HackerLangs
トップ新着トレンドコメント過去質問紹介求人

aldielshala

no profile record

投稿

Show HN: Llm.sql – Run a 640MB LLM on SQLite, with 210MB peak RSS and 7.4 tok/s

8 ポイント·投稿者 aldielshala·3 か月前·2 コメント

コメント

aldielshala
·3 か月前·議論
Haven't tested on a Pi yet, llm.sql is still in alpha, focused on validating that SQLite can actually work for LLM inference and profiling memory usage. That said, 210MB peak RSS should fit comfortably on a Pi. In theory, any device that runs SQLite (which is almost every device) could run llm.sql. Planning to benchmark across different hardware as the project matures.
aldielshala
·3 か月前·議論
Nice project. I'm also working on something that pushes SQLite well beyond its typical use case. It's encouraging to see more people exploring what SQLite can really do.
aldielshala
·3 か月前·議論
Curious how it handles 10K+ notes performance-wise, does it index everything or lazy-load?
aldielshala
·3 か月前·議論
Trying to use human attention, instead of Transformer attention.
aldielshala
·3 か月前·議論
Intent debt is a useful framing. A few comments explaining "why" instead of "what" would have saved hours of guessing.
aldielshala
·3 か月前·議論
Finally an AI that takes someone's job and nobody's upset about it.
aldielshala
·3 か月前·議論
My contribution today: fewer LLM calls, fewer GPU hours, less CO2.
aldielshala
·3 か月前·議論
Everyone's focused on Meta employees, but the real concern is normalization. If Meta does this and gets away with it, some companies may quietly roll out the same thing.
aldielshala
·3 か月前·議論
Yes, maybe context engineering (prompting is just one part of it) and soft skills.
aldielshala
·3 か月前·議論
Honestly, I doubt this data is as useful as they think.

Half my workday is me browsing random tabs while an AI agent does the actual work. They're going to train a model on alt-tabbing and scrolling HN/Twitter/Reddit.
aldielshala
·3 か月前·議論
$60B for a VSCode fork with AI integration... It may show the value of the gap between vanilla LLM output and production-ready applications.
aldielshala
·3 か月前·議論
Communication, with both human and AI.