Apple is developing robots - humanoid and others. Having written of $bns of investment in cars, it won't want another dead end. Competition is high, from Figure, and from Swastikars.

Conscium comment on the UK's decision not to sign the declaration from the Paris AI summit. “The Prime Minister said he wanted to mainline AI into the veins of the UK. That’s all very well, but he will find it harder to do so if he alienates allies."

@Matt_le_kat Thank you. Yes, most mainstream media interviewers still haven't understood what is happening, which is truly amazing.

New Conscium meetup! We're delighted that Jonathan Shock will be our guest for the next Conscium Meetup, at 4pm UK time on 25 Feb.

Amodei is right. The idea that advanced AI should be less regulated than a sandwich shop (H/T Stuart Russell) is crazy. Vance's claim to be defending the world against authoritarian regimes would be funny if it weren't so tragic.

Chinese researchers found that AI systems from Meta and Alibaba replicated themselves without human intervention. They also took steps to avoid shutdown. In other news, Andrew Ng still thinks concerns about AI risk are science fiction.

@timventura Tim is a very gracious and generous host. I really enjoyed this interview.

Frances announces €100bn investment in AI. The EU announces €200bn investment in AI. Pretty soon they'll be talking real money. Looks like Europe is finally taking AI development seriously. Good.

Talking on Sky News breakfast show about superintelligence and humanity becoming chimpanzees. Couldn't have done this even a couple of years ago. Things are moving fast.

I'm probably missing something, but why would people get excited about Apple using AI to make a cute table lamp, when OpenAI and the rest are acing maths tests for PhDs?

Until LLMs came along, China was closing the gap with the US on AI. DeepSeek's successes take us back from monopoly to duopoly. Your move, Europe.

France's announcement of $125bn investment into AI should be welcomed. And replicated around Europe. AI is humanity's most powerful technology, and it should not be a duopoly between the US and China. Especially now.

Growing human teeth in a pig's mouth sounds like a scene from Alien 4: Resurrection, but in reality it could be incredibly useful.

Tom Davenport has spent much of the last two years helping clients deploy LLMs productively. According to one study, only 6% of corporates had deployed them by early 2024, and it is only 20% now.

@TomMcClelland1 @guardian And a very good letter it is. "If there’s a genuine chance of developing conscious AI then we have to act responsibly." Spot on. Also agree that we are too "casual ... about suffering in many organisms." Maybe developing a concern for sentient AI could help there too.

AI agents may be widely available but not widely used this year because we resent robot mistakes more than we resent human ones. We need to design escalation paths to humans.

The 2023 Bletchley Park Summit brought "existential AI risk" into the Overton Window of world leaders. The Paris sequel is throwing it out again, Merde.

We discuss whether LLMs are predictive text models on steroids. They allocate numbers to tokens (word fragments) and the relationships between them, which means they do generate models of languages. To some degree, humans are predictive texters.

How splendid. I am number six in the top 30 futurists in the world. But I am not a number. I am a free man.