Why things will eventually fall apart: 1. Everybody, even Google, seems to be treating AI as if it were some kind of winner take all competition like web search was, in which Google taking over 95% 2. But everybody is building essentially the same technical solution with essentially the same data, so there is no moat. 3. If there is no moat, nobody is going to take 90% of the market. 4. With no clear winners, nobody can charge monopoly prices; instead, you get price wars and commodity pricing. 5. Which means everybody will wind up overpaying compared to the modest profits they will be able to make in an intensely competitive regime. Am I missing something?

Researchers found our current approach to making AI smarter over time has a giant blind spot. AI is not actually understanding or applying high-level abstract lessons at all. Developers spend massive amounts of time building systems that condense past AI mistakes into neat little rules for the future. This paper proves that the AI essentially throws those rules in the trash and only looks at raw historical logs. Modern LLM systems try to get better over time by storing past tasks as either raw step-by-step histories or condensed summary rules. The study tested if these agents actually use their stored memories by secretly swapping the correct tips with random garbage text. - When the step-by-step histories were messed up, the AI failed hard, proving it heavily relies on copying exact past actions. - But when researchers completely corrupted the condensed summary rules, the AI kept acting normally and showed zero performance drop. If an AI cannot apply an abstract lesson to a new situation, it is not truly reasoning or learning. This raises the question if the entire AI industry need to rethink how memory works because right now these agents are just mimicking instead of understanding. ---- arxiv. org/abs/2601.22436 "LLM Agents Are Not Always Faithful Self-Evolvers"

Make no mistake: post-Mythos, the United States has a licensing regime for AI. It’s just informal, with no consistent rules or firm boundaries on state power or public transparency. Cobalt mining in the Congo is vastly more institutionalized than frontier AI licensing in the US.

The problem @DavidSacks is that the Administration has a credibility gap on whether decisions are being made with political or retributive motives or for the national interest. Trump's mantra has been, until now, accelerationism and no regulation. Why not work with Congress to set up an independent agency for AI safety, like the nuclear commission or FERC, so that the public has confidence in these decisions?

Wow! OpenAI subpoena’d big time! “documents related to a range of its activities and impact on users, including advertising, user engagement and retention, handling of consumer data and health data, activities related to minors and seniors, deep learning models, model sycophancy and company policies” What a wild night!

Peter Thiel and Balaji Srinivasan have funded a platform called Objection.ai that allows anyone to file a complaint against a journalist's story for a starting price of $2,000. A team of human investigators examines the story, then submits findings to a "jury" of AI models - OpenAI, Anthropic, Grok, Google - which publish a "verdict" on the story's truthfulness and rank individual journalists on metrics including truth-telling and corrections. If the journalist doesn't respond to defend their reporting, the verdict is issued and published online anyway. The platform is being sold as "letting anyone fight the press like a billionaire." The creator is Aron D'Souza, who led the Thiel-funded lawsuit that bankrupted Gawker in 2016. The design choices tell you what this is. The system treats anonymous sources as less trustworthy and ranks anonymous whistleblower claims near the bottom. Anonymous sources are how most significant accountability journalism happens - they're how the Pentagon Papers got out, how the CIA's black site program got exposed, how the HHS stories we've covered this week were reported. The people who most need protection from powerful interests are specifically deprioritized by Objection's scoring system. The creator calls it "the same as Community Notes." A civil rights and defamation attorney calls it "a high-tech protection racket for the rich and powerful." One of those descriptions is accurate. The AI models being used as the "tribunal" were trained on journalists' work without consent or compensation. They hallucinate. They amplify bias. They are being deployed here specifically to issue verdicts on the work of the people whose labor built them. Thiel killed Gawker with a lawsuit. This is faster and cheaper.

Taken at face value, the Trump admin. ban on Fable calls into question the entire AI industry. The underlying technology is non-deterministic and therefore inherently unpredictable. It’s impossible to build in safeguards to reliably prevent jailbreaks. This is a turning point.

Interesting article with new details on the lead up to the export controls. Both sides are telling very different stories. Anthropic says they were given a 90 minute hard deadline to pull both models. The administration says their concerns were not taken seriously.

AI's Core Flaw: "Mass Regurgitation Of Misinformation" zerohedge.com/ai/ais-core-fl…

This might be a come to Jesus moment if the regulators suddenly realize that jailbreak prevention isn't real.

Everyone can see Anthropic's hypocrisy in pushing toward RSI while warning that doing so creates catastrophic risk. These All-In guys might not know the insiders' party line, the galaxy-brained “someone else will build it if we don't” logic. But they're still right: It's crazy.