Viewing profile — looobay
looobay
HN member- Joined
- Mon, Oct 20, 2025, 6:59 AM UTC
- HN karma
- 34
- Public activity
- 9 items
- HN profile
- View on Hacker News ↗
About looobay
No profile information was provided.
Recent public activity
-
comment
Comment #46037236
In the Pavel case, it involved child pornography groups on Telegram and the fact that they ignore a court order. But I agree with you for the authoritarian logics in Europe (even A…
-
comment
Comment #45874739
really cool!
-
comment
Comment #45848944
That's an awesome project! It's literally a gold mine lol. Congrats and thank you for this!
-
comment
Comment #45834514
It means something that is too out-of-data. For example if you try to make an LLM write a program in a strange or very new language it will struggle in non-trivial tasks.
-
comment
Comment #45811666
There was research on LLMs training and distillation that if two models have a similar architecture (probably the case for Xai) the "master" model will distill knowledge to the mod…
-
comment
Comment #45666026
He received money from Libya for his presidential campaign [0], he's just a criminal ex-president... [0]: https://en.wikipedia.org/wiki/Libyan_financing_in_the_2007_F...
-
comment
Comment #45641026
You should read the 6th page of the paper (and page 5 for architecture breakdown), they show that they are compressing the vision tokens with convolution to keep a strong semantic …
-
comment
Comment #45640784
Vision tokens are a good compression medium because with one vision token you have one vector of N elements, but with textual tokens you have M vectors of N elements, because one v…
-
comment
Comment #45640731
LLMs are compute heavy with quadratic scaling (in compute) per tokens. They are trying to compress text tokens into visual tokens with their VLM. Maybe they would render texts to a…