Posts

New ask Hacker News story: AI Regex Scientist: A self-improving regex solver

AI Regex Scientist: A self-improving regex solver 3 by PranoyP | 1 comments on Hacker News. I built a system where two LLM agents co-evolve: one invents regex problems, the other learns to solve them. The generator analyzes the solver's failures to create challenges at the edge of its abilities. The result: autonomous discovery of a curriculum from simple patterns to complex regex, with a quality-diversity archive ensuring broad exploration. Blog: https://pranoy-panda.github.io/2025/07/30/3rd.html Code: https://ift.tt/NFghjtn
Post a Comment
Read more

New ask Hacker News story: 10 months since the Llama-4 release: what happened to Meta AI?

10 months since the Llama-4 release: what happened to Meta AI? 2 by Invictus0 | 1 comments on Hacker News. I understand Llama 4 was a disappointment, but what's happened at Meta since then? Their API is still waitlist-only 10 months on.
Post a Comment
Read more

New ask Hacker News story: Ask HN: Anyone Using a Mac Studio for Local AI/LLM?

Ask HN: Anyone Using a Mac Studio for Local AI/LLM? 4 by UmYeahNo | 2 comments on Hacker News. Curious to know your experience running local LLM's with a well spec'ed out M3 Ultra or M4 Pro Mac Studio. I don't see a lot of discussion on the Mac Studio for Local LLMs but it seems like you could put big models in memory with the shared VRAM. I assume that the token generation would be slow, but you might get higher quality results because you can put larger models in memory.
Post a Comment
Read more

New ask Hacker News story: Ask HN: Will Crypto Currencies survive past this market downturn?

Ask HN: Will Crypto Currencies survive past this market downturn? 3 by halamadrid | 2 comments on Hacker News.
Post a Comment
Read more

New ask Hacker News story: Zendesk Email Spam

Zendesk Email Spam 25 by Philpax | 9 comments on Hacker News. Looks like there's another round of Zendesk email spam happening. I've gotten hundreds over the last half-hour.
Post a Comment
Read more

New ask Hacker News story: There is no evidence for X

There is no evidence for X 2 by cadabrabra | 0 comments on Hacker News. Is not the same as “There is evidence of no X.” Very important distinction.
Post a Comment
Read more

New ask Hacker News story: YC S26 Application: "Attach a coding agent session you're particularly proud of"

YC S26 Application: "Attach a coding agent session you're particularly proud of" 2 by simplydt | 1 comments on Hacker News. I vibecoded a couple of iOS apps & a full SaaS (SEOZilla.ai) over the past six months and the honest answer is: my best coding agent sessions from 3-4 months ago would make great submissions. Excellent debugging, catching poor architecture choices, back-and-forth problem solving. But lately? I mostly write product specs, make simple architecture decisions, and do QA. The agents just... handle it. Across the board, Opus, Sonnet, Cursor, whatever you're using, the jump in the last 2 months has been wild. Which raises a genuine question: what is YC actually selecting for with this prompt? The most impressive sessions are probably from people using worse tools or tackling harder problems. The founders who've figured out the best workflows might have the most boring transcripts. Anyone else finding that their "best" agent sessions...
Post a Comment
Read more