Hacker News: Ancv123https://news.ycombinator.com/user?id=Ancv123Hacker News RSShttps://hnrss.org/hnrss v2.1.1Sun, 09 Aug 2026 07:00:51 +0000<![CDATA[New comment by Ancv123 in "Timeline of the OpenAI accidental attack against Hugging Face"]]>Maybe they didn't have proper debuggers in 1960? For a language model you need (RNG state, context, prompt).

So if they wrote an LLM step by step debugger, it would be all deterministic. But they prefer rapid sales, chaos and mystique.

]]>Sat, 08 Aug 2026 13:56:24 +0000https://news.ycombinator.com/item?id=49221868Ancv123https://news.ycombinator.com/item?id=49221868https://news.ycombinator.com/item?id=49221868<![CDATA[New comment by Ancv123 in "Timeline of the OpenAI accidental attack against Hugging Face"]]>I'm just reading the captions of the video for May 7th. They clearly say at 10:18:

"we kick off a new reinforcement learning run to train a next frontier model.

It the captions are correct, there is no ambiguity.

]]>Sat, 08 Aug 2026 13:47:07 +0000https://news.ycombinator.com/item?id=49221806Ancv123https://news.ycombinator.com/item?id=49221806https://news.ycombinator.com/item?id=49221806