Re:Astra can read and write weights to improve Qwen
by ghosty141 on Monday November 02, 2026 @04:41PM
(#50344212)
My understanding is that the underlying part is a next token predictor and that's what is "actually" editing the code but the part that decides what to do (to achieve a successful unit test result for example) is a mix of all the things including post training.
[ Parent
| on HN ]
Re:Astra can read and write weights to improve Qwen
by Vaushite on Monday November 02, 2026 @08:04PM
(#50346633)
It's all a next token predictor, which is trained at first to predict the next token. RLVR etc. are also training the next token predictor.
Re:Astra can read and write weights to improve Qwen
by hypfer on Tuesday November 03, 2026 @02:22AM
(#50349062)
I think what breaks people's minds is just how far one gets with probabilistic approaches.
Which is interesting, given that life itself is merely that. It's just that humans decided that that ain't good enough and invented something alien to nature - logic.