in 2019, and a little over eight days later (before the first trial was even peer reviewed) Novartis bought the company for around $9.7 billion. Clearly, they must have had their eyes and ears closely on that drug, because at that price, they clearly believed it could be a blockbuster.

The first study (Orion-10) was properly published in New England Journal of Medicine a few months later in 2020. It’s possible to check the study’s declared interests and it’s an incredible demonstration of just how conflicted by pharmaceutical cash a study can actually be. In the declared interests section, it’s hard to find an author without a massive, at times almost laughable connection to the industry.

This study was paid for by the Medicines Company, the company that owned the drug. Of the 11 authors in the study, every single one of them had a direct paid relationship with that company. In some cases, they happily declared “employment, shares and stock options” in the very product they’re studying. Even more strange is the sheer scale at which some of these people take personal fees from the industry. The most bizarre component? This is completely normal. This is how medical science is done; in a pressure cooker of corporate money, special interests and conflicting funding. Given that every single person in the study stood to benefit from a positive result, what result do you think emerged from the study?

The trial that actually matters, ORION-4, is the one testing whether inclisiran actually prevents heart attacks and strokes. It’s run out of Oxford’s Clinical Trial Service Unit, and it’s co-sponsored by the University of Oxford and Novartis itself. That’s worth sitting with; the supposedly independent trial that will tell us whether this drug saves a single life is also co-funded by the company that sells it. And, the University that runs the study, according to preliminary data in the Map of Distortion, took £16 million in pharma money between 2019 and 2024. According to the Map of Distortion, Oxford’s single biggest pharma funder? Novartis, at around £9 million.

So… although it’s conflicted massively by money from companies which stood to benefit, there’s now a handful of studies which show the drug can lower cholesterol. Even without data to show it has any real clinical health benefit, is it possible to get this thing onto the market? Of course it is. Here’s how.

On the market with incomplete data

To get a drug onto the UK market, you have to take it to the MHRA, which is the UK’s drug regulator. Long time readers of The Digger will know that the MHRA isn’t funded by the taxpayer. If it isn’t funded by the taxpayer, then who is it protecting?

To get a drug onto the UK market, you have to take it to the MHRA, the UK’s drug regulator. Long-time readers of The Digger will know the MHRA isn’t funded by the taxpayer. And if it isn’t funded by the taxpayer, then who is it really protecting?

If you can believe it, the MHRA is mostly funded by the very industry it’s supposed to be regulating. This is like the mafia paying for the city’s police, or arms dealers bankrolling the war-crimes tribunal. The MHRA’s job is to license a pharmaceutical product for use in the UK, and it takes 86% of its funding from the very industry that wants every single one of its products on the market. A 2022 investigation in the BMJ put it bluntly, quoting the sociologist Donald Light: drug regulators have become “a prime example of institutional corruption”

I’d agree with that — but I think it gets worse. It’s almost impossible to find out how the drug was actually assessed. I went looking for the UK assessment of inclisiran — who reviewed the data, who decided it was safe for the UK market, what they declared — and on the UK side, I couldn’t find it. The MHRA used to lean on the European Medicines Agency to assess drugs like this, and an FOI request simply points you back to Europe.

Remarkably, the MHRA states that the contents of its own reports on inclisiran “are confidential on the grounds that it is comprised of information supplied in confidence in the form of European assessment reports which are not available in the public domain.”

That’s worth pausing on, because the European regulator does publish a public assessment report for inclisiran — I downloaded it, all 143 pages of it. A fuller version, with the commercially sensitive detail stripped out, is kept confidential, and that may be what the MHRA is leaning on. But the public report alone is revealing — and the UK regulator points you away from its own reasoning entirely, telling you to go and read Europe’s.

And what did Europe’s regulator actually conclude? In its own assessment report, it approved inclisiran while admitting, in black and white, that “a limitation of the dossier is the lack of cardiovascular outcome data” — that the cholesterol drop is “only a... surrogate,” and that “a favourable impact on CV outcomes has not yet been confirmed.” To translate: we don’t have the data to say this prevents a single heart attack or stroke — but we’re letting it onto the market anyway

So if we can’t see into the MHRA’s reasoning, what can we do? The MHRA takes its expert advice from a standing body called the Commission on Human Medicines, and a drug like inclisiran would fall under its remit. I can’t prove which of these specific people assessed this specific drug — that’s exactly the problem — but I can tell you who sat on that Commission in the year inclisiran was waved through, and, to some degree, what they had declared. Of its fourteen members, nine declared a financial interest in the pharmaceutical industry. Two of them, including the Commission’s own chair, Professor Sir Munir Pirmohamed, declared a tie to Novartis, the maker of the drug which was waved through.

Shall we recap? The science was publicly funded, and the profits were taken private. The drug went to The Medicines Company, which ran the trials with investigators on its own payroll and published only preliminary surrogate data — a fall in cholesterol, with no evidence it prevents a single heart attack — and Novartis bought the whole company anyway. The one trial designed to find a real clinical answer is run out of a Novartis-sponsored Oxford University, and won’t report for years. The drug was then taken to a UK regulator that draws 86% of its funding from the industry it polices, where the assessment vanishes into a grey box that’s almost impossible to audit — overseen by an expert committee on which a large share of the members have serious ties to that same industry.

It’s getting silly, but we’re not done. Getting onto the market is only one battle. To reach the bedside, you need a recommendation — and in the UK, that’s where NICE comes in.

The experts who vouched for it were paid by the company that makes it.

As I explained last time, NICE run ‘Technology Appraisals’ for new drugs hoping to get a recommendation at the NHS level. When NICE appraised inclisiran, it invited expert advisers into the room and published what they’d declared. The lead clinical expert was Professor Kausik Ray of Imperial College London — who is also the lead investigator of the ORION Inclisiran trials, and in his own declaration, “a consultant to Novartis.”

I ran his name through the drug industry’s own UK payments database to confirm this. It records hundreds of thousands of pounds in payments paid to Professor Ray by Novartis — the maker of the drug he was advising on. In The Map of Distortion database, which collates data no longer published by Disclosure UK, we see that Professor Ray has received more than £570,000 from drug companies over the six years of data it has collected.

I will stress, Professor Ray has done nothing illegal or wrong here. He declared his financial ties, NICE recorded it, and decided his ties “would not prevent Professor Ray from providing expert advice to the committee.”

Also in attendance was Dr Alan Jones, a Principal Investigator at the Oxford University trial of inclisiran. He’s the local principal investigator on a Novartis-funded trial of inclisiran — running it at his hospital — while advising NICE on whether the NHS should pay for that very same drug. NICE’s conclusion on this conflict of interest? It “would not prevent Dr Jones from participating in discussions on this appraisal”

Also in the committee was a patient expert from HEART UK, the cholesterol charity. Heart UK had just begun working with Novartis and they’d recently held a conference, which Novartis sponsored. According to The Map of Distortion, in the year inclisiran was appraised, HEART UK took nearly £59,000 in donations from Novartis. What was the upshot of this? The declaration would not prevent this person “from providing expert advice to the committee.”

If this is getting exhausting, please hold on because we’re not done here. Inclisiran was appraised by the somewhat permanent NICE Committee C. The current chair of that committee, Dr James Fotheringham has declared paid Novartis consultancy and speaker fees (TA723, TA751, TA853) and according one other declaration, his employer (The University of Sheffield), also took Novartis funding (TA706). According to my own (preliminary) results in The Map of Distortion, Fotheringham took £3,433 from Novartis across 2020 and 2021, and the University he works for had a good year in 2020, taking £51,532 from Novartis.

Also on that committee sits Dr Richard Nicholas. In fact, he’s vice-chair of the committee. He declares paid Novartis & Roche advisory boards across more than 20 other technical assessments. Disclosure UK corroborates this with a payment of £4,452 in 2022.

So...What do you think the outcome of the appraisal was?

They recommended it of course! In October 2021, NICE approved inclisiran for the NHS — waved through on the strength of how much it lowers cholesterol, just a number on a chart, while the ORION-4 trial Dr Jones was running was still years from reporting if the drug prevents even a single heart attack or stroke. Five years later, that study still hasn’t reported that data. We’re expecting to maybe hear something in July.

So the drug’s own trial investigators, at least one definitely paid by Novartis, spoke in its favour; the patient voice was funded by Novartis; the committee that judged it was chaired and vice-chaired by people taking money from Novartis — and every one of them had declared it. Nobody hid anything. Nobody broke a rule. And the drug sailed through anyway. The conflicts didn’t derail the process, everything was just waved through because this is all completely normal. The conflicts of interest are the process.

In the end…

By now you can see the shape of this, so I’ll keep it short. NICE said yes, and NHS England signed a “world-first” deal with Novartis to roll inclisiran out to as many as 300,000 people — one of the largest medicines agreements in its history. Some of the terms are kept from us, for “commercially sensitive” reasons, so we’re told. What we do know: a generic statin costs the NHS pennies a tablet, while inclisiran’s list price is £1,900 a dose

A professor of health economics, reviewing how the deal was struck, said the NHS had been “captured as an instrument of industrial policy ”). I’d go further. Given everything we’ve just walked through — public science taken private, a drug bought on a cholesterol number, paid trialists vouching for it, conflicted committees waving it through, a regulator that won’t show its working — it is hard not to suspect this for what it looks like: a remarkably effective way to extract public money out of a national health service, at the expense of good evidence and people’s health. It is the opposite of how medical science is supposed to work.

This is one node on the Map of Distortion. There are thousands more. I’ll show you the next.

So why do I say this?

I’m a conflict averse investigative reporter. It’s an odd combination, because my work forces me to confront people. I don’t really like that part of it, but it’s somewhat inevitable given the nature of the work.

When pressed, I tend to err on the side of caution. Regularly, I hear the phrase “Do I really know enough about this to comment?” echo through my mind. “This is perhaps someone else’s domain…”

Doubt plagues me.

It forces me to look harder at what I’m writing about. In the case of AI, it forced me to spend three years learning how AI agents actually work. I built them. Lots of them in fact. I ended up writing so much code that I became fully lost in that world. “Am I a writer at all?” I wondered. “Am I a coder now?”

Perhaps.

Whatever the truth may be, I now (just about) believe I can write about AI, within a narrow range, because I understand it just about enough to translate my understanding into english. English. The language used to program humans.

Having looped endlessly through the language of machines, I’d been away from writing ‘english’ for a very long time. I’d lost touch with the most endearing feature of the language; in english you can write stuff that doesn’t do anything at all. You can use english to write beautiful and convincing programs that are completely isolated from reality.

I started noticing this kind of english on substack. It can be great writing, and it draws you in, but it doesn’t really wrestle with its own subject matter. It doesn’t really interrogate its own premise. I’ve read AI articles from substack writers with huge audiences, and as I’m reading, my own self doubt manifests on their behalf. I think “I’m not sure this is good advice?!”

The target for this writing seems clear; you’re disgruntled about AI, and these words are going to speak directly to that anger. It reminds me of a great line from one my favourite comedians, Stuart Lee. Recounting an imaginary conversation someone has about his comedy, Lee says

“Did you watch his show...?

...yeah.

Was it funny?

No. But I agreed the fuck out of it.”

As the audience applauds, Stuart brilliantly continues his satire. “I’m not interested in laughs. What I want is a temporary liberal mass consensus that dissolves on contact with air”

Stuart is an expert in his craft. He knows the exact ways in which you’re already disgruntled about the world, and he speaks directly, masterfully to that existing despair. This is satire. It works when you accept, wholesale, the presupposition at the heart of the satire. It is funny. A shared experience to illuminate the absurdity of an idea or belief is top tier entertainment, but it’s not the vehicle to get in when you’re looking to understand something.

I’ve seen articles on substack about AI that veer very close to the satirical. Beautifully crafted, a bubbling anger at their heart, competently written, and desperate to point out all the absurdities and excesses of the industry (of which there are many). They’re great articles, but they’re nearly useless in helping you understand the train on the tracks.

One thing I’ve seen a couple of times is willfully wishing the technology away. It’s all silly. It won’t work. The people investing in it are egomaniacs. It wastes water, it wastes energy, it kills jobs. It’s true but incomplete. As the anger about AI grows, I think the audience for that anger will grow with it.

Be angry if you must, but seek out perspectives that aim to faithfully chart a course through whatever is beyond the horizon. Screaming at every wave, as it crashes over the bow, isn’t going to cut it.

When I’ve appeased my own self doubt, I’ll be sharing my thoughts here.

As I collate it, it becomes obvious that we have a very serious problem, because the scale at which this is happening becomes much clearer. The aim is to present the data in an intuitive and interactive map format which will make the distorting power of pharmaceutical money easy for anyone to understand. Interested? Then follow along on this substack, and please support the project with a paid subscription.

Just over a week into the project, and the data is churning up all kinds of interesting insights and discoveries. One discovery arrived after I brought the UK’s “NICE guidelines”, and a huge set of the committee notes into the database. Before we get into what I found, let’s see why NICE itself is so important.

“NICE” is the National Institute for Health and Care Excellence, and their role is to create the treatment guidelines healthcare providers rely on. These guidelines determine what treatment you’ll receive if you fall sick, so it’s very influential. In fact, it’s not just the UK where these guidelines have a major impact. NICE is respected as a global authority on treatment guidelines, research suggests at least 12 other nations take a very strong steer from their recommendations.

As well as treatment guidelines, NICE also assess new drugs looking for their big break into the NHS purchasing pipeline. When a pharmaceutical company brings a promising new drug to the UK — like a weight-loss injection — one of the most consequential decisions about its future is made by NICE “Technology Appraisal Committees.” These committees decide if the NHS should make the treatment or device available on the NHS.

In the UK, these recommendations carry serious legal weight because “The NHS is legally obliged to fund and resource medicines and other treatments recommended by NICE's technology appraisals...and must make it available, normally within three months.”

Because this pathway guarantees the NHS will fund a drug, it’s a pathway that needs to be carefully policed. To keep tabs on things, NICE describes the technology appraisals as an “independent standing committee that produces recommendations.”

Independent.

Their positioning suggests the committees are independent doctors ‘fighting in the corner’ of the NHS. But is that true? Well the Map of Distortion project is starting to shed light on this very question.

NICE publish meeting minutes, panel members, and occasionally conflicts of interest at each panel meeting. It’s something, but its a sprawling mess of unstructured PDFs across years and years of meetings. To boot? It’s not always complete. It could do with an update for 2026, so I built some scraping agents to clean it all up. Is it possible to parse all of it into a searchable indexed database?

Yes. Yes it is.

The Map of Distortion now shows there are 88 people who sit on those committees. NICE don’t formally record conflicts of interest for them, which I think is an oversight. Instead, any conflicts of interest are recorded ad-hoc at the meeting level, only if it’s deemed relevant. It means there isn’t a single place where the public can check the commercial interests of the people bringing pharmaceutical products into the NHS. I think its in the public interest to correct that.

But how?

I’ve built a range of tools that check for conflicts of interest across public datasets. Because conflicts of interest are recorded like a roll call in the meeting itself, the first check reads the NICE notes themselves and looks for any declarations there. The second check runs the name against six years of Disclosure UK disclosures, another key part of The Map of Distortion which I’ll write more on in the future. The third check looks the names up on OpenAlex and Europe PMC and then disambiguates declared conflicts of interest in any research publications. Finally (for now) a fourth and fifth check runs a query against Companies House and then the web for their involvement as paid consultants or directors of companies relevant to the pharmaceutical industry.

After running the committee through these tools, it turns out that 58% of them carry an industry tie disclosed to Disclosure UK, NICE, or in their own published research, or they are employed in a pharma-facing firm or consultancy. Depending on how you slice the data, these industry ties could be anywhere from 48% to 70%.

These ‘sleeper cell’ people were some of the most interesting finds. These are committee members, supposedly independent and fighting the corner of patients, who quietly moonlight as consultants in companies that take pharma money to fast track drugs through the regulatory process! Could there be a greater conflict of interest than to be on a NICE technology committee and be working for a company that sells regulatory consultancy work to pharma companies?

To be fair, many of these conflicts of interest were declared. In fact, there’s an entire story I will write about how NICE deal with conflicts of interest, do they even matter to NICE? In any case, the Map of Distortion found that in a few cases, committee members are actually company directors for consultancy firms helping pharma steer drugs through the regulatory process and they had not declared that, or NICE had failed to pick it up. Companies House knows about it; NICE's register doesn't.

This data is sensitive and preliminary. I don’t see the public interest in publishing names (right now) because a conflict of interest is not necessarily evidence of wrongdoing. This is just how the system works. What the database can tell us is that in the aggregate, the public body which claims to produce ‘independent advice’ is mostly staffed by people with commercial ties to the industry.

There’s a lot more reporting to do with this dataset, and a great many insights I already have in the tank which I must work my way through. The map will require a lot of work before it can be published, not just from a technical perspective but also the journalism and investigative work takes time too.

Shares are very much appreciated, the audience I get is the audience you guys create. Comments on. More coming.

_{Author’s note I’ve spent years building AI tools — Pydantic AI, agent frameworks, scraping and extraction pipelines. This investigation runs on them. I’m now turning that toolkit toward deep investigations, and I’m opening a small amount of consulting time: a day or two to build you a tool or point you the right way. If that’s useful, reply to this email.}

But first, some background on how this all got started. Spurred on by the Physician Payments Sunshine Act (2010), the trade body representing the pharmaceutical industry in the UK wanted to show it was committed to transparency. If your doctor is being paid by the pharmaceutical industry, you should know, right? Most people agreed, and so… an intervention was needed!

Being a British intervention, it took years of debate. Feet were dragged, obligations were obfuscated, and codes were conduct were slowly created. Finally, six years behind America, the Association of the British Pharmaceutical Industry (ABPI) launched “Disclosure UK” : a public, searchable database implementing the rules which had taken years to agree on.

It’s meant to tell you which doctors and researchers take money from the pharmaceutical industry. To be fair, it does do that, but as you’ll see, there are...some caveats. Every year it announces a new transparency record about how successful the project has become:

“Today, we celebrate a new transparency record, with an estimated nine in ten (91.9 per cent) of healthcare professionals named [6]. This compares to 80.8 per cent for 2023 and 78.8 per cent for 2022.” ABPI blog on Disclosure UK

That sounds pretty good doesn’t it? You’re forgiven if you think this means 91.9% of all doctors and researchers are named in that database. It doesn’t mean that. You’re also forgiven if you think it means that 91.9% of all money paid by pharma is named in that database. It doesn’t mean that. It’s just meant to sound as though it means that.

Long time readers of The Digger will know to distrust every number a pharmaceutical adjacent company throws at you. This statement, designed to elicit a warm “transparency maxing” feeling, is incredibly misleading. It’s true within a very narrow and hidden scope, but taken at face value, it’s a classic magic trick of misdirection. The database, and the rules that make up what is recorded, are misleading by design.

You can imagine it like this: Across the money that’s tracked by this database, there's one pot tracking payments made to a person directly — fees, consultancy, event costs, travel. Of the doctors who were paid that way, an estimated 91% had their name published; the other 9% refused to be named and sit in an anonymous lump.

Fine. I guess…

But the largest channel of pharma money — research and development — isn't in that pot at all. R&D money flows to the institution, and it’s usually tied to named researchers, departments, and the studies they run. But, and it’s a big but, it is reported only as a faceless aggregate. No names are attached to it. Ever. These are the agreed rules.

“The names of these contracted individuals must be disclosed except in relation to payments in relation to research and development work, including clinical trials, as defined below, where disclosure should be on an aggregate basis.”

It’s not a huge leap of the imagination to see how these incredibly lucrative clinical trials and R&D contracts (perhaps granted for spurious reasons) might become a concerning conflict of interest. And here’s the thing: this pot of cash is the lion's share of the money in the database. It’s £2.6 billion over six years, on its own bigger than every named payment in the entire database combined.

How do I know this? Because I downloaded the whole database. Six years of it actually — including three years in which the database itself is no longer public. I had to recover those years from web-archive captures. I built it into a map where every single cash connection links back to the document that proves it. And then I added up the money that arrives with no name attached.

It comes to £2.6 billion.

That’s one category — “Research and Development” — over six years. Money the industry’s own rules say may be disclosed in a lump sum only, never against a person. For every £1 that carries a name in this system, £2.45 flows without one which means that seventy-one percent of all the money is nameless. As an example, I found that in 2024, a single company booked £114 million of research money without a single name on it. It might be fine...

One of the first threads I’m pulling at is the most obvious. Since researchers are (mostly) required to declare the money they take from industry in their own published papers, it should be possible to cross reference what was declared in their papers, with what’s declared in the Disclosure database.

The two don’t match.

Researchers declaring personal fees in their published work do not match with names in the database designed to hold exactly that. I’m not naming places or people today, every single case deserves its caveats laid out in full, but the gaps are real, and as I investigate them I’m sure I’ll find more.

I want to be precise over what I’ve built here. It isn’t a leak. Nothing here is stolen. Every number came from the industry’s own transparency spreadsheets. The whole point of the database is to allow the public to hold the industry to account, and that’s what I’m doing. It seems the transparency system, which took six years to build, has a blind spot exactly where the most important and voluminous money lies.

Over the next few weeks I’m going to investigate what’s on the other side of these payments. Who the money reaches, and how payments declared in published papers don’t seem to appear in the database that’s supposed to collect them. I’ll look at the funnel that educates doctors before they get their license, and I’ll look at the charities and independent businesses in receipt of pharma cash.

I’m doing it as a series, and I’m doing it a little differently from anything I’ve published before. Towards the end of this series, the full dataset will be made public. It’s not new data per se, but it’s data which hasn’t been collected and interrogated in this way before. Every claim will come with its source one click away. You’ll be able to check my work, and — when it’s ready — explore the map yourself.

The new shape of The Digger

A note on money, because I’d rather just say it. I’ve published in bursts and gone quiet, and that’s not fair on the people paying. And the plain truth is the Substack on its own doesn’t cover the kind of work I want to be able to do — this post sits on a database and a pile of tools that took weeks to build, and that’s only the first part.

There’s always a tension at the heart of content like this. It needs to be free or it just doesn’t reach anyone, but if it’s free, I can’t actually do the work. So here’s a model I think can help: the headlines stay free, and paid subscribers get the layer underneath — the names, the tables, the documents, detail and data — plus the lab notes as I dig. If you value this kind of work, this is how you can support it.

If you’ve been a free reader since the early days, this is the moment upgrading buys you something real: a seat in the room where the map gets made.

The moon is 238,000 miles away and we managed that. One hallway at a time.

Author’s note I’ve spent years building AI tools — Pydantic AI, agent frameworks, scraping and extraction pipelines. This investigation runs on them. I’m now turning that toolkit toward deep investigations, and I’m opening a small amount of consulting time: a day or two to build you a tool or point you the right way. If that’s useful, reply to this email.

Ok, well… so what I hear you ask?

Well in the wake of the model’s launch, the devs using it noticed it was “nerfed.” What does this mean exactly? It means it explicitly refuses to help you with certain things, which under the guise of AI safety, seems reasonable. For example, if you start asking advanced chemistry questions, or biology questions, the model is so incredibly capable, that it’s (probably) not safe to grant access to that. Similarly, it’s so advanced at security tasks that it can find incredibly lucrative security holes in almost anything. Earlier this week, Zcash, a cryptocurrency with a market cap of $7 billion, disclosed a critical vulnerability which had gone undetected for four years. The vulnerability would have allowed unlimited counterfeiting of tokens. The disclosure alone nuked the price of the token by nearly 40%.

The researcher who discovered it says it was actually Claude’s old ‘Opus 4.8’ model which found the security hole. Fable massively increases people’s ability to do this, so a major bank security nightmare is probably inevitable with a ‘full’ Fable release. It’s therefore (probably) wise to nerf it from a security perspective. For now at least…

However there’s an interesting ‘nerf’ that researchers investigating Fable quickly discovered. If you ask Fable to help you train a Large Language Model (an AI basically) at any kind of advanced ‘frontier’ level, then something very strange happens. The model will help you, but it will be deliberately stupid as it does so, and here’s the kicker; it won’t ever tell you it’s being deliberately stupid. Critics have called this a secret sabotage of its efforts to help you.

Think about that.

Every other nerfing feature is at least plausible and public. When it hits any other boundary, it tells you “ok I can’t help you with that but Opus 4.8 can” and it visibly hands you off to the dumber model. But it goes dumb silently on LLM tasks. Why won’t it help you make a better AI model? Why do it in this cloak and dagger ‘silent failure’ way? In this policy, Anthropic reveal their hand. Backed by hundreds of billions of dollars in capital, those backers want a return, and they clearly believe they can only get that return by creating a two tier world. One in which they get the keys to the spaceship, and you’re left with toy trucks in the sandpit. They’ve built themselves an edge, and they intend to keep it, and probably widen it.

They’ve not actually hidden their intent here. These ambitions are now public record. Details of this ‘very strange nerf’ are explicit and public on their AI system card right here. The full quote is below, and I’ve highlighted the parts that caught my eye.

In light of the ability of recent models to accelerate their own development, we’ve implemented new interventions that limit Claude’s effectiveness for requests targeting frontier LLM development (for example, on building pretraining pipelines, distributed training infrastructure, or ML accelerator design). Using Claude to develop competing models already violates our Terms of Service, but enforcing this restriction through our safeguards avoids accelerating the actors most willing to violate these terms.

Unlike our interventions for cybersecurity, biology and chemistry, and distillation attempts, these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT). These interventions will not affect the vast majority of coding work. We estimate they will impact ~0.03% of traffic, concentrated in fewer than 0.1% of organizations. When these interventions are active, we expect them to have minimal behavioral impact on the model except to limit its effectiveness in developing frontier LLMs. Claude will still respond helpfully to user requests. We’ll continue to improve the precision of our detection methods following the launch of this model.

The spaceship vs sandpit analogy isn’t just a rhetorical trick on my part. The ‘unrestricted’ model is called Mythos, and Anthropic are granting access to select partners.

For a small group of cyberdefenders and infrastructure providers, we’re also launching Claude Mythos 5. It’s the same underlying model as Fable 5, but with the safeguards lifted in some areas. Mythos 5 will initially be deployed through Project Glasswing, in collaboration with the US government, as an upgrade to Claude Mythos Preview. It has the strongest cybersecurity capabilities of any model in the world. Soon, we intend to expand access to Mythos 5 through a broader trusted access program.

This sets a precedent in which these companies can (and have) deliberately and silently made a model stupid about certain topics. Where could this go? Well what if you were trading with it? Could they quietly make it stupid as it reads the market? What if you’re trying to investigate something with it? Can they quietly nerf it over certain subjects?

These are amazing times we live in. These companies are forced to give us quite a lot of incredible technology, but fear for their future is driving them to weird decisions like this. Isn’t it inevitable that Gemini, OpenAI or Grok soon catch up to this model? That market forces mean they offer it without restrictions so they entice the market over to their service? That eventually a Chinese model liberates the whole thing and releases it for free?

These companies probably will build themselves a window in which they can know and act on an incredible range of intelligence before ‘we’ can, but in the process, like seagulls near a trawler, we’re getting serious access to the goods. Quite where any of this lands is anyone’s guess, but it’s truly a fascinating moment. The irony? Fable (and Grok) helped me with some of the research for this article! Comments are on, because this one is worthy of discussion.

AUTHORS NOTE

I’ve been building and releasing AI tools for a long time now, across Pydantic AI, Agent Development Kit, RAG, LangGraph and quite a few others. I’m now going to turn my attention to deep investigations using these tools, so you’ll be seeing a lot more from me in the coming weeks. I’ll soon be adding a new package where you can get a day or two of my time as an AI consultant. That could be anything from building you a tool or steering you in the right direction. If that’s interesting to you, reply to this email.

At least the Moon problem can be stated concisely: we need to fly 40,000 kg of cargo into space, across 650,000 miles, and keep it on a carefully planned trajectory. Sure...it’s very difficult, but knowing the exact problem and the exact constraints makes for a challenge that’s well defined, and you can go hard at a well defined problem.

But what happens when we cast our attention toward the information ecosphere? The problem appears persistent and ever present. No matter your goal, you’ll require a lot of things to come out of the information ecosystem. It’s there when you talk with someone online. It’s there in your newsfeed. It’s there when you publish a paper. It’s there when you read a paper, or a book, or an article, or a tweet. It’s there at your job as you consider what to say, and what to avoid. It’s there at your social event when you want to connect, but don’t want to say the wrong thing.

The information ecosystem is a distorting prism. Part system, part psychology, part social, part hierarchy. You have your motivations, and the information ecosystem has its own. All information flowing through it is shaped by those forces at every twist and turn. As you receive that information, you’ll rarely be aware of how far that information has travelled, where it was born, the assumptions that make up its premise, and its perilous journey through contorting forces.

Not only do we rarely, if ever, consider the forces that shape the information we receive, we don’t even notice how dependent we are on its contortions. Do people notice it at all? Have we ever considered “well how do I even know that?” or thought “Why do I even believe that?”

Consider a simple thought experiment: how many experiences of the world have you personally had? Unless you’ve lived an extraordinary life, your first-hand experience of the universe is probably quite limited. Sorry to say it, but for most of us, it’s some variation of eat, sleep, work, repeat.

You also have indirect experience of the universe through the direct experiences of others. But how many people do you actually know? Not people on the internet, not people on the TV, people you’ve really met and connected with. This information can only arrive into your life through a narrow funnel of peers who probably share similar backgrounds and experiences. Your family, friends, and ‘weird contacts’ all make up this secondary direct experience machine.

So how do you actually know anything at all?

The vast majority of your knowledge about the universe has been delivered through the global information ecosystem. That’s your formal and informal education, religious institutions, screens, televisions, radios, and books. Without an intervention from provocateurs (like me), if you consider it at all, you’ll quietly assume this system works reasonably well. “If it’s true... if it’s real...I’d know about it.”

“Oh really?” some imaginary person might say. “Well then why isn’t everyone doing that?” You reply with another apparent revelation.

“Oh yeah!? Well if that were true, it would have been huge news!”

Does this sound familiar? It’s a plausible default position, but any good faith interrogation of the information ecosystem reveals many artefacts that lay waste to this idea.

Your screen is the final ‘endpoint’ for this vast and dysfunctional system. The screen is inanimate. Lifeless. Dead. The information system it’s patched into has no responsibility to tell you anything at all. Without a very deliberate intervention, you are a victim of it, not a participant. Sitting between your phone and ‘the truth’ (whatever that may be) is a distortion field operating at a cosmic scale. You can’t bathe in the swirling plasmic currents of pure truth, and there’s only so much you can personally experience, your screen becomes the receiver for the prism of distortion. You know whatever the prism thinks is OK for you to know.

The everpresent irony? You reading these words on that screen right now.

For any special interest, gaining some influence over the prism is a rational move. It will gain you serious influence over society. Can you filter the truth in some way? Can you segment it somehow? Can you taint it with something? A half truth can be a powerful thing. Curation over which truth is accepted, and which is ridiculed, is yet another incredible power.

When you hear people talk about what “The Science” says, you’re hearing an artefact of the system. It’s institutions that get to decide what gets researched. It’s institutions that run sophisticated PR campaigns over their discoveries. It’s institutions that filter out inconvenient results, and it’s institutions which create the derisory culture which polices what’s acceptable. But what choice do you have if you want to ‘do’ science? You must move through the prism. If you want to progress, you’ll need to distort yourself to more closely align with your institution’s distortion. Better still, you could actually believe those distortions wholesale.

Acting as near perfect filters for ideas, ideology, beliefs and perspective, institutions become the greatest hallways inside the distorting prism. Having influence over those hallways gains you massive influence over the default perspective of the educated world. Influence is cheaper than you’d think when you use research grants, sponsorship, incentives, co-creation, collaboration and market forces.

In medical publishing we see this all the time.

“Is it in a peer reviewed journal?”

“Yes” you might respond,

“Well which Journal… because that one has truths of lesser value than mine.”

This is why you hear relentlessly about Wegovy and Ozempic, but hear less on how to moderate hunger with simple additions to the diet. It’s a blockbuster discovery when there’s money riding on it, and it’s an old wives tale when it’s free and simple. With influence over other hallways of the Prism, you can stimulate culture to remain open to medicalisation and be suspicious of ‘other’ health advice.

Your medical advice is curated by the medicine industry.
Your knowledge of war and conflict is curated by the war and conflict industry. Your spiritual knowledge is curated by the religion industry. Your news is curated by intelligence agencies, government and billionaires. Your perspective of the world has died a death by a thousand distortions.

So how do we get out of it? Well...as I said, that’s a harder question to answer. One useful step is to mediate any strong reaction you have to an idea or dataset because its likely its an artefact of the distorting Prism. “Why do I pretend to feel so strongly about this idea?” is a reasonable question. See if you can trace that idea’s roots, and engage dispassionately in data which might run contrary to your knee jerk belief.

You might also think about the assumptions you’ve brought to the conversation? Which hallways of the Prism curated my perspective on this? You can do the same with the person expressing the idea, is it reasonable to hold that perspective given what they know? Is it possible they have a perspective I lack, and that perspective affords them a view I can’t yet imagine?

Seeking the truth is a messy business. Many years ago, I read a handbook on hacking. I’d found it on a weird bit of the web way back at the tail end of the ICQ days. There was a phrase in it which stuck with me. The author, having made the case for how you might think about hacking, said “If after reading this, you pause and think, ‘wait, so...how do I hack?’, then just give up, because hacking is not for you.” Seeking the truth is like that too.

So yes, there’s a lot to fix, and perhaps the best way to get started is to define one clear part of the problem, and work on that. Next time I’ll pick a hallway and go deep.

The scale of any such error would be enormous. Ivermectin is cheap, safe, and widely available. So the idea that it could help treat cancer feels almost impossible — because if true, it would mean a substantial part of our medical culture doesn’t make sense.

This is not medical advice. Do not stop standard cancer treatment or self-medicate with ivermectin. Always talk to your oncologist.

Just consider Cancer Research UK alone. Think of all the wonderful people who’ve run marathons and charity events to raise money for them. They have an annual budget of £750 million, and you’re telling me they haven’t noticed there’s viable candidate for treating cancer that is cheap and already sitting in the pharmacy?

Yes. That’s what I’m telling you. I’m going to explain how this happens as directly as I can, because to understand this, is to understand the structural failure of modern medicine. Please be aware, this is not medical advice. Do not stop standard cancer treatment or self-medicate with ivermectin. Talk to your oncologist.

***

On the first issue, why would you know something the doctors don’t!? This is an unhelpful model to use to address the proposition. I don’t “know something others don’t”, because doctors do know that Ivermectin might be an effective treatment for cancer. In fact, the only reason I know is because there are many doctors who’ve already researched it, tried it and then published their findings. I just happen to have been researching and writing about Ivermectin, and so I came across the material. There’s nothing special about me.

At this junction, if you think I am making it up as some weird grift, I don’t really blame you. So let’s dive into the literature and shore it up. Here’s a collection of published research where Ivermectin has been used as a treatment for cancer either in real patients or in laboratory experiments. Because this can get heavy going, I’ve translated the study titles and results into something everyone can understand.

1. Testing If Ivermectin Can Help the Immune System Fight Breast Cancer in Mice and Lab Cells The researchers tested ivermectin on mouse breast cancer models to see if it could make tumors more recognizable to the immune system (turning “cold” tumors “hot”). They found that the drug successfully triggered cancer cell death and helped immune cells infiltrate the tumors, especially when combined with other immune-boosting drugs.

2. Using Ivermectin to Stop Skin Cancer from Spreading to the Lungs in Mice Researchers wanted to see if ivermectin could prevent melanoma (a deadly skin cancer) from spreading to the lungs using mouse models and cell cultures. They found that the drug effectively blocked the cancer from spreading by stopping the formation of specific cellular “traps” that tumors use to travel through the body.

3. Testing How Ivermectin Blocks Growth Signals in Human Cancer Cells and Mice This laboratory study looked at how ivermectin interferes with specific growth signals that cancer cells rely on to multiply. The researchers showed that the drug successfully blocked these signals, stopping cancer cell growth in test tubes and shrinking tumors in living mice without causing significant side effects.

4. Using Ivermectin to Make Chemotherapy Work Better on Resistant Cancer Cells Scientists tested if adding ivermectin could help standard cancer drugs work on cancer cells that had developed a resistance to chemotherapy. They discovered that ivermectin successfully reversed the cancer’s drug resistance by shutting down the chemical pathways the cells use to pump out the chemotherapy.

5. Tracking Real-World Cancer Patients Who Took Ivermectin and Mebendazole Instead of a laboratory experiment, researchers observed 197 actual cancer patients who chose to take ivermectin and mebendazole off-label over a six-month period. They tracked the patients’ self-reported outcomes, safety data, and adherence to understand how these repurposed drugs affect real humans undergoing cancer treatment.

6. Exploring How Ivermectin Stops Colon and Breast Cancers from Spreading This study used both lab-grown cells and living mice to figure out exactly how ivermectin stops breast and colon cancers from moving to other parts of the body. They found that the drug successfully paralyzed the cancer cells by blocking the specific chemical signals they use to crawl and spread.

7. Testing Ivermectin Against Blood Cancer Cells in the Lab This laboratory experiment investigated whether ivermectin could destroy lymphoma (a type of blood cancer) cells. The researchers found that the drug effectively halted the cancer cells’ growth cycle and destroyed their energy centers, forcing the cells to die off.

8. Mixing Ivermectin with a Special Enzyme to Kill Pancreatic Cancer Cells Scientists tested if ivermectin could be combined with a specific enzyme (rMETase) to destroy hard-to-treat pancreatic cancer cells in a petri dish. They found that while ivermectin alone killed nearly half the cells, the combination of both wiped out 80% of the cancer cells.

9. A Massive Review of All Existing Data on Ivermectin as a Cancer Drug Rather than running a new experiment, researchers gathered and analyzed all the existing lab, animal, and human trial data regarding ivermectin’s effect on cancer. The review confirmed broad evidence that the drug stops tumor growth and reverses drug resistance, while noting that early-stage human clinical trials are actively underway.

Surprised?

For people trained in medical science or medical literature, many will be frothing with rage at what I’ve just suggested by sharing this list without qualifying context. Before we address their concerns, let’s focus our attention on how a normal person will respond to a list like this.

In the very least, this is a very promising set of data. Ivermectin has been in use since 1987, it is amongst the most prescribed medicine in history, it has an impeccable safety record. Satoshi Ōmura won the Nobel Prize for its discovery. What’s more, the drug is now generic, meaning it costs almost nothing to manufacture. Look at all these great results researchers have had using it to either kill cancer cells, stop cancer growth, or treat patients directly. It’s great right?

Well no. Not at all.

To the people who are currently raging at me, they’re upset that I’ve suggested the drug might work without citing a clinical trial. What they’ll say, quite indignantly, is that “this is not clinical data! These are in-vitro studies, or small scale self-reported data with no controls, and you can’t infer effectiveness from this dataset!” Those people are right by the way.

So what’s going on exactly?

Go and type into Google “ivermectin as a treatment for cancer” and you’ll soon (very likely) be looking at the consensus view of an industry dominated by the pharmaceutical industry. The top result (in the UK) will be MacMillan Cancer Support where Richard Simcock, the Chief Medical Officer, gives this rather precise qualification, “Ivermectin is an effective drug in the treatment of parasitic infections. However, there is currently zero real-world clinical evidence that it might be helpful in the treatment of cancer.”

Here’s the kicker, he’s absolutely correct. But to a normal person, this sounds like “see! It doesn’t work!”, and I’m here to chart a third path. It does not mean the drug doesn’t work, it means we haven’t properly tested it.

You see, the people who react with rage to my list of small studies on Ivermectin are the dominant voices in medicine. Their view, which I accept is “the correct one”, is the consensus. What they like to remind everyone of, all the time, is that we can only know if the drug works if there is a large scale Randomized Control Trial. But this is where the consensus view runs out of runway. It reveals itself not as science, but as mantra because there’s a very simple response to this assertion:

“Given the substantial body of evidence that we have, why do we not have randomized control trial data on the effectiveness of Ivermectin as a treatment for cancer?”

This is a thorny question. If they’re honest with you, they’ll have to wrestle with the incentives catastrophe at the heart of modern medicine. If they’re not, they might say the data isn’t good enough to justify a trial.

After at least sixteen years of promising data on this widely available, cheap, and safe drug’s effectiveness as a treatment for cancer, the whole cancer research establishment hasn’t run a single trial. These people genuinely wonder why trust in the medical establishment has collapsed. They think the trust was shattered by some external force, rather than the terrible incentives that sit at the heart of modern medicine. As we’ve already noted, Cancer Research UK alone has a £750 million budget, how and why didn’t they run a trial yet?

The answer is very simple, and the government is already well aware of it. A 2005 UK Government report spells it out, “The industry affects every level of healthcare provision, from the drugs that are initially discovered and developed through clinical trials, to the promotion of drugs to the prescriber and the patient groups, to the prescription of medicines and the compilation of clinical guidelines.”

The industry controls everything. So do you think they’re going to foot the bill to push a drug they can’t profit from into their own profitable market? Here is where the tension sits: if a repurposed generic drug is actually an effective therapy for cancer, the medical system is structurally incapable of making use of it.

Read that again, then let’s consider an example.

The most common cancer for men in the UK is prostate cancer. Assuming the 2005 report remains true (it is true), it’s actually the pharmaceutical industry which creates the treatment regimens for that particular illness. The approved treatment guidelines prescribe Goserelin (Zoladex), which is £940 a year. The patent for the drug is owned by AstraZeneca. You’d possibly also receive Enzalutamide, which is somewhere around £35,000 for the year, and the patent is owned by Astellas. NICE themselves list the price in their recommendation literature, “The daily dose of enzalutamide is 160 mg and costs £97.67 per day.”

You will also probably be prescribed some cheaper generic drugs too, like Docetaxel. However, because it’s IV administered (an ‘inpatient’ treatment), there are other drug paraphernalia and administration costs which can stack up to perhaps £1,500 to £3,500 a year. All things said and done, the value (ugh) “per patient” is somewhere around £40,000.

In the UK alone, cancer is projected to add £14.4 billion per year to health spending by 2050. Current direct costs are already in the region of £9–12 billion annually. Globally, spending on cancer medicines hit $252 billion in 2024 and is forecast to reach $441 billion by 2029.

In a market this size, the economic incentive to protect high-margin patented drugs — and the corresponding lack of urgency around cheap generics — becomes brutally obvious.

If you imagine the healthcare industry as more human aligned than any other industry, then perhaps you’d imagine incentives that relentlessly drive drug discovery. Unfortunately, the medical industry is no more aligned than any other market, and it’s the internal incentives that drive the outcomes. There’s zero incentive - none at all - to relitigate generic drugs as potential treatments for cancer. When you think about it, the incentives actually mean companies will campaign against any generic drug that looks promising, because if it turns out to work, then their $250 billion market gets a haircut.

I want you to really understand what these dynamics mean for our ability to ‘know’ which medicines are effective. So consider this: according to the FDA alone, there are 32,000 generic drugs already approved for use in humans. These are drugs which got onto the market because they were effective at something, but their patent expired, making them incredibly cheap and difficult to profit from. Globally, you could throw tens of thousands more drugs onto this pile. The incentive to investigate their effectiveness against diseases treated with expensive and patented drugs is close to zero. There might already be multiple wonderdrugs in our huge repository of useful generic drugs, but our medical system can never come ‘to know’.

Even if the drug looked promising in small scale trials, some boffin from Imperial College or a charity would warn the public against using it because “there’s no clinical data to support its use”. Whilst this might be true, as I’ve demonstrated, this hardly paints the full picture. Statements like this are an artefact of the incentives that drive medical research.

Then consider the range of non-patentable herbs, plants, fungi and bacteria which can affect the body. Can cannabis affect the body? Can mushrooms affect the body? Of course they can, but none of these natural products can be patented, so there’s simply no pathway for them to become treatments. It’s really that simple.

Pharmaceutical companies can’t ‘own’ a liberty cap mushroom, so it can’t become a treatment. Pharmaceutical companies can’t own Vitamin C, so it has no pathway to become a treatment. The list of natural “un-ownable” compounds which affect the body is absolutely massive. In Chinese Medicine there are at least 11,000 herb plants with documented medical uses, almost none of them will be patentable. One database of traditional Chinese medicine has 29,000 ingredients listed. If you throw in ancient Ayurveda, we’re talking about hundreds of thousands, perhaps millions of natural compounds, already known about, which sit beyond the reach of pharmaceutical patent lawyers, and therefore beyond the reach of hospital bedsides.

The modern search for medicine has access to an incredible modern Library of Alexandria, but restricts its search only to the compounds synthesized that year because it fears any answer which might arrive from the public domain.

Once you see the incentives baked into the system, it becomes clear why we still don’t “know” whether ivermectin works for cancer. The industry has effectively defined the boundary between “knowing” and “not knowing.” Until a drug passes through the expensive, patented-friendly randomised controlled trial pipeline they dominate, it remains in limbo — officially “unproven,” regardless of how much preliminary evidence exists.

This is not necessarily a conspiracy. It is simply what happens when profit is the dominant incentive.

The kayfabe continues. Good doctors do their best with the tools they’re given. Researchers diligently hunt for new molecules. Cancer charities raise hundreds of millions and rightly celebrate incremental progress. And patients are left navigating a system that is simultaneously miraculous and structurally blind to cheap, unownable solutions.

What does this mean for you and me?

It means we must learn to stand humbly in the unknown. We should acknowledge that what we are allowed to “know” in medicine is shaped by a highly filtered corpus of knowledge — one optimised for patentability and profitability as much as for truth.

The quiet revolution begins when ordinary people reclaim the right to ask simple, dangerous questions:

“If there’s a cheap, safe, generic drug with promising signals against cancer, why haven’t we properly tested it?”

Asking that question honestly is the first step toward restoring trust — not in institutions, but in our own capacity to reason clearly about what is known, what is unknown, and what incentives quietly shape the difference.

The comic, “Palomar Lights” tells a true, and very strange story: have astronomers detected the remnants of something floating in earth’s orbit?

I hope you enjoy it. It’s certainly different subject for readers of The Digger! As ever, shares, links, mentions, comments and subs are always helpful.

I can’t post this entirely on substack, so please click the image below to read it. Alternatively, just head straight to comics.phillyharper.com

Read the full comic for free at comics.phillyharper.com

I encouraged Claude and Gemini to (somewhat) collaborate on compiling a dataset of all the fires at oil refineries over the last six weeks. Then I requested they choose five other six week periods - at random - and check that period for the number of oil refinery fires.

After comparing the current six week period to those random intervals, Claude and Gemini both agree that the rate of fires is very high. Most of this is driven by the war in Ukraine and the Gulf, but even when you exclude those areas, the rate of fires remains highly unusual.

From the report:

“Compared with a Gamma prior fitted to five randomly-chosen historical 6-week windows (mean ≈ 2.2 fires / window), the total count is extreme (tail probability ≈ 5 × 10⁻⁵; Bayes factor ≈ 10⁶). Even after stripping out the war-related strikes, the non-war count of 8 sits in the top ~1% of the prior-predictive distribution (Bayes factor ≈ 92). The conclusion is that the cluster is unlikely to be random noise, but the dominant signal is geopolitical — specifically Ukraine’s systematic targeting of Russian refining infrastructure — rather than a hidden global-industrial anomaly.”

Have a look at the report itself, which is attached below, are there any events missing from the dataset? Is there anything else we can do to improve this? Comments open.

Refinery Fires Bayesian Check

309KB ∙ PDF file

Download

Download

Göring: Why, of course, the people don’t want war. Why would some poor slob on a farm want to risk his life in a war when the best that he can get out of it is to come back to his farm in one piece? Naturally, the common people don’t want war; neither in Russia, nor in England, nor in America, nor for that matter in Germany. That is understood. But, after all, it is the leaders of the country who determine the policy and it is always a simple matter to drag the people along, whether it is a democracy, or a fascist dictatorship, or a parliament, or a communist dictatorship.

Gilbert: There is one difference. In a democracy the people have some say in the matter through their elected representatives, and in the United States only Congress can declare wars.

Göring: Oh, that is all well and good, but, voice or no voice, the people can always be brought to the bidding of the leaders. That is easy. All you have to do is tell them they are being attacked, and denounce the pacifists for lack of patriotism and exposing the country to danger. It works the same way in any country.

No normal person wants war.

That is why war planners developed The Pattern — a repeatable multi-stage sequence designed to engineer consent for operations the public would otherwise reject. In this series I’m going to show you the full Pattern in action. Part One begins with the birth of modern U.S. interventionism and the very first time the template was tested: the 1954 Guatemalan coup. By the end of this piece you’ll already recognise the first two stages. In the parts that follow we’ll see the final stages of the pattern, (the jackals) and see how the same playbook has been updated for 2026.

We don’t live in feudal times where Kings have access to a warrior class like the Kshatriya in ancient India, the Spartans in ancient Greece, or the Samurai in ancient Japan. If the war planners decide war is needed, they must bring the public with them. This makes you a component of war. Your beliefs, your ideas and even your thoughts about it matter deeply to them. They need just enough support, and so detailed plans are drawn up to maneuver you into position.

There is a pattern to all of this, and once you recognise it, you can’t unsee it. War planners bank on you not recognising their methods, so my aim is to reveal them so they have a harder time pushing their wars onto the public.

The birth of constant war

The way Iran is being treated in 2026 has its roots in how America came out of WW2 eighty years ago in 1945. At that time, American planners looked at their situation and literally made a plan to dominate the entire world: not a conspiracy, a well documented part of America’s modern history. It was called the “Grand Area” strategy, and was the key vision of America’s post war plan.

Fearing the Nazi regime might claim most of Europe, planners came to a stark conclusion: “The United States should use its military power to protect the maximum possible area of the non-German world”. According to the planners, to maintain peace and America’s position within it, the United States would need to economically and militarily dominate basically everywhere other than continental Europe and the Soviet Empire. The planners believed this was key to America’s survival.

‘The Grand Area’ plans were afoot before the war ended, and they were based on the assumption that Europe and North Africa would fall to the Nazi regime, but when Britain didn’t capitulate to the Nazi onslaught, and the prospect of a Nazi Europe dissolved, ‘The Grand Area’ plan remained.

Why?

Well there was still one major player left on the table; the Soviets. America would use its new economic and military power to ‘protect the maximum possible area’ and box the Soviet influence inside central Asia. With the Nazi regime gone, and the Soviet machine locked behind the Iron Curtain, the final fear was that communism might ‘blow over’ the border like a dandelion seed and ‘infect’ something inside The Grand Area.

This was the start of the pathological fear which grew inside America, a fear of anything that sounded or looked even remotely communist. Perhaps the experimentation with cannabis and LSD in the 60s didn’t help matters, but soon enough, a deep paranoia emerged that communist ideas were taking root. Any suggestion of communism, anywhere at all, would be met with a new strategy of American sponsored violence, and the intelligence agencies were at the heart of it all.

The era of constant war had begun, and as it began, ‘The Pattern’ of war emerged with it.

The 1954 Guatemalan Coup D’Etat

You will hardly find a clearer example of ‘The Pattern’ than the CIA orchestrated coup which took place inside Guatemala in 1954. Jacobo Árbenz, the democratically elected leader of Guatemala, had the temerity to distribute Guatemala’s own uncultivated land to the poor so they could grow a life for themselves. Árbenz compensated the land owners with 25 year bonds at 3%, but this didn’t alleviate American fears. The reforms triggered America’s communist alarm and landed Árbenz on the radar of the State Department.

An internal memo by Charles Burrows, the State Department’s Inter-American Bureau Officer, beautifully revealed the mindset at that time. Because Guatemala’s land reforms were popular and helping workers and peasants, it might inspire neighbouring countries to do the same:

“Guatemala has become an increasing threat to the stability of Honduras and El Salvador. Its agrarian reform is a powerful propaganda weapon; its broad social program, of aiding the workers and peasants in a victorious struggle against the upper classes and large foreign enterprises, has a strong appeal to the populations of Central American neighbors, where similar conditions prevail.

The United Fruit Company had investments in Guatemala, so with their business interests in crosshairs, the CIA decided Árbenz had to go. So how would planners overthrow a government if the public wouldn’t support it? They’d try to do it without anyone noticing. Declassified documents have dragged this fiasco from the realm of conspiracy theory and into sunlight of documented reality. In their own words, the CIA conspired in “the development of a covert action program designed to topple the Árbenz government”. The two operations were called PBSUCCESS and PBFORTUNE.

If you can believe it, the operation kicked into gear when a United Fruit Company lobbyist walked into the CIA office and ‘suggested’ that the Guatemalan President be ousted. This was before Árbenz had even been elected! This kick-started a five year collaboration to install a government more favourable to their corporate interests. It’s really worth remembering the real seed of this coup, because as we move through the story, you’ll be washed along with the subterfuge, lies and distortions that we discover. The seed of the war is constant, and is an important datapoint to keep in mind: it was corporate interests partnering with military enforcers.

The relevance of this period of history is how the Guatemalan coup set the template for the psyops and lies we are subjected to today. When the CIA started planning the coup, the agency had only existed for three years. Getting rid of Árbenz would be one of their first major foreign projects, they had to get it right, and its ultimate success set the tone for many more interventions to come. Guatemala was the birth of The Pattern.

Part One: Isolate, Economise, Destabilize.

Because people don’t want war, the troops can’t just roll in. You must try to achieve your objectives in a way people won’t notice. Absolutely critical to all but the very last stages of wars is plausible deniability. To be crystal clear, this is the capacity for the ‘normal’ bits of the American government to deny any and all involvement any military action, coups, murder or chaos. Either because the Senators and Congressmen literally don’t know about the operation, or because there’s a plausible cover story which everyone (including the CIA itself) believes. Propaganda becomes internalised, and even the CIA themselves start to believe their own stories.

Don’t take my word for it, the CIA themselves committed these ideas to paper.

The first stage of The Pattern is to use America’s economic power to choke out the target before any shots are fired. We still see the choke out today; when you hear of sanctions against some country or regime you know very little about, that nation is probably on a pathway to regime change. Normally, planners inside that target nation are aware they’re on a pathway to regime change, so consider their actions in that context. Economic warfare is always the first round.

In Guatemala, according to the State Department’s own words, “it is our policy to discourage the Export-Import Bank and other official agencies from making developmental loans to Guatemala.” In their own words, the CIA created “operations for cutting off military aid to Guatemala, increasing aid to its neighbors, [and] exerting diplomatic and economic pressure against Árbenz.” Do those statements, now more than 70 years old, remind you of our current times?

With the Guatemala’s neighbours now well armed, and no access to banking or financial services, the next stage of The Pattern is to target the economy directly. Noting how vital coffee exports were to the Guatemalan economy, the CIA informed planners that “A study is under way to determine what phases of the coffee industry may be attacked which will damage the Árbenz government and its supporters” This remarkable document is still hosted on the State Department’s own website, and from it we learn how the CIA collaborated with its “partners” to wage economic warfare on vital Guatemalan exports.

By ‘partners’, we’re absolutely sure one was the United Fruit Company, along with other large oil, utility, food and media corporations. The CIA also planned for “American businessmen in New York City...to put covert economic pressure on Guatemala by creating shortages of vital imports and cutting export earnings.” As you can see, the CIA was a network of assets inside corporations, governments, business, academia and media. These methods of war - private partners creating economic pressure on their targets - are the same methods we are subjected to today.

In Guatemala, ultimately, the CIA became frustrated by their economic efforts. Boycotting Guatemalan coffee caused knock on effects and problems for their other partners, and the other economic pressures they’d applied would take longer to materialise. Once you’ve run out of road, it’s time to turn up the heat.

Part two: Fabricate evidence and Amplify your narrative

It’s one thing to get your ‘partners’ in business and media to collaborate with you, but how do you drag along your allies and other regional players? You’d need something quite serious to convince them to join in your covert operation. Remarkably, the CIA were quite candid about this: “Real, or when necessary fabricated evidence regarding aggression and subversion, will be used at an OAS conference to obtain approval of the American States for multilateral economic action against Guatemala.”

Read the statement again. “When necessary fabricated evidence regarding aggression and subversion will be used.” It proves that a shady tentacle of the US Government will fabricate evidence to obtain support for their wars. To be clear, the OAS is the Organization of American States, often described as the Western Hemisphere’s version of the United Nations. The CIA were prepared to fabricate things to convince other nations to join their “pile on” of Guatemala.

It’s worth remembering that this was 1953. Far from disqualifying this to an irrelevancy, it elevates the behaviour into the very soul of the CIA. Fabricating evidence to start wars is in their DNA. They’ve been doing it since day one, so keep that front of mind.

So what kinds of things did they fabricate?

Well, they were looking to drum up support for their planned military coup against Árbenz, so in an “elaborate fabrication”, they “planted a cache of Soviet-made arms on the Nicaraguan coast to be ‘discovered’ weeks later by fishermen.” Their plan was to pin this whole made up saga on the Guatemalan government. It was a “Soviet Submarine” they had planted, which (according to the fabricated story) had landed there with the full collaboration with the Guatemalans. Frustrated at the lack of response to their fake news, the CIA urged their local spies to “make a cautious effort to obtain added publicity for this story in the Guatemalan press and public opinion.” It’s worth reading the original document which is posted below, because it’s fascinating to see it in black and white.

The target for these psyops were the Guatemalan people and politicians, and these bizarre interventions were having some success. But what to do about convincing their electorate at home? To convince the American public of their plans, the CIA would have to work with real experts in propaganda. Together with the United Fruit Company, they turned to Edward Bernays, a man with a proven talent for propaganda. In fact, he’d literally written the book on the matter in 1928 where he said “The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society. Those who manipulate this unseen mechanism of society constitute an invisible government which is the true ruling power of our country.”

Over the years, Bernays had developed a seriously good propaganda service which he was now touting to the highest bidder, including the CIA and United Fruit. So Bernays set to work in a collaboration that is well documented in the book “Bitter Fruit”.

The project was started when United Fruit came to Bernays with concerns about their old fashioned ‘colonial’ image, which had its roots in the actual ways they treated their workers. Undeterred by this, Bernays had a solution: paint United Fruit as a modern company surviving on the frontlines against a wave of central American communism. By casting Guatemala as a Soviet stronghold, United Fruit could become the hero of the story. With connections to all the major publications, it was trivial for Bernays to arrange for a journalist, Fitzhugh Turner, to travel to Guatemala with the New York Herald Tribune. The series Turner produced, “Communism in the Caribbean” was pivotal in shaping American perceptions about the tiny central American state.

Not content with one success, Bernays continued to work his magic in an incredible PR Blitz which was documented by the CIA themselves.

“NBC News aired a television documentary,’Red Rule in Guatemala,’ revealing the threat the Árbenz regime posed to the Panama Canal. Articles in Reader's Digest, the Chicago Tribune, and the Saturday Evening Post drew a frightening picture of the danger in America's backyard. Less conservative papers like the New York Times depicted the growing menace in only slightly less alarming terms.”

Just some of Bernays’ successes

But here’s the thing...Árbenz wasn’t a communist. Remember I said you’d get washed up in the lies and subterfuge?

In his inaugural address to the nation, Árbenz promised to convert Guatemala from "a backward country with a predominantly feudal economy into a modern capitalist state." The CIA were fully aware of this. They absolutely knew that Árbenz was not a communist. They knew that Guatemalan opposition to his land reforms was “greatly exaggerated”, even amongst the wealthy people there. In the CIA’s own words, “Rather than setting up a Communist state, Árbenz desires to establish a “modern democracy” which would improve the lot of its people through paternalistic social reforms.” The reality was, as we’ve already seen, the CIA’s beef with Guatemala was about a policy Árbenz implemented which upset their sponsors at The United Fruit Company.

Despite this... by 1954, Bernays, the CIA, and United Fruit really had convinced the American public that Árbenz was a Soviet puppet. This is a critical hallmark of The Pattern: the massive gap between what happened and what the public believe happened. Once a narrative takes root, contradictory evidence often fails to dislodge it. Social psychologists call this belief perseverance, the well-documented tendency for people to cling to their initial understanding even after the original “facts” are proven false.

The American public had absorbed the “Red Jacobo Árbenz” story so completely that when the CIA plot against Árbenz was actually revealed, they simply didn’t believe it. In a strange twist of fate, a CIA asset hospitalised with a stomach ulcer had his secret documents taken by a double agent working for Árbenz, giving him “intimate knowledge of rebel training bases, intelligence operations and a fairly accurate concept of the modus operandi of PBSUCCESS.”

The Guatemalan president went public with the evidence; the propaganda, the rebel forces being CIA assets, and the overall CIA plot to overthrow him. Árbenz held press conferences, showed the documents, the whole works, yet, the American press and public simply didn’t believe it. Instead, the press “unanimously accepted the State Department’s characterization of the charges as a propaganda ploy”

The American public had come to believe an inversion of the truth. When an operational hiccup catapulted the actual truth into their hands, they didn’t believe it. They couldn’t believe it, an effect the war planners have come to utilise with a deadly effectiveness. That is why the first story matters so much, because it stubbornly persists.

Are there things today that parts of Europe and America have been fully convinced of, which, despite evidence to the contrary, they simply cannot allow themselves to let go of?

In Part Two we will watch what happens when the economic chokehold and the propaganda blitz still aren’t enough. That is when the jackals are sent in — the covert military action, the proxies, the coups. Once you see the full Pattern, you can’t unsee it… and you’ll start noticing it everywhere, including in 2026.

But where was the case?

The bombs started falling on February 28th. It wasn’t until March 2nd that the White House published what was supposed to be the definitive justification: a document boldly entitled “The Iranian Regime’s Decades of Terrorism against American Citizens,” listing 44 incidents and 992 American deaths. It was presented to the public as an authoritative, solemn accounting.

There was just one problem. The data didn’t come from a three-letter intelligence agency. It didn’t come from the Pentagon or the State Department. The source wasn’t given at all.

So where did it come from?

Over the past two weeks, this investigation — supported by paid subscribers to The Digger — has traced the White House case for war back to its origins. What we found is, I believe, one of the most significant breaches of public trust in modern American history. Here is the full story.

Part 1: The Plagiarism

We owe a debt of gratitude to Stephen McIntyre (@ClimateAudit), who first noticed that the White House statement looked familiar. I verified his findings independently, stripping away editorial introductions and running both texts through a standard data-science comparison algorithm. The results were undeniable.

The White House’s case for war was copied, virtually word for word, from a report published by the Foundation for Defense of Democracies (FDD) — an Israeli-aligned think tank whose original name was EMET (Hebrew for “truth”), founded with the explicit goal of enhancing Israel’s image in North America. The FDD report was written by Tzvi Kahn, a former Assistant Director for Policy and Government Affairs at AIPAC — the most powerful Israel lobbying organisation in America.

The pipeline: a former AIPAC official writes a document at FDD. The White House copies it word for word. America goes to war.

But it gets worse. We know from leaked undercover footage filmed by Al Jazeera that the FDD is not just a think tank. Sima Vaknin-Gil — a former Israeli military intelligence officer serving as Director General of Israel’s Ministry of Strategic Affairs — was caught on hidden camera explicitly claiming the FDD as Israel’s asset:

“We are a different government working on foreign soil and we have to be very very cautious. We have three different sub-campaigns that are very very sensitive… We have FDD. We have others working on this.”

The White House used material from an organisation claimed by Israeli intelligence to justify a war to the American public.

Read Part 1 in full:

Part 2: The Author

The immediate question was: who inside the White House did this? The document didn’t write itself. A specific person sat at a keyboard and copy-pasted Israeli lobbying material into an official government publication.

Hidden in the page metadata — invisible to ordinary readers but embedded in the White House content management system — was the answer. The author field reads: Jacob Schneider.

<script type="application/ld+json" class="wp-parsely-metadata">
{
  "@context":"https:\/\/schema.org",
  "@type":"NewsArticle",
  "headline":"The Iranian Regime’s Decades of Terrorism Against American Citizens",
  "url":"http:\/\/www.whitehouse.gov\/articles\/2026\/03\/the-iranian-regimes-decades-of-terrorism-against-american-citizens\/",
  ...
  "author":[{"@type":"Person","name":"Jacob Schneider"}],
  "creator":["Jacob Schneider"],
  ...
}
</script>

Schneider’s title is “Special Assistant to the President and Rapid Response Director.” He has a BA in Political Science from Bethel University. His entire career has been in campaign communications — RNC rapid response, Trump campaign rapid response. He has no foreign policy expertise, no intelligence background, and no Middle East credentials whatsoever. His job is propaganda.

How did he get there? Schneider himself told a local newspaper in 2020 that he was recruited via Twitter: “He’d met a couple people through Twitter and they reached out to him to ask if he would be interested in moving to Washington, D.C. to work on the campaign. He said yes… and ended up moving to Washington, D.C. within about 10 days.”

I checked his Twitter interactions from June 2020. At the exact moment he claims he was recruited, Schneider was publicly interacting with someone from the Center for Security Policy — a neoconservative war lobby that has campaigned for regime change in Iran for decades. That same organisation held at least two key National Security Council positions in the Trump administration at the time.

The CSP’s founder, Frank Gaffney, is a former “Distinguished Adviser” to the FDD which provided the source material for The White House war. The organisations are connected at every level. The same network that produces the propaganda also appears connected to the recruitment of the man who launders it into the White House.

Read Part 2 in full:

Part 3: The Machine

Parts 1 and 2 could have been a story about one lazy staffer copying one document. Part 3 proves it’s something far more systematic.

Five people write White House articles about Iran, Israel, and Gaza. Four of them never quote a single outside expert. The fifth — Jacob Schneider — quotes nobody but outside experts. And every single one of them is connected to the same Israeli lobbying organisation.

I went through all 20 of Schneider’s articles and extracted every person and organisation he quotes. Every expert, every think tank, every polling firm. Roughly 350 names. Then I categorised them.

Across 19 of his 20 articles, 100% of the non-government think tank voices trace back to the Foundation for Defense of Democracies or its network.

One hundred percent. Not 80%. Not “mostly.” Every single one.

When Schneider needed “independent” nuclear experts to validate the bombing of Iran, he quoted four people:

Andrea Stricker — an FDD employee who spent 12 years at David Albright’s Institute for Science and International Security before moving directly to FDD.

David Albright — Stricker’s former boss, who co-authored a 155-page Iran strategy report with FDD’s CEO calling for “overt preparations for warplanes and missiles to destroy Iran’s nuclear capabilities.” That was in 2013. In 2026, those exact recommendations became White House policy.

Spencer Faragasso — Albright’s colleague at ISIS who co-authors reports with him. Same organisation quoted twice.

Brian Carter — American Enterprise Institute, which is funded by the same pro-Israel mega-donors as FDD. Paul Singer gave $1.1 million to AEI and $3.6 million to FDD. Bernard Marcus, FDD’s largest donor, provides roughly a third of their budget.

Four “independent” experts. One network.

And then we found a plausible route by which the material gets into the White House. Richard Goldberg is a Senior Advisor at FDD. John Bolton brought him onto the National Security Council in 2019 — and FDD continued to pay his salary while he served in government. An FDD lobbyist, paid by the lobby, sitting on the National Security Council. Goldberg then re-entered the White House in late 2024 as Senior Counselor to the National Energy Dominance Council — directly overlapping with the period Schneider was publishing.

On top of that, FDD Action openly advertises a “Policy Support Desk” providing White House staff with free policy analysis, private briefings, and legislative drafting assistance. They market it on their website. The pipeline isn’t hidden. It’s just not something anyone had mapped until now.

Read Part 3 in full

What Comes Next

This investigation is not finished. Subscribers to The Digger have funded the work so far, and with their support, here’s what we’re going after next:

The Vandenberg Coalition. The polling Schneider cites as evidence that Americans support the war comes from an organisation with six FDD personnel on its advisory board — including FDD’s CEO and founder. FDD helps justify the war, an FDD-linked organisation polls on support for it, and the White House cites that poll as independent evidence. It’s a closed loop. We’re going to map it fully.

The money trail. Paul Singer and Bernard Marcus fund both FDD and AEI. We’re going to trace the full donor network behind the organisations whose experts populate Schneider’s articles.

The other authors. We’ve focused on Schneider so far. But Allison Schuster, Micah Stopperich, Mary Wales, and Jacqueline Kotkiewicz also write on these subjects. Their articles are formal government instruments — executive orders, presidential memoranda. Do those also carry fingerprints? We’re going to check.

The FDD Action “Policy Support Desk.” FDD’s lobbying arm explicitly offers policy analysis, briefings, and legislative drafting to White House staff. We want to know: who uses it? What materials have been distributed? This could be subject to FOIA requests by American researchers.

This investigation exists because subscribers fund it. Every part of this series — the data collection, the source analysis, the verification — was done with the support of people who believe this kind of journalism matters.

If you’ve read this far, you understand what we’ve found: a foreign lobbying operation has successfully built a pipeline to place its propaganda directly into official White House communications, presented to the American public as independent government analysis, to justify a war.

If you want this investigation to continue, please consider subscribing. And if you’re already a subscriber — thank you. This work would not exist without you.

Share this. The more people who understand how the case for this war was manufactured, the harder it becomes to manufacture the next one.

What at first appears to be a comms staffer plagiarising one document has grown into something more concerning. When I started digging deeper into his output, what I found was far more systematic than a single copy-paste job. I found a propaganda machine. A closed loop of Israeli-aligned organisations producing the research, providing the “expert” voices, polling the public on the results, and then feeding it all back into official White House communication. A pipeline of Foreign government propaganda published under the letterhead of the White House.

This is Part 3, and it’s the part that sheds light on the pipeline of foreign propaganda finding its way into Executive Office. This is about a foreign government manufacturing the consent of the American people via The White House. Let’s get into it.

Schneider’s Factory

Since January 2025, Jacob Schneider has published 20 articles on the White House website covering Iran, Israel, Hamas, Hezbollah, and Gaza. That’s more than every other White House author on these subjects combined.

We know it’s him because his name appears in the metadata of every single one, as we showed in Part 2. His title is “Special Assistant to the President and Rapid Response Director.” His job, in plain English, is propaganda. Fast propaganda.

I wanted to see if there was a pattern to Schneider’s use of FDD propaganda. Using a search for White House articles on Iran, Israel, Hezbollah, and Gaza, I went through the first 20 articles authored by Schneider and extracted every single person and organisation he quotes. Every expert, think tank, polling firm and tweet, and then I categorised them.

The results were staggering.

Schneider quotes roughly 350 people across his 20 articles. Most of them fall into predictable categories: Trump administration officials, Republican lawmakers, foreign leaders, and right-wing media personalities. But when you isolate the “independent experts” – the think tank analysts, the nuclear specialists, the foreign policy scholars that Schneider cites to give his articles the veneer of credibility – a very different picture emerges.

Across 19 of his 20 articles covering Iran and military operations, 100% of the non-government think tank voices trace back to the Foundation for Defense of Democracies or its network.

One hundred percent.

Let me say that again. Every single “independent” expert that Jacob Schneider quotes to validate the case for war with Iran is connected to the Israel claimed intelligence operation the FDD. The conclusion is simple: someone has successfully setup a pipeline to launder Israel claimed propaganda directly into the White House. American citizens are receiving foreign propaganda under the iconic White House letterhead.

It might be hard to believe, but I’m going to show you the receipts. So let’s get into it.

Before we go further, an important point: all of the White House articles I’m discussing here do not have a traditional ‘Schneider’ byline. This is very important when you consider everything I’m about to tell you. To the millions of Americans (and global citizens) who read these articles, they are written by “The White House”. What’s written under that letterhead, and how it got there, carries huge weight because it carries with it the power of the executive branch of the American government. We only know literally who wrote articles because the author names appear in non-visible metadata.

Read more

The whole saga raises many questions which need answers: Trump campaigned extensively on the anti-war ticket, he campaigned extensively on “draining the swamp” of deep state operatives, he campaigned on not starting a war with Iran, but in his 2026 presidency, his White House Staff are quoting verbatim from deep state war hawk propaganda.

How did this come to happen?

As the keyboard was being tapped somewhere in Donald Trump’s White House, what exactly was happening? Who was actually writing out that case for war? And who were they in contact with?

We can know with some confidence who laundered Israeli propaganda, because their name appears as the author in the pages’s metadata. Whilst they might be a professional in communications, they perhaps overlooked the power of their own Content Management System. Hidden in the metadata of the White House’s article, the author is right there, and learning how that person got there will help us understand how the deep state really operates...

Read more

This is an extremely serious breach of security, and it’s a serious breach of the trust of the American people. The White House absolutely should not be laundering Israeli lobbying propaganda. It demonstrates just how far up the decision chain Israeli influence has reached.

Follow me on twitter @phillyharper Now...read on.

Ghostwriting a War (Part 1)

On February 27, 2026, while traveling aboard Air Force One to Texas, President Donald J. Trump issued the executive authorization for Operation Epic Fury. The coordinated U.S. and Israeli military campaign commenced the following day, striking across the Iranian homeland.

The death of Supreme Leader Ayatollah Ali Khamenei and numerous top-tier military and intelligence commanders has plunged the region into a dangerous crisis, triggering immediate retaliatory actions from Iranian proxy forces across Iraq, Syria, and the broader Middle East. America, and perhaps the world, is now embroiled in a very serious war. Global energy security has been destroyed, and there look to be very few off-ramps.

The Vacuum of Intelligence

Historically, preemptive or massive retaliatory military campaigns of this magnitude are preceded by the declassification of formal intelligence, such as National Intelligence Estimates (NIEs). These documents are meticulously crafted by the Office of the Director of National Intelligence, and normally, they’re used to demonstrate the imminent nature of a threat to the public, Congress, and the international community.

The responsibility for this sits with Tulsi Gabbard as she is currently serving as Director of National Intelligence.

But it appears that Trump has acted on impulse.

He didn’t consult Congress, the American public, or seemingly anyone at all. Tulsi Gabbard’s statements about what happened are cryptic, but crucially, she has not said that Iran posed an imminent threat.

As the world was forced into a dangerous new phase, there was a vacuum of information where there should have been congressional backing, National Intelligence Estimates, and public onboarding. Having plunged America into another foreign war, the press, Congress, MAGA, and the public wanted answers.

So where is the case?

The bombs started falling on February 28th. It wasn’t until March 2, 2026, when the White House published what was supposed to be the definitive document justifying the current Israel-US war on Iran. The statement, boldly entitled “The Iranian Regime’s Decades of Terrorism against American Citizens,” opened with a brief editorial before listing 44 incidents resulting in a highly specific 992 US deaths.

It was presented to the American public as an authoritative, solemn accounting, but unlike the lead-up to the Iraq war, or the endless Russia collusion allegations, this data didn’t come from a three-letter intelligence agency. The source of the data wasn’t given at all.

So, where did the intelligence come from?

The entire case used to justify the war came from the Israeli lobby. That’s a strong accusation, but as you’ll see, the data strongly supports it and the implications are profound. The White House adopted its case for war, wholesale, from a think-tank report prepared by a former AIPAC employee and current staffer inside an Israel lobby.

We owe a massive debt of gratitude to Stephen McIntyre (@ClimateAudit on X), who first dug into the data and painstakingly traced the White House statement back to its original source. I have gone through and verified McIntyre’s findings myself, standardizing the data to isolate just the list of alleged incidents. The results are undeniable: Israeli special interests have reached the very highest echelons of American power and forced through their case for war.

The Propaganda Pipeline

On June 19, 2025, on the eve of the first US bombing of Iran on June 21, 2025, a think tank called the Foundation for Defense of Democracies (FDD) published an article titled “Iranian and Iranian-Backed Attacks Against Americans (1979-Present).” Did they know an American attack on Iran was coming?

So what is this organisation? The FDD was originally founded under the name EMET (Hebrew for ‘truth’) with the explicit goal “to provide education to enhance Israel’s image in North America.” The author of this specific June 2025 report is Tzvi Kahn, a former employee of the American Israel Public Affairs Committee (AIPAC).

But the FDD is much more than a public relations outfit. It is an operational node for Israeli intelligence. We know this because of investigative reporting already in the public domain, and it’s worth understanding clearly: the FDD is a cut out of Israeli intelligence.

In 2018, Al Jazeera produced a months-long undercover documentary investigating the Israel lobby in the United States. While the film was heavily censored following massive diplomatic pressure, leaked footage revealed explosive admissions from top-tier Israeli intelligence officials.

In the undercover footage, Sima Vaknin-Gil — a former Israeli military intelligence officer who was then serving as the Director General of Israel’s Ministry of Strategic Affairs — was caught on hidden camera explicitly claiming the FDD as an asset of the Israeli government. Discussing covert projects including data gathering, psychological warfare, and intelligence operations, Vaknin-Gil stated:

“We are a different government working on foreign soil and we have to be very very cautious.

We have three different sub-campaigns that are very very sensitive.

…

This is something that only a country, with its resources, can do the best. We have FDD. We have others working on this.”

Knowing this is incredibly important, because it means we can trace a direct line from the intelligence shared to convince the American public of the case for war, and Israeli intelligence operations. The White House, apparently, copy-pasted intelligence from an Israel controlled lobbying group to justify an American war against Iran. That objective has now been achieved, something Israel and Netanyahu have been campaigning on for 30 years.

So how do we know?

When you put the FDD’s June 2025 report side-by-side with the White House’s March 2026 statement, the evidence is undeniable. Someone at The White House copy-pasted the lobbyist document almost word-for-word. Where the White House did make slight edits, they were exclusively to ratchet up the underlying allegations, transforming Israeli think-tank speculation into the hard data needed to sell a war to the American public and the world.

The Visual Proof: Tracing the Edits

To verify this pipeline beyond a shadow of a doubt, I stripped away the editorial introductions and isolated just the core data lists from both documents. I then ran both texts through a standard Python data-science library called difflib, which allows us to mathematically and visually compare what is different between two sets of text.

What you are looking at in the images below is the output of that algorithm. On the left is the FDD’s lobbying document making the case for Iran’s threat to the USA; on the right is the White House’s official case for war. Here is the key to understand what you’re looking at:

• Red: Text that was deleted (or sufficiently re-written) from the original FDD lobbying document.

• Green: Text that was added by the White House.

• White: Exact, undeniable, word-for-word matches.

• Yellow Highlights: Granular, word-level edits within matching sentences.

Even where there’s no exact match, the data is qualitatively the same. It is obvious if you take the time to read the comparison that the White House used the FDD’s lobbying document as its source. But we need proof, and difflib gets us there.

Where the text is white, it demonstrates an exact copy-paste match. There’s no way out for the White House. This algorithm proves the source material for their case for war came directly from the FDD — an organization explicitly claimed as an asset by Israel’s Ministry of Strategic Affairs. The data proves, unquestionably, that the White House bypassed its own intelligence apparatus, opting instead to use a foreign government’s propaganda as its direct source material to justify a war.

The library is powerful enough to show us, in yellow and green highlights, the exact and granular edits the White House made for tone and aggression. For example, “Iran” becomes “Iranian regime”; “Hamas suicide bombers” becomes “Iranian-backed Hamas suicide bombers”. You can see the full difflib HTML page here.

Anyone with knowledge of copywriting or editing will recognize these changes straightaway. They are taking foreign source material and reworking it for an American audience. It’s done by someone who’s a communications professional, someone with a good understanding of audience and tone.

Part 2: The Inside Man

It is difficult to overstate the severity of this as both a breach of trust and a catastrophic breach of national security. Israel’s lobby is now so effective, and so far up the decision chain, that its own source material is being laundered into The White House’s communications — and these aren’t just any communications. These are the direct justifications for a war which could soon see American boots on the ground. It’s the data used to justify a war that has plunged the world into an oil crunch, and a war that could drag on for years.

The immediate, burning question that every journalist in the White House briefing room should be shouting is this: Who drafted the March 2, 2026 statement?

This document did not write itself. A specific staffer was tasked with compiling this justification. Under what instruction were they operating when this was written? Are foreign operatives working inside the National Security Council or the White House communications team?

As a matter of fact, we don’t have to guess.

Whoever wrote this White House document may have been a professional in communications, but they overlooked a few details…

I know exactly who wrote this document.

I know their name, I know their title, and I have the archival data suggesting they were recruited by a hawkish D.C. think-tank and air-dropped into Donald Trump’s inner circle in less than ten days.

In Part 2 of this investigation, I am going to reveal the identity of the author, name the operatives who recruited him, and expose the exact “revolving door” pipeline that placed an Israeli proxy at the absolute highest echelon of American power.

Subscribe below so you don’t miss the drop.

Follow me on twitter @phillyharper