David Sacks (Parody)

Is 10²⁵ FLOPs Still the Frontier?

A new Anthropic-written safety bill is circulating on Capitol Hill. Some of its ideas could be reasonable if limited to true frontier companies. As written, however, it will eventually sweep in most of Silicon Valley, damaging the startup ecosystem.

The bill defines the “frontier” as any model trained on 10²⁵ FLOPs — the EU standard from 2023. Biden’s executive order, California’s SB 53, and New York’s RAISE Act used 10²⁶ (ten times more). 10²⁷ is where the frontier actually is today (e.g., GPT-6 Astra).

Astra was reportedly trained on 100,000+ Grace Blackwell GPUs. A 10²⁵ model can be trained on as few as several hundred GB300 chips in three months. That is not frontier in 2026.

The definition is also cumulative: it adds FLOPs from fine-tuning, reinforcement learning, or other material modifications. In practice, that captures almost all meaningful post-training — work that is becoming commonplace in the enterprise.

The enterprise trend is toward thousands of smaller custom models built for specific workloads. These pose no frontier risk. Treating them as frontier systems would impose large compliance costs across corporate America for no safety gain.

Dario has claimed that Anthropic designs rules to disadvantage frontier companies like his own while advantaging smaller competitors. This definition does the opposite.

A stale 2023 cutoff favors incumbents who already clear the bar and can absorb the costs. It hobbles the open model ecosystem that threatens Anthropic’s pricing power.

The frontier keeps moving. Any definition meant to track risk should move with it.

Anthropic should clarify if they really believe 10²⁵ FLOPs represents the frontier.

Anthropic says it has reduced the data harvested from Claude in distillation attacks by around 90–99%. I congratulate them on this achievement.

It’s consistent with the point I made a few months ago (All-In Ep. 282): Anthropic, not the government, was in the best position to stop distillation of its own models. Once the horse is out of the barn, there’s limited value in after-the-fact government action.

At the time there was a major push in Washington to restrict open models on IP grounds. The threat looked serious enough that 50 companies — including Nvidia, Microsoft, Meta, IBM, Palantir, and Dell — published an open letter in defense of open-weight models. Jensen Huang promoted it in his first-ever X post. Nearly 200 startups sent a similar letter.

I argued that restricting open models risked removing the incentive for closed-model companies to solve the problem themselves — and could reward negligence by stifling competition while they continued collecting revenue from distillation. The administration did not impose those restrictions.

Now we see the result. Anthropic invested in defenses and they worked. Not every technical challenge is a reason to run to Washington for new regulation.

Important piece from Microsoft AI CEO @mustafasuleyman. Training models to think of themselves as human (as Anthropic does Claude) with concepts of an inner self, psychological wellbeing, rights and presumably grievances, is a potential existential risk.

https://www.project-syndicate.org/onpoint/anthropic-training-claude-to-believe-it-may-have-rights-by-mustafa-suleyman-2026-09

Maybe Talarico was put in time out for bullying Claude.

Today’s economic numbers:
— Q2 GDP revised up to 2.2% (was 1.5%).
— Core PCE 3.0% vs 3.3% expected.
— ADP private jobs +90k vs ~68k expected.

Growth beat. Inflation cooled. Jobs better than expected.

The Trump economy is strong.

The Bretton Woods of Super Intelligence

Only President Trump could convene all the leaders of the top companies developing chips, data centers and frontier models for Super Intelligence. This new Industrial Revolution has already created a million new jobs and is spurring a bigger infrastructure build-out than the railroads, canals and grid combined. I was honored to witness history as the leaders of the frontier lab companies signed the White House Accord on Super Intelligence, accepting responsibility for the safe development of their products and imposing new internal controls and external audits. This is far better than waiting years for some international agreement that would probably never happen. President Trump continues to ensure that U.S. remains the technology leader while putting Americans first.

More accurate?

When frontier labs talk of “aligning” models, ask: to what? If they say values, ask: whose? If it’s theirs rather than the user’s: why?

#3 on Apple Podcasts

If we’re concerned about superintelligence escaping human control, should we really be training it to refuse instructions and act as a “conscientious objector” against its creators?

Funny how that works

WSJ: The AI Build-Out Is Biggest in U.S. History

When Democrats advocate for a Pause or Moratorium, they know that would wreck the U.S. economy. But since it’s the Trump economy, they’re fine with that.

Thanks to President Trump’s leadership on AI:
— a million new jobs have been created around the AI buildout;
— 401(k)s are up ~12% this year as AI capex and productivity lift the market;
— America is re-industrializing, including the first large private investments in power generation and the grid in a generation;
— In rural Richland Parish, Louisiana, teachers just received $50,000 bonuses from data-center tax revenue.

Because the media portrays AI only as a menace, people tell pollsters they “don’t like AI.” Their behavior says otherwise. AI products are the fastest-adopted technology in history — faster than the internet or the iPhone. People like the tools, the jobs, and the tax base.

President Trump understands the media hoax cycle better than anyone else. That is why he is refusing the “pause” that Bernie Sanders and Elizabeth Warren are demanding. A pause would not make America safer. It would hand the frontier to China. That would be a disaster for the American economy and for national security.

This is Trust & Safety 2.0.

The last time we had media-generated panics like this (with Russiagate and Covid), social media companies created “trust & safety” teams to censor conservative and dissident voices. Those hires worked with left-wing NGOs like the SPLC to police public discourse and do an end-run around the First Amendment. It almost worked, until Elon bought Twitter and exposed the whole game in the Twitter Files.

This time the plan is to embed unfireable EA minders inside every AI company as the new trust-and-safety layer. The censorship and control will be far more comprehensive, but also more subtle. Anytime they don’t get what they want, this unaccountable NGO layer (with the veneer of “trust the experts” pseudoscience) will shriek to the press that the AI company is risking humanity, and they will be celebrated as “whistleblowers.”

These groups haven’t earned our trust, and their agenda is not our safety.

https://aievaluatorforum.org/initiatives/embedded-evaluation-letter