nick m Sabine Hossenfelder I indeed think that whenever even just an extremely slight change in the physical brain state that IS (not encodes, or gives rise to, or generates, but literally IS) your world simulation, occurs, then yes, there is a change in your world simulation as well (because again, the

"Addiction potential" or "abuse potential" are terms that are not meaningfully applicable to the next-generation drugs I am developing, which will have negligible to non-existent side effects. I am designing these compounds to be suited precisely for long-term daily use, which

When I finally create an outstanding valence enhancement drug, it will constitute abuse to NOT take it. Only those hopelessly addicted to puritanical masochism will abuse themselves (and their hedonic setpoints) like that.

I am looking forward to watching GPT-7 doing a line by line analysis and rebuttal of Gary Marcus’s line by line analysis and rebuttal.

The more AI capabilities advance, the more salient the below becomes, and the more I hope that xAI wins the race. Most other AI companies think that engineering technologies for voluntary valence enhancement constitutes "misuse" or "abuse". This is unspeakably tragic. For all I

If you plan on continuing to sleep in a bed and not on the floor indefinitely, that means you are addicted and the bed must be taken away from you, and the more you resist the more it proves how addicted you are and that the bed really really really must be taken away from you

I am happy to report that this has indeed been remedied with o3 Pro, which oftentimes thinks for 15+ minutes even after dozens of messages. I have not detected any "dumbing down" as the chat length increases.

GuilleFix — k/acc | /dd ⏩ I think that ultimately all you want is the positive valence, and being able to easily produce extremely positive valence with advanced neurotech (perhaps accompanied by suffering too, in a kind of a bliss-to-suffering-to-bliss-to-suffering-to-bliss etc… "roller coaster"

Kris Captain Pleasure, Andrés Gómez Emilsson Perhaps enchantment is not the right word. I think what Andrés was trying to say is that for the most part, intelligent people (to whatever extent they can even decouple qualia from world models, which they mostly can’t, which is another huge problem, but I digress) feel that

It’s legit insane that nobody sees the gargantuan money-making potential of valence enhancement. Like okay, even if you think it’s perverse and unnatural an "raises important questions about what it means to be human" and wireheading bad because Brave New World, surely you must

Free will, information-rich non-qualia, and experiencers separate from experiences are good examples of things people implicitly believe in that are "not even wrong". They are downright incoherent as concepts.

Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) Good question. A LLM is a high-level abstraction that strictly speaking only exists in the map, not in the territory outside of the map. (Again, very strictly speaking) there are no such things as LLMs in the territory outside of our skulls - there is only the hardware of the

Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) An LLM is software, it is by definition substrate-independent. It can not exist in the territory outside of the map because the real territory does NOT support substrate independence, otherwise reductionism would be violated. In your view, what ARE the LLMs that exist in the