niplav @niplav@schelling.pt

@Paradox not sure I understand

BRHS: BCHS but with Rust?

@Paradox They claim that human learning is a lot like current AI training: A lot of self-supervised pre-training+some fine-tuning+a little bit of RL (and in this view then multi-agent RL on top)

@Paradox If you're interested: https://www.lesswrong.com/posts/iCfdcxiyr2Kj8m8mT/the-shard-theory-of-human-values

If I were a Shard Theory person, I'd say that constitutional AI is a next step in training AIs in the similar way that humans are trained: Reinforcement learning from interacting with other agents, starting with a simple set of values

[[Cause X]] Create ontological libertarian free will (if it doesn't exist)
- Same but for objective moral truths & God

@rime Blocked in my country

@augustus I *also* have a twitter account, but sometimes I quarantine myself over here

@augustus Yeah nitter dying is a real shame

Ah well there was an attempt

inb4 it's that Stripe in general doesn't like me credit card

actually why am I planning to give money to *those* guys

Claude surely has an open API too…

GRRR why is OpenAI rejecting my credit card

juuust checking

$$instance$$

@puckipedia Hang on…

Using GPT-4 to fix an encoding bug in a program I wrote years ago, hope this doesn't bite me (I don't think it will)

But man X11 is *cursed*

Unsure whether the human mind is the best or the worst Linux distribution