Europe 2031 is well-intentioned but regrettably timid. The authors hide the existential risk by AI in a fold-out FAQ section instead of addressing it head-on and this is just one case of them not taking their own premises seriously. Decision-makers need the truth. They won't make better decisions if you sugarcoat how dire the situation is. For Europe itself, the train has likely already left the station anyway.

Hard to say what Anthropic’s motivations are but it is true that half of all AI researchers think there are double digit odds that the technology will cause human extinction

Peinlich. Einfach nur peinlich.

If leading AI companies are indeed approaching the point of recursive self-improvement, a coordinated, verifiable, and universally applied pause is probably the only responsible solution to mitigate several major AI risks; at least until safety guarantees are developed and demonstrated. Ensuring that such a moratorium is respected would require sincere collaboration between various countries and companies, but I definitely believe it is achievable if others follow in @AnthropicAI's footsteps.

I've got a lot of quibbles with Anthropic's "When AI builds itself" blog post, but I appreciate them coming right out and saying this.

Corporate astrology, context dumps, conspiracy theories and how to make forgery really expensive - explained using fish Here are the worlds AI created for humans

Could an AI company lose control of its own agents? To find out, Anthropic, Google, Meta, and OpenAI let us (1) test their best internal models with CoT access, (2) review non-public info about capabilities, alignment, and control. The result: our first Frontier Risk Report.

The guy (@BigMeanInternet) blocked me immediately after asking this question, so I'll answer here instead: Any use of current AI tech, however alarming you might think it is and however despicable it might in fact be, is not one of the most alarming hazards of the technology when we talk about AI x-risk. Hinging your belief in credible expert warnings on them saying something untrue in order to force them to take your side in your favorite political hot-button topic is bad form and, worse, stupid.

Developing a superintelligent AI that does what we want without killing everyone may be extremely difficult. In this video, we explain why, using arguments from "If Anyone Builds It, Everyone Dies" by @ESYudkowsky and @So8res.

RT @TaylorLorenz: SCOOP: A pro-AI dark money group backed by a powerful super PAC funded by execs tied to Palantir and OpenAI, has been sec…

Friedrich Merz ist die Sorte Mensch, denen z.B. Anwälte oft energisch bis verzweifelt erklären müssen, dass sie zu ihrem eigenen Besten bitte einfach mal die Klappe halten müssen. Seine persönliche Tragödie ist, dass er mittlerweile die mächtigste Person im Raum ist und sich entweder keiner traut, ihm das zu verklickern, oder er denkt, dass er auf die nicht mehr hören muss.

Endlich ist es mir auch mal gelungen, bei irgendsoeinem irrelevanten Hype-Thema mal so richtig vor die Welle zu kommen (haha, Welle, versteht ihr?). Vielen Dank an den Suhrkamp Verlag dafür, dass er mir das nicht ermöglicht hat und natürlich an die anderen deutschen Diskurswale für die gute und schnelle, dafür aber auch frei erfundene Mitarbeit und natürlich große Entschuldigung an Richard David dafür, dass ich seinen Text leider um rund 120 Seiten kürzen musste. Gerne wieder! Den Sammelband findet ihr ab sofort in keiner gut sortierten Buchhandlung.

Never send to know for whom the gotcha gotchas; it gotchas for thee.

And OpenAI is on track to professionalize infowarfare like in this example.

actually that’s not impressive the concept of a Dyson sphere was already in the training data

Machine superintelligence would extinguish Democrats, Republicans, British, Chinese, scientists, cab drivers, and polar bears. It is a sign of hope that all of those now seem to be saying they'd prefer otherwise (except the polar bears).