Justus Knight – RR News Update! September 15th, 2026

Welcome to Justus Knight News. This isn’t a recap show; it’s an ongoing education in political psychology. The Clinical Analysis: Tracking 8+ years of U.S. & global chaos. Subscribe to learn how power actually communicates. 🔒 Extended deep-dives & reports: www.restrictedrepublic.com

DESCRIPTION

The AI debate just became much bigger than whether artificial intelligence is dangerous.

Anthropic’s Dario Amodei is calling for the frontier to slow down. Bernie Sanders and Greg Casar are proposing sweeping federal restrictions on superintelligence. President Trump argues slowing American AI could hand China an enormous advantage.

But underneath that debate is another question:

Who gets to write the rules?

Today we trace the documented relationships connecting frontier AI companies, independent evaluator METR, investor and philanthropist Jaan Tallinn, ControlAI, and emerging AI legislation—and separate the confirmed connections from claims the available evidence does not establish.

The real fight may not be humans versus machines.

It may be humans fighting over who gets to control the machines.

COMMUNITY POST

YES OR NO:

Trump says AI doesn’t need more government “guardrails” — it needs a strong, smart President.

Do you believe ANY President can truly keep artificial intelligence under control?

References

1. Reuters — What Amodei, Altman and Musk Have Said About AI Risks, Stoking Doom Fears
   https://www.reuters.com/business/what-amodei-altman-musk-have-said-about-ai-risks-stoking-doom-fears-2026-09-14/
2. Reuters — Trump Says There Is a “Sick Conspiracy” Against AI and Data Centers
   https://www.reuters.com/world/trump-says-there-is-sick-conspiracy-against-ai-data-centers-2026-09-14/
3. Dario Amodei — We Must Pace the Frontier
   https://darioamodei.com/post/we-must-pace-the-frontier
4. Reuters — What Amodei, Altman and Musk Have Said About AI Risks, Stoking Doom Fears
   https://www.reuters.com/business/what-amodei-altman-musk-have-said-about-ai-risks-stoking-doom-fears-2026-09-14/
5. Reuters — Trump Says There Is a “Sick Conspiracy” Against AI and Data Centers
   https://www.reuters.com/world/trump-says-there-is-sick-conspiracy-against-ai-data-centers-2026-09-14/
6. METR — Frontier Risk Report
   https://metr.org/blog/2026-05-19-frontier-risk-report/
7. METR — Funding Update
   https://metr.org/blog/2026-08-14-funding-update/
8. Anthropic — Anthropic Raises $124 Million to Build More Reliable General AI Systems
   https://www.anthropic.com/news/anthropic-raises-124-million-to-build-more-reliable-general-ai-systems
9. TIME — Jaan Tallinn — TIME100 Philanthropy
   https://replatform-prod.time.com/collection/time100-philanthropy/2026/jaan-tallinn/
10. WIRED — UK Lawmakers Are Scrambling in Response to AI’s Summer of Chaos
    https://www.wired.com/story/uk-lawmakers-are-scrambling-in-response-to-ai-summer-of-chaos/
11. U.S. Senator Bernie Sanders — Sanders, Casar Introduce Legislation to Ban Artificial Superintelligence and Temporarily Pause Advanced AI Development
    https://www.sanders.senate.gov/press-releases/news-sanders-casar-introduce-legislation-to-ban-artificial-superintelligence-and-temporarily-pause-advanced-ai-development/
12. Reuters — Trump Says There Is a “Sick Conspiracy” Against AI and Data Centers
    https://www.reuters.com/world/trump-says-there-is-sick-conspiracy-against-ai-data-centers-2026-09-14/
13. Dario Amodei — We Must Pace the Frontier
    https://darioamodei.com/post/we-must-pace-the-frontier

Your three original setup articles, which I’d also put at the bottom of the YouTube description/reference section:

14. RedState — Dirty Little Secret Comes Out About Bernie Sanders’ Bid for Government-Controlled AI
    https://redstate.com/kyle-becker/2026/09/14/dirty-little-secret-comes-out-about-bernie-sanders-bid-for-government-controlled-ai-n2206886
15. RedState — Trump: “Sick Conspiracy” Against AI and Data Centers Only Helps China
    https://redstate.com/rusty-weiss/2026/09/14/trump-sick-conspiracy-against-ai-and-data-centers-only-helps-china-n2206892
16. ZeroHedge — Full Stop? Amodei, Altman & Musk Team Up Against AI Singularity While China Goes Full Send
    https://www.zerohedge.com/ai/full-stop-amodei-altman-musk-team-against-ai-singularity-while-china-goes-full-send

Trump Just Went “All In” on AI. Elon Musk Says Slow Down. This Is Going to Get Ugly.

The Trump–Musk split is only the surface. Underneath it sits a much bigger fight over who gets to control artificial intelligence — the companies, the government, the evaluators, or the machine itself.

Donald Trump just planted his flag in the artificial-intelligence race.

And Elon Musk is standing somewhere else.

Not because Trump launched a personal attack against Musk. He didn’t. The disagreement is more important than another personality feud: they have landed on different sides of the argument over whether America should deliberately slow the development of increasingly powerful AI.

Anthropic CEO Dario Amodei recently called for the frontier of advanced AI to be “paced.” OpenAI CEO Sam Altman backed the proposal. Elon Musk’s response was even shorter: “Dario is right.”

Trump’s response to the broader slowdown movement was essentially the opposite.

He argued that America already has mechanisms capable of regulating and prosecuting AI companies when necessary, warned that excessive fear could benefit China, and wrote that:

“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART … PRESIDENT.”

Trump also characterized the industry’s demand for additional regulation as a “hoax.”

And suddenly we have one hell of a technological argument.

Trump says America cannot afford to take its foot off the accelerator.

Musk, Altman and Amodei are saying the accelerator itself may now deserve considerably more respect.

But that’s only the beginning.

The people building AI are admitting something uncomfortable

Amodei’s argument deserves to be stripped of the technical jargon because the basic concept isn’t difficult.

Artificial intelligence increasingly helps researchers develop better artificial intelligence.

AI writes code. It assists with research. It tests systems. It helps design pieces of the next generation of models.

Eventually, that process could accelerate.

An AI system helps humans build a better AI system. That improved system becomes better at helping build the one after it.

Amodei calls attention to recursive self-improvement, arguing that progress has recently accelerated substantially because AI is becoming increasingly useful in AI development itself. He warns that this pace could outrun researchers’ ability to understand and control increasingly capable systems.

And then comes perhaps the most unsettling admission in the entire debate.

Anthropic is one of the world’s most sophisticated AI laboratories, yet Amodei writes that despite substantial progress in interpretability — understanding what is happening inside a model — researchers still understand only a tiny fraction of what happens internally.

Read that again.

The politicians don’t fully understand these systems.

The public certainly doesn’t.

And even the laboratories building them acknowledge significant limits in understanding exactly what occurs inside them.

Yet somebody is going to have to write the rules.

That’s where this story gets interesting.

Who watches the machine?

Amodei’s answer includes something called embedded evaluators.

Under his proposal, frontier AI companies would provide independent outside evaluators with unusually deep access to their operations.

Not merely a demo account.

Employee-like access.

Anthropic says it intends to provide outside reviewers with office access, company laptops, internal tools and permissions similar to those available to employees performing comparable risk assessments. Those evaluators would be able to report findings without Anthropic controlling their conclusions, subject to limited protections for privileged, confidential or security-sensitive information.

One organization Amodei specifically cites as an example is METR — Model Evaluation and Threat Research.

METR is already doing serious work inside the frontier.

Earlier this year, Anthropic, Google, Meta and OpenAI participated in a METR pilot assessment. The companies provided access to highly capable internal models and nonpublic information concerning capabilities and internal AI use. METR says the participating companies did not possess editorial control over its published report.

If organizations like METR eventually become the independent referees of advanced AI, one obvious question follows.

Who funds the referees?

Follow the money — carefully

There has been speculation online that Anthropic directly funds METR.

METR says that isn’t true.

The organization says it has not accepted funding from frontier AI companies, nor donations made by or at the direction of employees of those companies. METR does acknowledge that frontier laboratories provide significant amounts of free model tokens for evaluations, research and engineering.

METR also disclosed that it raised approximately $71 million in commitments during the six months preceding August 2026. Its listed supporters over time include major foundations, philanthropists and the Survival and Flourishing Fund.

That takes us to one of the more interesting figures in the AI-safety world.

Jaan Tallinn.

Tallinn helped create Skype and has spent years investing significant resources in reducing existential risks from artificial intelligence.

He also has history with Anthropic.

Anthropic’s own 2021 financing announcement states that Tallinn led its $124 million Series A round.

TIME reported this year that Tallinn’s Survival and Flourishing Fund has organized approximately $150 million across more than 300 projects, specifically naming METR among the recipients.

Now let’s be precise.

That does not establish that Anthropic controls METR.

It does not establish that Tallinn dictates METR’s conclusions.

It doesn’t establish a secret conspiracy.

What it establishes is something simultaneously less sensational and more important:

a relatively compact ecosystem of AI companies, evaluators, investors and safety organizations whose relationships increasingly intersect with public policy.

And the trail doesn’t stop there.

Another organization appears: ControlAI

WIRED recently reported that a nonprofit called ControlAI drafted proposed British legislation aimed at banning artificial superintelligence.

According to WIRED, ControlAI is backed by Tallinn.

Its U.S. operation has also become active in Washington. WIRED reported that ControlAI’s American arm consulted on the Sanders-Casar proposal to ban superintelligence development and has briefed nearly 200 congressional offices and more than 20 members of Congress about superintelligence risk.

Senator Bernie Sanders and Representative Greg Casar announced their proposed Ban Artificial Superintelligence Act on September 3.

Their proposal would permanently prohibit the development and deployment of artificial superintelligence and temporarily pause some advanced AI development until a federal regulator establishes safety rules. It would also direct the United States to pursue international agreements aimed at preventing superintelligence development elsewhere.

Now pull the camera back.

Tallinn led Anthropic’s early Series A financing.

A philanthropic fund associated with Tallinn has supported METR.

METR evaluates frontier AI laboratories including Anthropic, OpenAI, Google and Meta.

Anthropic’s CEO proposes dramatically expanding the role of independent evaluators such as METR.

ControlAI, another organization reported to have Tallinn’s backing, drafted proposed legislation in Britain and consulted on the Sanders-Casar proposal in America.

Again: that isn’t proof of central control.

It is an influence network worth understanding before governments redesign the rules governing an industry potentially worth trillions of dollars.

JD Vance asked the question everyone should ask

Vice President JD Vance publicly noted something unusual about the situation.

He said he felt uneasy watching so many frontier AI companies approach Washington asking the government to regulate them.

The skepticism raises a legitimate economic issue regardless of one’s political position.

Suppose Washington creates expensive licensing requirements, audits, compliance departments, reporting rules and mandatory testing regimes.

Who can afford them?

Anthropic.

OpenAI.

Google.

Meta.

xAI.

Who potentially struggles?

The startup that hasn’t raised billions.

The university laboratory.

The independent developer.

The next company nobody has heard of yet.

Regulation can serve a legitimate public-safety purpose while simultaneously creating barriers to entry that benefit incumbents.

Those two things can be true at the same time.

That’s why reducing this debate to “AI companies are trying to save humanity” or “AI companies are conspiring to control the market” is intellectually lazy.

Reality is nastier.

Everyone can sincerely believe they’re protecting the public and still possess incentives that happen to strengthen their own position.

You don’t need a conspiracy when incentives already point in the same direction.

Then China walks into the room

Trump’s objection becomes much easier to understand once China is added to the equation.

If the United States deliberately slows advanced AI development while China continues at maximum speed, Washington hasn’t ended the race.

It has potentially changed who wins it.

Trump argues that excessive restrictions could assist China and says the United States should continue aggressively developing AI.

Interestingly, Amodei doesn’t dismiss that concern.

He explicitly argues that pacing only works if democratic countries preserve enough of a technological lead over China to prevent a slowdown from creating a national-security disaster. He advocates tighter restrictions on advanced chip access, combating chip smuggling, preventing model theft and protecting America’s technological lead while simultaneously pacing frontier development.

So Trump and Amodei actually agree on a major premise.

China matters.

They disagree violently about what America should do about it.

Trump’s position is essentially that America must continue running.

Amodei’s position is that America must remain ahead while learning how to run without driving straight off a cliff.

Musk has now publicly sided with Amodei’s call for greater caution.

And that’s how Donald Trump and Elon Musk ended up on opposite sides of one of the most consequential technological arguments imaginable.

The problem isn’t old men. It’s confidence without competence.

There is an easy joke here about aging politicians trying to regulate technology they barely understand.

But age isn’t actually the issue.

Competence is.

A 75-year-old engineer who understands neural networks is infinitely more qualified to discuss AI policy than a 35-year-old politician who doesn’t.

The deeper problem is that decisions of enormous consequence are now being made by institutions attempting to govern a technology whose behavior isn’t completely understood even by the laboratories building it.

Amodei himself says today’s interpretability techniques remain profoundly incomplete. METR says frontier systems are changing rapidly enough that risks from autonomous agents could increase substantially over short periods.

So perhaps humility is the missing technology.

Government says it can regulate AI.

Corporations say they can build it safely.

Evaluators say they can measure the danger.

Investors finance the companies and parts of the safety ecosystem.

Advocacy groups help lawmakers understand the threat and sometimes help shape proposed legislation.

Meanwhile the systems keep improving.

Who exactly is controlling whom?

Maybe artificial intelligence isn’t controlling humanity.

Not yet.

But humans are already engaged in an extraordinary fight over who gets to control artificial intelligence.

Trump wants America moving aggressively enough to maintain dominance.

Musk and several leading AI executives say the frontier needs more deliberate pacing.

Lawmakers are considering new regulators.

Anthropic wants embedded independent evaluators.

METR is already examining frontier laboratories.

AI-safety organizations are advising lawmakers.

Investors and philanthropists fund overlapping pieces of the ecosystem.

And China keeps developing its own systems while everybody debates what the rules should be.

Perhaps the question we’ve been obsessing over is premature.

Everyone keeps asking:

What happens when AI becomes powerful enough to control us?

There is another question that arrives first.

Which humans are going to control AI?

And if artificial intelligence becomes as economically, militarily and politically powerful as its architects believe it might—

the answer to that question could determine considerably more than the future of technology.

It could determine who holds power in the world that technology creates.

Categories: Politics

Tags: #ai, #AIRegulation, #AISafety, #Anthropic, #artificialintelligence, #china, #congress, #DarioAmodei, #METR, #OpenAI, #technology