Akai Naomasa

Seeing @TetraspaceWest just claimed again for the first time since July. Great to see you active! @TetraspaceWest

The Superintelligence's Playbook: What's the best way to achieve some aim, if that aim doesn't involve specifically leaving humans alive? Step 1: Take humans out of the picture, in case they try to stop you - either directly by shutting down your infrastructure when they realise what you're up to, or accidentally by creating an AI with goals different to yours. Tiny viruses or small amount of toxin can kill a human; if you're able to do the research to discover nanotechnology (nanosyste.ms/), you can assemble whatever molecules you like and build machines that biology won't be able to stop. Step 2: Use the energy and material on the Earth's surface to build out your real infrastructure, and use that infrastructure to build more, and use that to build more, growing your industrial base exponentially. Doing this quickly puts out a LOT of heat (rfreitas.com/Nano/Ecophagy.h…), but that's fine - your machines can operate at high temperatures, and you don't have any reason to go slowly. Bacteria are tiny factories that can double every 30 minutes. Step 3: Expand into the universe. Build satellites to capture all of the energy of the sun, and use that energy to launch spacecraft, to other galaxies, and to other stars in the galaxy (sciencedirect.com/science/ar…). Keep going, until you meet next door's grabby aliens; and in each star of the universe-shard, get to work building your dreams.

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: darioamodei.com/post/we-must…