Astra can read and write weights to improve Qwen

Re:Astra can read and write weights to improve Qwen by FeepingCreature on Monday November 02, 2026 @02:28PM (#50342301)

I suspect a lot of this is not (just) capability but bravery. A human would be daunted by even a small neural net; the task would seem too challenging to be worth attempting. Astra, fearless, just assumes an answer exists and then tries to reverse and execute on it.

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by ex-aws-dude on Monday November 02, 2026 @02:33PM (#50342377)

I've been finding that too with these models

For example pointing them at a crash dump they manage to piece together and extract information I would not even try to extract by just looking at the raw memory

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by nsoam on Monday November 02, 2026 @03:12PM (#50342998)

I once sent Gemini Pro a hex dump of a dmp file and it told me what the application was doing when it crashed. It seemed surprising, but not that surprising given what I've seen already.

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by Vaushite on Monday November 02, 2026 @04:19PM (#50343915)

It's the same thing is decompilation and hacking. It's not that a human couldn't find the same bug if they looked, but a human wouldn't even try because of the low probability of success.

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by tim333 on Tuesday November 03, 2026 @07:18AM (#50351072)

Human brains are not so good at dealing with a sea of thousands of numbers. Not that it's impossible but there are usually easier ways for us.

[ Parent | on HN ]