logoalt Hacker News

m3kw9 • today at 1:47 PM • 1 reply • view on HN

A better idea is to have Astra modify it's own weights and let it try if it can RSI, but you gonna need to closely monitor for actual leap in capabilities so you can keep a check on alignment.


Replies

Aboutplants • today at 2:04 PM

Someone has to already be doing this right?

➕ show 1 reply