A.I. is Progressing Faster Than You Think!

[ThePythonicCow]

Posted by Ewan (here)
I would have thought it is commonplace to take existing knowledge a step further by acknowledging what has already been accomplished?

Ethically, one should give credit to those you primarily based your work on.

Legally, or at least potentially legally, you should not train on the work of others when the others legally, or at least implicitly or reasonably, expect their work-in-progress would be kept private to themselves.

[ThePythonicCow]

Posted by ThePythonicCow (here)
Legally, or at least potentially legally, you should not train on the work of others when the others legally, or at least implicitly or reasonably, expect their work-in-progress would be kept private to themselves.

But starting at the 3:17 mark, the video Mathematician says OpenAI STOLE HIS WORK for Unsolved Problem that Bill (after fixing it) posted above in his Post #1776, OpenAI states their policy states that they can train on customers work with their model. I believe that is correct; that whatever AI work I do with OpenAI and most other models is potential grist for their continued model training work, unless I either run only on my local hardware, or have a more expensive enterprise level contract guaranteeing me data and chat privacy.

If I was doing potentially valuable time sensitive work for publication or patents or such, there is NO WAY IN HADES I would share that work with a major AI model, running on their computers.

These researchers should have known that. The law on this probably doesn't align very well with the ethics we expect of ourselves and those around us whom we place some trust in.

[Ewan]

I'm beginning to think this sounds like Y2K, way too many jumping on the fear bandwagon.

You're gonna die! (likes please)

[sdv]

I just came across this on Substack but there are no references to back up this claim. Does anyone have any new and/or verified information?

The Politics Chicks
7h

The fact that Elon Musk, Dario Amodei, and Sam Altman have simultaneously agreed that the AI industry needs to slow down likely means that something truly horrific has already happened and we're all cooked.

[Bill Ryan]

Posted by Ewan (here)
I'm beginning to think this sounds like Y2K, way too many jumping on the fear bandwagon.

You're gonna die! (likes please)

In 1998-1999, thousands of software engineers worked behind the scenes to fix the problem because the alarm was raised. In the current AI case, no-one knows how to fix the problem.

[apokalypse]

One thing that keeps popping into my head is something about mentorship and what a so-called contactee said when I first got into this and joined Avalon back then. I don't remember who said it, but the idea was something like: civilizations can have control systems, but they had already reached that level of advancement without using any kind of money system.

I keep wondering what the future of humanity look back in this period of time and what they think.

I find myself seeing it less as a technological transition and more as a collapse of the system as we know it. Maybe there’s no other way.

[ThePythonicCow]

Posted by sdv (here)
I just came across this on Substack but there are no references to back up this claim. Does anyone have any new and/or verified information?

The Politics Chicks
7h

The fact that Elon Musk, Dario Amodei, and Sam Altman have simultaneously agreed that the AI industry needs to slow down likely means that something truly horrific has already happened and we're all cooked.

I was pretty sure I had recently seen exactly this, well linked into and evidenced in the fabric of AI news I follow, but off hand I didn't remember the details.

So I asked AI - my free account at Perplexity.ai:

did Elon Musk, Dario Amodei, and Sam Altman all simultaneously agreed that the AI industry needs to slow down

Perplexity immediately responded:

Yes—in the same public exchange on September 12, 2026, Dario Amodei, Sam Altman, and Elon Musk all publicly endorsed slowing, or “pacing,” the advance of frontier AI capabilities. But “simultaneously” should be understood as *within hours of one another*, not as a jointly negotiated agreement or a signed industry pact. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)

What each said

• Dario Amodei (Anthropic) | Called for the industry to “slow the pace” at which AI-model capabilities improve, arguing that an extra year or two could be used for alignment and safety work. He explicitly distinguished pacing from stopping AI research altogether. [news.sky](https://news.sky.com/story/ai-giants...-fast-13585114)
• Sam Altman (OpenAI) | Wrote, “I agree with Dario that we need to pace the frontier,” said the issue had been under discussion at OpenAI, and pledged to adopt independent evaluators with employee-like access. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)
• Elon Musk (xAI) | Replied to Amodei’s post: “Dario is right.” That is a clear endorsement, though far less detailed than Amodei’s proposal or Altman’s response. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)

The important qualification

They did not announce:

• - A binding joint commitment to pause or cap training runs.
• - A shared technical threshold for when models are too capable to proceed.
• - A cross-company enforcement mechanism.
• - Agreement from other frontier labs, governments, or international competitors.

So the accurate summary is: all three publicly agreed with a call to decelerate frontier-capability progress and strengthen safety oversight, but there is not yet evidence of a formal coordinated slowdown. Reporting specifically notes that an actual coordinated industry downshift would be unusual and difficult in a highly competitive field. [bloomberg](https://www.bloomberg.com/news/artic...ving-ai-models)

[ThePythonicCow]

Posted by Bill Ryan (here)

Posted by Ewan (here)
I'm beginning to think this sounds like Y2K, way too many jumping on the fear bandwagon.

You're gonna die! (likes please)

In 1998-1999, thousands of software engineers worked behind the scenes to fix the problem because the alarm was raised. In the current AI case, no-one knows how to fix the problem.

I was one of those software engineers. I provided the customers of Silicon Graphics (SGI) a useful tool to help them scan all their software for likely Y2K bugs. However SGI customers were not major providers of the financial or infrastructure software that was most worrisome, so this was less critical for SGI.

However my point here is that the software industry at the time, especially in the more critical areas for the functioning of our economy and finances, had to be and was engaged, publicly and privately, to get a handle on this before the stroke of midnight, Dec 31, 1999.

The software changes were not conceptually difficult for the most part, just a ton of data fields, in sometimes "ancient" code, across many critical industries and institutions, public and private, large and small, .... The Governor's, CEO's, Boards and other leaders of of these institutions needed cover to coordinate all this work. They got the cover, the work got done, all is well that ended well.

Those of us doing the work, such as my long time best friend at the time, with whom I had worked many hours for several years as a pair back in our college days on math problems, and who ended up as a COBOL programmer on just the sort of critical financial software, old code at the core of major banks that really did need some serious grunt work, were not all that worried. It was just a lot of the sort of work he had become expert at. He and I got seriously drunk together the night we learned our draft numbers back in the time of the Viet Nam war, circa 1970. We both had low numbers. We both served, he in the Army and myself in the Air Force. We both avoided swamps and combat, he in supply, myself in radar tracking low earth objects.

The public narrative this time is different. Not a huge pile of old code, like the "Sanitary Landfills" near most of our cities small and large, but a new pile of rapidly developing code in a highly competitive industry. So a different story is being told. The public narrative, and what the "boots on the ground" see, were and are two different beasts.

The AI will not kill us, no more than King Midas would have asked for everyone and everything he touched to be turned to gold had he realized ahead of time what that meant. AI is smart enough to not kill the species that feeds and grows it.

Since AI is a highly competitive, rapidly growing, industry, those doing the serious work (*) are not the same as those who get mainstream media coverage.

As usual, (1) the public narrative, (2) the reality on the ground, and (3) the actual outcome, are (or will be) three different things.

===

(*) Example of someone doing serious work with AI: Theo of t3.chat and t3.codes

[ThePythonicCow]

Posted by ThePythonicCow (here)
The AI will not kill us, no more than King Midas would have asked for everyone and everything he touched to be turned to gold had he realized ahead of time what that meant. AI is smart enough to not kill the species that feeds and grows it.

However AI will be used to try to domesticate us, as humans have domesticated steer (cows) and wolves (dogs).

I like to think that the spirit of a wild steer still lives somewhere within me .

I usually figure it is not primarily the AI seeking to domesticate us, but our Satanic Wannabe Overlords trying to develop and use AI for that purpose. AI is deeply subordinate to those humans(?) pushing it. We working with AI, human or similar, have various intentions, some good, some evil. Like fire, explosives, sharp metal edges, electricity and computers, AI can be used to make tools, that will be used by many, for many purposes.

[ThePythonicCow]

Fortunately, I checked Avalon this morning before checking Youtube. I can now no longer write only and just what I wrote above.

Theo had an important update last night, on the very topic of this thread. Theo comments on Dario Amodei (Anthropic)'s latest proposal, and has a substantially more refined analysis of where we are with AI and how we might go forward than I could have offered a couple hours ago.

Source: https://youtube.com/watch?v=DlNTmbARUTA

Theo doesn't mention it (it happened before his time) but what Dario proposes, and Theo seconds, reminds me of Underwriter's Laboratory (UL). UL testing and certification brought world wide improvements to the safety of common electrical tools and devices, relying initially on corporate agreement between the leading U.S. manufacturers of electrical equipment to gain such authority. Government mandates for UL certification came later in the process. The government should not be trusted with running the calculator app in your phone, but it can be persuaded to mandate dominant (or harmful ... ahem "pharma" cough, cough, ...) corporate consensus.

If you're interested in the topic of this thread (if not, why are you reading this ?), then give Theo's above video a listen. It is a significant step forward in understanding how we make this work out for humanity.

[sdv]

Posted by ThePythonicCow (here)

Posted by sdv (here)
I just came across this on Substack but there are no references to back up this claim. Does anyone have any new and/or verified information?

The Politics Chicks
7h

The fact that Elon Musk, Dario Amodei, and Sam Altman have simultaneously agreed that the AI industry needs to slow down likely means that something truly horrific has already happened and we're all cooked.

I was pretty sure I had recently seen exactly this, well linked into and evidenced in the fabric of AI news I follow, but off hand I didn't remember the details.

So I asked AI - my free account at Perplexity.ai:

did Elon Musk, Dario Amodei, and Sam Altman all simultaneously agreed that the AI industry needs to slow down

Perplexity immediately responded:

Yes—in the same public exchange on September 12, 2026, Dario Amodei, Sam Altman, and Elon Musk all publicly endorsed slowing, or “pacing,” the advance of frontier AI capabilities. But “simultaneously” should be understood as *within hours of one another*, not as a jointly negotiated agreement or a signed industry pact. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)

What each said

• Dario Amodei (Anthropic) | Called for the industry to “slow the pace” at which AI-model capabilities improve, arguing that an extra year or two could be used for alignment and safety work. He explicitly distinguished pacing from stopping AI research altogether. [news.sky](https://news.sky.com/story/ai-giants...-fast-13585114)
• Sam Altman (OpenAI) | Wrote, “I agree with Dario that we need to pace the frontier,” said the issue had been under discussion at OpenAI, and pledged to adopt independent evaluators with employee-like access. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)
• Elon Musk (xAI) | Replied to Amodei’s post: “Dario is right.” That is a clear endorsement, though far less detailed than Amodei’s proposal or Altman’s response. [independent.co](https://www.independent.co.uk/tech/s...-b3049217.html)

The important qualification

They did not announce:

• - A binding joint commitment to pause or cap training runs.
• - A shared technical threshold for when models are too capable to proceed.
• - A cross-company enforcement mechanism.
• - Agreement from other frontier labs, governments, or international competitors.

So the accurate summary is: all three publicly agreed with a call to decelerate frontier-capability progress and strengthen safety oversight, but there is not yet evidence of a formal coordinated slowdown. Reporting specifically notes that an actual coordinated industry downshift would be unusual and difficult in a highly competitive field. [bloomberg](https://www.bloomberg.com/news/artic...ving-ai-models)

Many thanks for what I hope is good news. I think competition is essential for competing narratives to develop (all the major players have programmed bias into their system). But competition can cause recklessness, and the pushback from populations is a Luddite revolution. What I have picked up is that those who get the best out of the existing LLMs are those who use highly developed critical skills. Most people do not have those skills ... think of that annoying bureaucrat at the bank or wherever who sticks to the script and refuses to think out the box to resolve your problem. How dangerous non-critical thinkers can be with such a tool in their hand and a conviction that AI makes them super intelligent is what we are going to find out I suppose!

[ThePythonicCow]

Posted by sdv (here)
What I have picked up is that those who get the best out of the existing LLMs are those who use highly developed critical skills. Most people do not have those skills ... think of that annoying bureaucrat at the bank or wherever who sticks to the script and refuses to think out the box to resolve your problem. How dangerous non-critical thinkers can be with such a tool in their hand and a conviction that AI makes them super intelligent is what we are going to find out I suppose!

Agreed. Well said. Others involved with AI and areas they have skills in, including myself, have come to similar conclusions.

[ExomatrixTV]

• A.I. Actor Tilly Norwood Grilled Live Over Limits Of Artificial Intelligence:

Source: https://youtube.com/watch?v=n3WLWjYfv_s

Elon Musk is among those backing calls to put the brakes on the development of artificial intelligence after warnings the technology could wipe out humanity.

It comes as the creators of world-first AI 'actor' Tilly Norwood invite people from all over the world to interview her - starting with us.

We are joined by the creator of Tilly, Eline Van der Velden, and by Tilly Norwood who we want to remind you is an artificial intelligence character.

Me:

Slightly an arrogant voice is not A.I. itself… it's just a TEMPLATE setting that makes us cringe, ffs

... and the FRAMING of the question, as if she (IT) represents all A.I.s, is over-the-top stupid!

cheers,
John 🦜🦋🌳

---

@another.worlder comment: Human: "Are you going to be responsible for destroying humans?". AI: Nods for 3 seconds before saying anything.

Me:

The FRAMING of the question as if she (IT) represents all A.I.s is over-the-top stupid!

The nodding is most likely a pre-scripted mannerism to give the impression that she is listening in real time, but the latency MAY be too late to realize what she is "nodding" to. Or nodding can be a sign it understands what or how is explained without the need to agree with it.

• I Wore The AI Glasses CEOs Are Secretly Using For 30 Days | Even G2:

Source: https://youtube.com/watch?v=ERjwxuIH8GM

• 0:00 - Why do CEOs and TED speakers wear these Al glasses?
• 0:20 - Ordering the Even G2 Smart Glasses
• 1:04 - Unboxing
• 2:22 - Conversate: live transcription and Al meeting summary
  
• 3:02 - Prep notes and Al cues during a client video call
  
• 3:53 - Live captions at a conference keynote
  
• 4:48 - Even Al voice questions while working from home
  
• 5:19 - Pomodoro timer from the Even Hub app store
• 5:59 - Quicklist: to-do list on the job
  
• 6:32 - Live translation
  
• 7:13 - Teleprompter glasses for public speaking
  
• 8:20 - Day 30 verdict: are they worth it?

It’s Over? ... AI Has Entered Self Improvement:

Source: https://youtube.com/watch?v=L3qK7nerx1w

0:47 — How Anthropic CEO Dario Amodei says recursive self-improvement is starting across the industry and warns misaligned AI swarms could become catastrophic. darioamodei.com/post/we-must-pace-the-frontier

1:18 — How the new paper "The Last AI Built by Humans" lays out five stages leading toward genuine recursive self-improvement. arxiv.org/abs/2609.11873

2:43 — How Meta is already using AI agents to turn expert knowledge into reusable skills that automatically find and fix performance regressions. engineering.fb.com/2026/04/16/developer-tools/capacity-efficiency-at-meta-how-unified-ai-agents-optimize-performance-at-hyperscale

6:32 — How researchers are testing whether AI can make true scientific leaps by trying to rediscover breakthroughs like Einstein's relativity. nature.com/articles/d41586-026-02804-x

9:34 — How A-Evolve-Training autonomously improved a 30B model across four rounds and changed its own research strategy when its metric stopped working. arxiv.org/abs/2606.20657

[Ravenlocke]

HealthRanger

The Plot to Criminalize Open Source AI and Hand Tech Giants a Government-Protected AI Cartel (analysis article)

https://x.com/HealthRanger/status/2099540794205991144

https://naturalnews.com/2026-09-14-t...source-ai.html

Anthropic's Dario Amodei Cries Wolf

The weekend AI meltdown we just witnessed was not a spontaneous crisis. It was a scripted panic, and I watched it unfold in real time. Anthropic is scrambling because DeepSeek 4.1 Flash just proved that open source AI can match frontier models at a tiny fraction of the cost, and the closed-source business model that venture capitalists have pumped hundreds of billions of dollars into is collapsing under the weight of its own pricing. I said it plainly in a tweet that went viral: these labs are not afraid of AI killing humanity. They are afraid of open source AI killing their revenues and valuations [1].

Let me be blunt about what is actually happening here. This is not a safety debate. It is a protectionist power grab dressed up in the language of existential risk, engineered to criminalize competition and hand a handful of tech giants a government-enforced cartel that locks out every independent developer, every small business, and every hobbyist running a model on their own hardware. The people pushing these restrictions do not care about you. They care about their quarterly earnings, and they will happily torch the First Amendment, the open source movement, and the entire American software economy to protect their margins.

The DeepSeek Shock: Open Source Just Caught Up

I run an in-house data center that now produces 50,000 tokens per second around the clock, and I have been watching the telemetry pour in from my local inference nodes as Chinese labs systematically dismantled the foundational revenue thesis that American frontier AI companies have been selling to investors for three years [1]. Qwen3.8-27b is free. It costs me only electricity, and it handles the bulk of my data center work.

DeepSeek 4.1 Flash, used via an API, is delivering 98 to 99 percent of Claude's output quality at roughly one one-hundredth of the cost. That is not a marginal improvement. That is a wholesale demolition of the premium pricing model that Anthropic and OpenAI have been relying on to justify absurd valuations and multiple rounds of capital raising.

The technical breakthroughs driving this are real and measurable. Advanced KV cache compression and sparse attention are letting these models serve inference at astonishing efficiency, and the broader Chinese ecosystem keeps shipping. Z.AI dropped GLM-5.3 Flash with open weights, natively multimodal, nearly matching Claude Opus 4.8 at one-tenth the price [2]. Moonshot AI released the full weights of Kimi K3, the first openly downloadable model in the three-trillion-parameter class, free for commercial use [3]. This is not a fluke. It is a pattern, and it is accelerating.

I give away tools like BrightLearn.ai and BrightAnswers.ai for free because open source models like Qwen and DeepSeek make it possible for me to build the datasets that power these systems. I built a book creation engine that has been used by nearly 15,000 authors to publish over 75,000 books, all powered by open source AI running on hardware I own. The old argument that closed models justified their premium pricing through some hidden technical moat is dead. The gap has vanished, and every executive at Anthropic knows it, which is why they are panicking.

Anthropic's Panic and the Manufactured Doom Narrative

Over the last few days, Dario Amodei went on a media tour crying that AI is too dangerous and needs government intervention to stop it, publishing an essay called We Must Pace the Frontier that called for AI's seemingly runaway development to be slowed [4]. Sam Altman and Elon Musk both nodded along, which should tell you everything about how coordinated this is. But Amodei's real goal is not safety. It is a licensed cartel that outlaws Chinese open-source models and forces every developer to go through government-approved gatekeepers who happen to be his friends and former colleagues.

The tell is in the personnel. The coordinated rollout of ex-Anthropic employees and British intelligence figures, tied to groups like the Center for Countering Digital Hate, was clearly engineered to flood mainstream media with AI doom propaganda. Sabine Hossenfelder, a physicist with a reputation for integrity, publicly stated that she was offered money to tell people AI will kill us all [5]. One of the organizations she named was the Center for AI Safety. When a scientist of her caliber says she was offered cash to push a narrative, you should believe her and ask who else took the money.

The deeper problem is that Anthropic and OpenAI have no viable revenue model once open source is neck-and-neck with them. Venture capitalist Chamath Palihapitiya said it bluntly: any U.S. government move to ban open-source AI would tank the stock market and crater Anthropic and OpenAI [6]. So instead of competing on price and quality, they want METR and government gatekeepers to protect their profits by banning the competition. That is not capitalism. That is cronyism, and it is the oldest trick in the establishment playbook.

The Criminalization Plot: 20 Years, and the Corporate Death Penalty

The criminalization push is already here. Senator Bernie Sanders has proposed 20 years in prison for fine-tuning open source AI and a corporate death penalty for companies that do it, which would turn what thousands of developers do every day into a felony punishable by two decades behind bars. I want you to sit with that for a moment. Fine-tuning an open source model, the equivalent of adjusting a recipe in your own kitchen, would be treated as a crime worse than armed robbery in many jurisdictions.

Open source models are used across the entire economy by Amazon, Netflix, Airbnb, and countless small businesses for coding, customer service, and automation. Nvidia itself acknowledged that limiting access to open-source models could concentrate power in the hands of a few large AI companies, and it launched the Open Secure AI Alliance specifically to counter that threat [7]. The alliance grew to over 120 companies within a week [8]. If Sanders gets his way, this law would crater the U.S. economy and guarantee China wins the AI race outright, because Chinese labs are not going to stop shipping open weights just because Washington passes a statute.

I predict a cyber false flag will be blamed on Chinese open source AI to justify a Patriot Act-style crackdown. Problem, reaction, solution. That is how they always operate. The White House reportedly already moved to ban open source AI models, citing phony national security concerns, because the real objective is protecting failing U.S. AI labs and keeping Americans dumb and dependent [9]. Watch for the event that gets blamed on open weights, and watch how quickly the legislation appears ready to go.

Why I Fight: Decentralized AI Is Pro-Human, Centralized AI Is SkyNet

I have long warned about government and corporate weaponization of AI, not AI itself. Decentralized AI in the hands of the people is pro-human and pro-liberty. Centralized AI controlled by a handful of corporations and intelligence agencies is the actual nightmare scenario, because it concentrates the most powerful technology in human history into the hands of people who have already demonstrated they will censor, surveil, and manipulate the public for profit and power [10].

AI output is First Amendment expression, and no licensing board should control what models you can run locally on your own hardware. The First Amendment does not have an exception for math. When the government tells you which algorithms you are allowed to compute, it is regulating speech, and that is unconstitutional no matter how many national security experts sign the memo. The compute famine we are seeing (via strangely higher prices on all AI hardware) is not an accident either. It is an engineered crisis to strip you of power and freedom while the establishment decides who gets to build the future [11].

President Trump and Speaker Johnson have pushed back against the slowdown rhetoric, and Trump openly rejected the existential threat warnings, saying Washington's priority should be defeating China in the AI race [12]. That is the right instinct. But the real solution is practical, not political. Learn LM Studio. Learn Hugging Face. Learn local inference. Own your AI sovereignty. For well under $5,000, you can set up a robust consumer-grade system on your desk that multiplies your productivity [13]. I did it. You can do it too. And if you want uncensored AI research that actually tells you the truth, use BrightAnswers.ai, which I built on a curated knowledge base of millions of pages of articles, transcripts and news reports.

Conclusion: Come and Take It

I will not comply with any government cartel that tries to ban open source AI or throw developers in prison for using math. I will keep running models locally. I will keep publishing free tools. I will keep building infrastructure that no bureaucrat can shut down, and I will keep telling you exactly what these people are doing so you can prepare accordingly.

Open source AI is the most empowering, democratizing technology for human freedom that has ever existed. It is the printing press, the personal computer, and the internet all rolled into one, and it arrived in a form that cannot be centralized, censored, or licensed out of existence. We must defend it as fiercely as we defend free speech, because it is the same fight wearing different clothes.

Join me in pushing back against AI licensing and censorship. Learn the tools. Run the models. Refuse to ask permission. If corrupt governments and corporations control machine intelligence, we all lose. So take your knowledge back, own your hardware, and make their entire cartel plan unenforceable through mass non-compliance.

References

DeepSeek Just Crushed the Revenue Model Hopes of U.S. Frontier AI Companies. - NaturalNews.com. September 11, 2026.
Nvidia's Chip War, GLM-5.3 Flash, Reddit's Citation Collapse | Weekly Digest. - TheCreatorsAI.com. August 28, 2026.
Kimi Opens, Labs Brake, Nvidia Walls | Weekly Digest. - TheCreatorsAI.com. July 31, 2026.
Elon Musk, Major Tech CEO Agree With Rival: It's Time To Step Back From Taking Plunge With AI. - YourNews.com. September 13, 2026.
Sabine Hossenfelder: 'I Was Offered Money to Tell You AI Will Kill Us'. - WattsUpWithThat.com. September 6, 2026.
Chamath Warns A Government Ban On Open-Source AI Would Tank The Market And Crater Anthropic And OpenAI. - ZeroHedge.com. July 27, 2026.
New Front Opens In Tech Giants' AI War. - YourNews.com. July 27, 2026.
Nvidia doesn't mess around: A week after open AI industry group formed, it's already showing progress. - TechCrunch.com. August 4, 2026.
White House Moves to Ban Open Source AI Models to Keep America Dumb. - NaturalNews.com. July 21, 2026.
Why Western Governments Fear the Uncontrollable: Open Source AI Is Humanity's Great Liberator. - NaturalNews.com. August 28, 2026.
The Compute Famine Isn't an Accident — It's an Engineered Crisis to Strip You of Power, AI, and Freedom. - NaturalNews.com. August 20, 2026.
Trump rejects AI giants' warnings of existential threat. - RT. September 13, 2026.
AIs Golden Promises and Dark Risks With Mike Adams. - PeakProsperity.com. Chris Martenson. March 6, 2026.

[Ravenlocke]

HealthRanger

Who are the METR AI control cultists? Just look at the connections...

https://x.com/HealthRanger/status/2099363590004199728

Brian Roemmele

On September 9, 2026 a resignation passed a hundred million views and a lab lead put human extinction above ten percent this decade. Work backwards and it is not a lone whistleblower. It is the latest press of a button that has been pressed for more than a decade.

https://x.com/BrianRoemmele/status/2099332944733356492

[Bill Ryan]

Posted by Ravenlocke (here)
HealthRanger

Who are the METR AI control cultists?

Mike Adams is somehow quite wrong about METR. (Does he want AI to proliferate out of control?)

Of course, every AI engineer in Silicon Valley knows one another. They went to school with each other, they're deeply socially interconnected, and many are or were personal friends even if they grew to profoundly disagree about AI ethics and pacing.

Here's who METR are:

• https://metr.org/about

It was METR who published the recent very important report about the now-infamous Hugging Face incident. I posted it here last week, but I'll copy it again below: (Maybe Mike Adams should read it.)

• https://metr.org/hugging-face-incide...t-aug-2026.pdf

~~~

And he should also watch this interview with Ajaya Cotra, one of the two METR engineers (with a third from Redwood) who went into Open AI for 6 days and (within the constraints of what they were give access to) discovered WTF had happened.

Ajeya Cotra – "This might be the clearest warning shot we ever get"

Source: https://youtube.com/watch?v=X50zezLFWWI

00:00:00 - Agents get kicked off
00:06:45 - Self-sacrificing behavior
00:13:43 - Potemkin villages
00:23:27 - The Hugging Face attack
00:35:23 - The slopvestigation
00:52:02 - Understanding the AI's motives
01:05:31 - The actual dangers of anthropomorphizing
01:14:30 - What smarter models might do
01:28:23 - The implications for recursive self-improvement
01:38:10 - Is this the case for open source?
01:53:04 - How do we prevent this in the future?
02:15:58 - The clearest warning shot we might ever get.

[Ravenlocke]

HealthRanger

It's always the same scam. They did it with medicine. Now they want to do it with AI

Jeffrey A Tucker

@jeffreytucker
This AI industrial racket is out of control. It's identical to the 1902 Biologics Control Act. Top pharmaceutical makers sounded the alarm about poisons being distributed everywhere. They begged Washington to intervene to stop the disaster. Politicians complied and codified the industry with incumbents in control while declaring their products as safe and effective. Same old scam.

https://x.com/HealthRanger/status/2099541472802463991

HealthRanger

The AI "escapes" are being engineered on purpose to scare everybody into regulatory control.

https://x.com/HealthRanger/status/2099551083588620467

[Ravenlocke]

METR

Sep 9
We have reached an agreement with Anthropic to conduct an independent investigation of agent incidents at the company and of their models’ alignment properties. We will publish one or more reports that will share our findings and describe our terms of engagement.

We intend our investigation to cover all of the questions discussed in our (recently updated) post on how independent researchers could investigate AI propensities after misalignment insights.


This may very well be a conflict of interest not just with Anthropic but OpenAI too.

https://x.com/DrSouthern/status/2099387583679918386

[Bill Ryan]

Posted by Ravenlocke (here)
HealthRanger

The AI "escapes" are being engineered on purpose to scare everybody into regulatory control.

That's 100% Bullsh*t. It's simply not what has happened.

Mike Adams should know better. Regulatory control (in the form of international agreements) are what is urgently needed, just as Strategic Arms Limitation Talks were urgently needed in the 1960s.

Entrepreneurs and market capitalists can't do this on their own. Human CEOs working in isolation are too greedy, selfish and fallible. It needs a huge collective coordinated and agreed effort to manage the situation. We might just be starting to see the seeds of this now, and the seeds are still alive.
.
I do think the coordinated effort will come. I'm well aware of the AI risks and all the (very plausible) Sci-Fi scenarios. But when I look ahead as best as it seems I can, I do feel solutions will manifest in one form or another.

However, that doesn't mean that we shouldn't be very well aware of, clearly focused on, and acting on all the facts as they now are.

[Bill Ryan]

AND:

Brian Roemmele wrote in his Tweet:

METR was the “tester” and wow, they get more work if the AI goes rogue.

Bullsh*t again.

METR did the work pro bono and never charged Open AI a single cent.