Skip to content

Self-hosting. Commercially available LLMs are increasingly hampered by cost

p4sc4l.substack.com
1 pointinferhaven2 comments
On HN

Comments

I wish I could run models locally. I just don't have the hardware for it. So I rent access to one that still works well enough and hasn't annoyed me yet. But "yet" is doing a lot of work there. The article is right that the trend is worrying. You're at the mercy of someone else's cost cutting and someone else's political compass. I just can't afford to opt out right now. That's the part that stings the most.

I believe local AI is the future of most daily workflows

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.