Hacker News Reader

yeah this is hacker news. we only share our data with Dario, Altman, Musk, and the CCP. Not untrustworthy people like zuckerberg.

I don't share my data with any of those either.

That leaves you with who, exactly, in the LLM game? Google? Mistral?

GLM 5.3 locally. Before that Deepseek, before that various Qwen versions and GPT-OSS.

"At a standard Q4_K_M quantization, the model footprint takes about 459.5 GB."