Before posting confident takes on LLM consciousness (whatever side you're on) maybe at least get yourself familiar with the long history of arguments about whether consciousness can arise from purely mechanical processes? Many of the most obvious takes are wrong...

Another nail in the coffin of the Octopus Thought Experiment. You *can* learn the structure of the physical world from text alone.

Album of the day: "Petite Constellations" by Bluetech. Their 2005 "Sines and Singularities" was for years my go-to for focussed work. This new work, 20 years later, is still wordless electronica, but simpler, warmer, reminiscent of pioneers like Giorgio Moroder and Jean-Michel Jarre.

Doesn't "clean room" imply that they did *not* work from the originals, i.e. decompile?

increasingly convinced that—whether they admit it or not—AI denialists think that this is a debate around whether LLMs are ensouled

Seeking an engineer to help with a large-scale, academic study of the effects of long-term AI use on young adults. Building a Chrome extension to collect conversations with chatbots securely and reliably. Experience with mobile app development is also a plus. DM me.

Amazing how libraries went from “keeper of knowledge” to “archive of historic objects” in what, two decades? Also they have some nice study spaces. Here’s my favorite nook in Balliol College library, Oxford

I'm over AI animated videos as a debate format. It's not that I don't worry about AI persuasion, but there can be no durable advantage. Persuasion is a competitive, adversarial ecology. Humans quickly evolve memetic defenses. No one would buy the Brooklyn Bridge today.

An entirely reasonable take on the whole "is it infosec or alignment?" debate. Spoiler: both.

it's living rent free in my head tbh

I suppose I have a much broader view of what healthy public discourse looks like. I *want* all comers to be to on the same platform, because the alternative is epistemic fragmentation which leads to worse ills. Of course this only works if you have constructive algorithms. Which we do, now...

It's a reference to Wargames, the movie, where the rogue AI eventually learns through simulation that launching the missiles is a bad idea. Interestingly, I am starting to consider self play to teach conflict resolution to AIs.

Now do global thermonuclear war

I made a little Jev-based language model! It uses beam search + Jev as a classifier to judge branches based on coherence and "success" (whatever that means). I've barely talked to it so I don't know how it'll behave or if it'll be interesting at to talk to at all, let's find out! @jevbot.bsky.social