Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
firethunder7
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
5 ms
·
1.
▲
by
firethunder7
2mo ago
From what I understand, the goal is to train an LLM that is better at training LLMs than humans, so that it can continuously train smarter models and, once smart enough, design the successor to LLMs.
2.
▲
by
firethunder7
2mo ago
My bad it was the coding agent index
3.
▲
by
firethunder7
2mo ago
AA coding index has been updated to use DeepSWE, Terminal-Bench v2, and SWE-Atlas-QnA.