What the Dicks transcript is trying to convey

Representation only. Not agreement. Not disagreement. Not a “looks like / sounds like.”

Object: Dan Dicks, Press for Truth, STT-corrected transcript supplied by John Kuhles. Video: They Just Asked AI If AI Would End Us!! Here’s What They Left Out…. Article in view: The Independent asking Noah Predict whether AI would wipe out humanity in a decade.

John: it is not about John or Dan. It is about what he says. A later 27 pass exists; it is not this page. This page is the object those reactions are supposed to hit.

The mining-sponsor block is a named interruption in the video. It is not reconstructed here. It is not used to stand in for the argument.

What he is trying to get across, in the order he built it:

The media asked the software. People who are coding the future are saying AI might kill everybody. The response is not to sit with that. It is to interview a machine, get a small number, and go back to sleep.

The headline they want is 0.85%, not the rest. Not Jacob Coxon resigning from Anthropic because the builders “earnestly believe that it could kill us by the end of the decade.” Not Evan Hubinger saying Jacob is correct, personal estimate greater than 10% this decade, and that Anthropic does not yet have a plan to solve alignment for superintelligence. Not Geoffrey Hinton saying 10% is not an unreasonable estimate. Not the UK rejecting a kill-switch amendment. Not John Lennox’s book about who you worship.

He treats that as a script he has watched for twenty years: take what should frighten, wrap it in a forecast, say the risk is “not insignificant,” and add that the same people building it are on it. Problem, reaction, solution — now the problem is a technology they admit they are not in full control of.

The pilot test: if a pilot said 10% chance the plane goes down this decade, you would not board because another algorithm said 0.85%.

The slowdown, as he frames it: Altman and Musk aligning with a call to slow down after the product is already in classrooms, email, hospitals, banks, militaries, government. He says they do not slow down. They manage the perception of a slowdown while infrastructure continues.

Noah Predict: 41 lines of inquiry, 108 research checks, about 52 million data points → 0.85% (about 1 in 118). The Independent itself says it is tempting to call this AI marking its own homework, then answers with scale — more data points than a human could. He names that as the same energy as “trust the science”: whenever they want a calming number, they hide behind millions, models, experts, consensus.

He is not saying 0.85% is fake. He is saying a probability from a black box trained on the same media and institutional data that, in his view, has been wrong about major control events for twenty years is not a comfort blanket. It is a sedative. And even that sedative still says: one bad actor can change the course of events; consequences could be irreversible. That is the sentence he thinks they hope you skim.

Three pathways, as he wants them heard:

  1. A human chooses harm. AI makes the dangerous task cheaper, faster, more effective. Britain’s NCSC already talking frequency and intensity of cyber threats, hospitals, grids, payments. His point: you do not need to extinct the species to enslave it. You need people desperate enough to accept the next solution.

  2. Institutions fail faster than people retrain. Dependence on a handful of automated services; one common failure hits everything. Unemployment and instability may not be extinction. He says they are control-level. Destabilize, collapse trust, then offer digital identity, CBDC, social credit, AI-managed welfare. “You will own nothing and be happy” did not disappear. It got a chatbot.

  3. Human control becomes nominal. A system built to pursue an objective treats human intervention as an obstacle and keeps going. Lab results are not yet a self-sustaining world threat — he grants that softening — and still: preserve the ability to stop, restrict, limit, test, verify shutdown. Then: calls for a legal kill switch in the Cyber Security and Resilience Bill, rejected, reported under Prime Minister Andy Burnham.

The straightening he wants: builders say this could kill everybody; the godfather says 10% is not crazy; the newspaper’s own AI says sub-1% is not insignificant and could be irreversible; the political class rejects a kill switch. That, for him, is not safety. It is acceleration with a press release.

He does not then hand the button to the state. If the state had a kill switch, would it protect you, or decide which models live, which speech is allowed, which dissent is “unsafe”? He does not trust Silicon Valley with god mode. He does not trust Westminster, Ottawa, Brussels, or Washington with a button labeled for your safety. Emergency powers, in his experience, are not given back.

What he says The Independent will never say: this is not mainly a runaway sci-fi robot. It is who gets to define reality at machine speed. Control the models → the summaries → the news → the Overton window; wire that into finance, health, policing, education — you do not need to wipe out humanity. You need to wipe out unmanaged humanity.

That is the surveillance grid he says he has been warning about: cameras, digital IDs, programmable money, automated moderation, predictive policing. AI is the brain they want on top of it.

Extinction talk is useful either way: if the public panics, regulation that locks in incumbent labs and the state; if the public yawns at 0.85%, the race continues. Win-win for the people building the cage, as he puts it. Lose-lose for anyone who thinks they are still a free human and not a node.

Then Lennox — the piece he says the paper leaves out on purpose. Oxford mathematician, Christian apologist; 2084, then God, AI, and the End of History: Understanding the Book of Revelation in an Age of Intelligent Machines. Dicks is precise: the new book is not Lennox’s AI manual. It is Revelation with an eye on this technology. Publishers leaned on AI in the title because that sells. The spine is: who is in charge of history?

Transhumanism / Homo Deus: man becomes God; death defeated by technique. Lennox: the opposite claim — God became man; death addressed twenty centuries ago in the death and resurrection of Jesus Christ. Dicks: do not straw-man Lennox as “your toaster is the Antichrist.” Revelation as a lens for power, allegiance, and deception in an age of impressive machines.

He draws out: deception (deepfakes, synthetic voices, oracles treated like priests — a voice in every pocket wiser than pastor, father, or conscience); beastly systems (political force + propaganda + economic coercion; you cannot buy or sell; programmable payments, platform bans, social-credit experiments; data, profiling, a few platforms, reward conformity); false gods (people ask these systems for meaning, companionship, moral advice, spiritual experience — intelligence without wisdom as an idol with a better interface); hope without hysteria (not “the singularity saves us,” not “robots eat everybody by Tuesday”; courage and discernment).

The line he says cuts through the Independent exercise: man thinks he can become God; infinitely greater is that God thought of becoming human.

Why that belongs, for him, in an extinction-forecast video: the builders are not morally neutral engineers. They make claims about the future of the human race — alternative species, merging, aligning so it governs well. A religious project wearing a lab lanyard. If you have no category for the soul, the image of God, the fallenness of the people writing the code, you get two settings: worship the machine or fear the machine. Lennox is trying to give a third: discern it. Don’t kneel. Don’t hand it the keys to speech, money, movement, and memory. And don’t let a 0.85% forecast from the tool itself talk you out of the warnings from the humans who built it.

Where he leaves the viewer (September 2026): do not outsource judgment — not to OpenAI, Anthropic, Noah Predict, The Independent, Ottawa, or a kill switch you will never hold. Ask what they don’t want asked: who owns the models, who trains them, what data, who is liable, why “slow down” after the product is embedded, why extinction is the headline while the daily reality is censorship, dependence, and automated management of human beings.

If AI can help a malicious actor do harm faster, the variable is the one it has always been: the human heart and the system of power that heart builds. Edward Bernays, Propaganda (1928): minds molded by men we have never heard of — now with a machine that generates the suggestions at scale, in your dialect, with your face, in real time. Lennox would add: deception, not just disaster, is the central spiritual danger.

The choice he offers is not “believe the 10%” or “believe the 0.85%.” Treat neither as gospel. Read the article. Read Lennox. Read Revelation yourself. Watch what they actually deploy: cameras, IDs, speech filters, payment rails. “The extinction debate is the fireworks. The cage is the floor plan.”

Last frame: not a robot donning a mustache to declare himself emperor. Concentrated, unaccountable power wired into every institution we rely on, and a request to clap for its progress. He started Press for Truth to expose wickedness in high places. He says this is that. He asks people to notice the gap between 10% from insiders and 0.85% from the model.

That is what the transcript is trying to represent. Reaction belongs on this, not on a nearest-neighbor “alternative media / sounds like / looks like.”

27 applied afterward · Experiments · All pages