Skip to content

Comment on Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB Mac with at ~12 tok/sparent

Comments

For what it’s worth, this comment was not targeted at you, but rather the model kinda forcing it. I get the sense that Anthropic did not think much of this, but it seems to have gotten worse with recent models and it really comes off as a kind of nails on the chalkboard writing style.

I have to image whatever style of writing this was trained on is a lot more pleasant to read and I feel bad for whoever writes like this now being associated as bad AI writing.

most of this stuff isn't really possible at this point _without_ models, so pissing and moaning about people publishing work and not humanizing is really silly.

AboutSource Built by g1lg1l

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