Claude vs Francis

Using AI to expose BS

I’m interested in people. I think they’re fascinating, as well as sometimes bloody irritating. I’m going through my own c**p at the moment, with former friends making absurd claims and basically going a bit loopy.

The conspiracy mindset can do that to you.

That’s not to say we shouldn’t explore the world’s shadow. However, if we choose to do so, we need to have our wits about us.

Far too many conspiracy-minded people get sucked into delusional and paranoid thinking. They externalise too much. They take very little responsibility for what is happening in the world. They paint a picture of reality in which we must be constantly on guard against every possible manipulation, while downplaying our innate capacity for discernment. It’s disempowering and ultimately serves only the fear-mongering loosh farmers.

We are all connected. We are powerful. Good deeds and good thoughts can change things.

I’ll give an example of why I don’t stress about who is “controlled” and who isn’t. (And you can bet your bottom dollar I’ll be accused of sowing complacency and putting everyone in danger by underplaying the seriousness of the situation—which is, of course, another symptom of the mindset in question.)

I used to listen to Alex Jones for three hours a day for a couple of years. I don’t know to what extent, if any, he is controlled, but I have no desire to listen to him now. At the time, however, I enjoyed his content. It was what I needed for where I was in my life. Because I was working on myself—developing discernment and trying to become a better person—I naturally moved on and sought a different form of nourishment.

I didn’t get bogged down in “exposing Alex Jones” because it seemed like a waste of time.

If someone came to me and repeated one of his claims that I disagreed with, I’d simply say I disagree and move on. To do otherwise is often just another form of shooting the messenger, and it works hand in hand with guilt by association. Both are ways of avoiding the actual issue at hand, whatever that may be.

There is plenty he has said over the years that strikes me as accurate or insightful. Do we dismiss all of that because he may be an agent? Of course not.

We use our discernment. We take each claim on its own merits and evaluate it accordingly. If Adolf Hitler gave me good advice on how to paint a watercolour, I wouldn’t dismiss it because of his politics. (And yes, I’ve watched Europa, so chill out.)

This is one of the dangers of much of what gets presented as “research”.

For example, in Francis O’Neill’s recent piece about Ed the Techie and his connections to Bob Moran, we can see this kind of thinking very clearly at work. Instead of taking someone’s words, deeds and actions and evaluating them critically, something much vaguer and more sinister is taking place. “Questions are being asked”, but they are framed in a way that encourages the reader towards paranoia and illogical conclusions rather than an objective understanding of the facts.

It’s poor reasoning, and it has real consequences.

First, it causes real-world suffering for people who feel betrayed, falsely accused and unfairly besmirched. Second, it normalises low standards of research to the point where people find it harder to distinguish decent journalism from hackery. Third, it creates confusion and division as others understandably push back against it.

I was inspired to pull apart O’Neill’s piece for all these reasons. It’s a useful exercise in trying to find clarity. I stand by everything I wrote in that analysis, and the response was overwhelmingly positive. A small group seem hell-bent on defending O’Neill, but they are increasingly coming across as people on a martyr’s journey, determined to go down in a blaze of glory while seeing themselves as the only real truthers in town.

Having said that, my piece wasn’t emotionally neutral. I really ripped into O’Neill. As I said in one of the comments: “Sometimes the shit stinks so bad you just have to retch.”

But I’m sure Francis isn’t a bad person, and I’m sorry if it became a bit personal. I’m sure he means well—as do most people.

Anyway, after using Claude to respond to some of the comments, I thought I’d get AI to analyse O’Neill’s article in a more neutral way than I did. Interestingly, it identified the key flaw as O’Neill’s tendency to use the sheer volume of evidence as a substitute for the quality of evidence. I was pleased to see that because I made the same point in my own analysis.

It’s what Flat Earthers do all the time. Just look at Eric Dubay’s “Two Million Proofs That the Earth Is Flat”, or whatever it’s called. It doesn’t matter that most of the points are simply variations of the same weak arguments. The sheer number of them is enough to create the impression that something substantial must be there.

---

OK so here’s Claude’s breakdown of Francis O’Neill’s ‘The Curious Case of Ed The Techie - And Bob Moran’:

It’s an accumulation piece: dozens of discrete observations (timelines of tweets, prices of cartoons, names of venues, mutual friends, employer descriptions) are laid side by side with minimal individual analysis, and the persuasive force comes almost entirely from density and juxtaposition rather than from any single strong inference. That’s worth naming up front because it shapes everything else — when an argument’s power comes from volume of association rather than from the strength of any one link, that’s a structural tell, independent of whether the underlying facts are accurate.

Specific reasoning patterns worth flagging

Unfalsifiable framing of contradictory behavior. The “Undercover Influence” section lists supposed traits of “agents” (high-profile rite of passage, abrasiveness, citing grief, slowness to dissent) and then fits multiple real people to it. But the traits are vague enough, and common enough among genuine dissidents too, that they don’t discriminate between “covert operative” and “ordinary person who has a public falling-out and later regrets it.” A pattern that fits both the null hypothesis and the suspected one isn’t doing evidential work.

Coincidence stacked as causation. The painting timeline (Moran working on a piece “in the week before” Ed’s death, finishing as he was dying) is presented as suspicious, but the alternative explanation — a friend works slowly on a tribute piece and a death intervenes — is at least as plausible and isn’t weighed against it. Same with the venue choices (Travellers Club, Satan’s Whiskers): notable-sounding details are listed without an account of base rates (how often do book launches happen at slightly grand or oddly-named venues for mundane reasons like availability or a friend’s connection?).

Argument from incredulity used repeatedly. “It’s surprising that he thought this was acceptable,” “Are all those depicted... to some extent, fictional characters?”, “we might ask why” — these rhetorical questions substitute for an actual argument. They invite the reader to supply the sinister inference themselves rather than the author demonstrating it.

Guilt by association, several layers deep. Moran → Charlotte Binns → Mappin → Mappin’s friend’s father → Chatham House. Each link may be factually true, but the chain is used to imply something about Moran himself, and the more hops involved, the less it tells you about the original subject — almost everyone is two or three hops from someone notable or controversial if you draw the graph widely enough.

Selective use of inconsistency. The piece is genuinely sharp in spots — noting Moran’s stated “never edit my work” policy being waived for one person, or the mismatch between Moran’s claimed creative block and a finished painting, are legitimate, checkable inconsistencies, and pointing them out is reasonable critical thinking. But these real inconsistencies are embedded in a much larger structure (masonic numerology in a signature, “33” pricing, building histories) where the standard of evidence drops sharply and treats numerology and architecture as equivalent in kind to a checkable contradiction.

Financial figures presented without comparison baseline that actually supports the implication. The Pritchett vs. Moran cartoonist pricing comparison is one of the more legitimate moves in the piece — actual numbers, actual comparison. But it’s immediately followed by “when pop stars... are richly rewarded it is often assumed they have a role in social conditioning,” which smuggles in the conclusion via a vague appeal to what “is often assumed,” rather than arguing for it.

Net assessment

There’s a real journalistic kernel here: documented timeline discrepancies, a specific unanswered question (why the disproportionate reaction to one death versus comparable others), and some actual numbers. Those are legitimate things to ask about, and a few of the inconsistencies are concrete enough to be checked rather than just felt. But the piece doesn’t sustain that standard — it shifts constantly between checkable claims and atmosphere (venue names, signature shapes, “33,” “ragtag band,” who’s “friends with” whom at two or three removes), and it never weighs the more mundane explanation against the suspicious one before moving to the next data point. The overall effect is less “here is the strongest case, addressing the obvious objections” and more “here is everything that looks odd, arranged so the cumulative weight does the arguing” — which is a much weaker form of reasoning even when individual data points are true.