Hacker News Reader

I run them all on an old Pentium J (Atom) NUC with 8GB RAM, so I don't even care. Some Chinese N100 mini PC for $100 is all one needs.

View on HN

Using local models or external?

If external, aren't they pricey? How much tokens they generate?

If local, what runs on such hardware that gives reasonable results?

Local models on different machines with multiple RTX Pro 6000 or multiple DGX Sparks or a 512GB RAM Macstudio; the agents themselves run on that Pentium J NUC and just use exposed endpoints for local models. Forgejo for Git runs on another server. Therefore I don't really care if that NUC goes kaboom and can test everything quickly (OpenClaw, Hermes, Claude Code, Codex, OpenCode, Pi etc.). Or I can just use OpenRouter API key and access 10-100x cheaper models than Opus.

Reply on news.ycombinator.com