Hacker News: imenani

New comment by imenani in "AirLLM 70B inference with single 4GB GPU"

imenani — Mon, 03 Aug 2026 13:02:56 +0000

For anyone wondering “how slow is this?”

IIUC, Kimi K3 on RTX 6000 Ada (48GB) takes 292 s/token

https://github.com/lyogavin/airllm/releases/tag/v3.1.0

New comment by imenani in "Explorative modeling: Train on the best of K guesses"

imenani — Sat, 01 Aug 2026 18:16:10 +0000

The relation to current RLVR methods I think is interesting, they do discuss it a bit but I would be curious to see more about this as well. Quote from the paper:

Exploration beyond Pretraining. The mode collapse XMs address during pretraining also often shows up in post-training, where RL fine-tuning is known to sharpen models onto a narrow set of behaviors [76]. Recent fixes such as pass@krewards [77] and best-of-N-aware fine-tuning [78] can be seen through our lens as Forward XM, with a verifier standing in for ground truth data. These fixes act only during post-training, though; pretraining with exploration may yield base models that capture more modes in the first place, leaving RL more to select among.

New comment by imenani in "30papers.com – Ilya's 30 essential ML papers, in a beginner friendly format"

imenani — Tue, 07 Jul 2026 17:26:50 +0000

Nice presentation of the list!

I'd recommend watching a few of his talks/podcasts before during reading these to get the overview and how all the bits in these works tie together.

https://www.dwarkesh.com/p/ilya-sutskever

https://simons.berkeley.edu/talks/ilya-sutskever-openai-2023...

https://www.dwarkesh.com/p/ilya-sutskever-2

New comment by imenani in "A 0-click exploit chain for the Pixel 10"

imenani — Fri, 15 May 2026 14:42:53 +0000

https://lwn.net/Articles/1065620/

New comment by imenani in "New arXiv policy: 1-year ban for hallucinated references"

imenani — Thu, 14 May 2026 20:59:08 +0000

https://xcancel.com/tdietterich/status/2055000956144935055

New comment by imenani in "Rewrite Bun in Rust has been merged"

imenani — Thu, 14 May 2026 12:59:26 +0000

The author discussed this here four days ago

https://news.ycombinator.com/item?id=48077663

New comment by imenani in "Significant raise of reports"

imenani — Thu, 09 Apr 2026 16:56:04 +0000

With the benefit of hindsight, perhaps much of this was Claude Mythos? The model was deployed internally since Feb

New comment by imenani in "The risk of AI isn't making us lazy, but making "lazy" look productive"

imenani — Sat, 28 Mar 2026 16:15:08 +0000

Agreed. LLMs have helped me achieve much deeper reading, _when directed to do so_. Asking an LLM to “Teach me Socratically about this paper/code. One question at a time”, usually allows me to get a much deeper reading of the material than I would otherwise.

New comment by imenani in "LLMs aren't world models"

imenani — Sun, 10 Aug 2025 22:00:07 +0000

Each of these models has a thinking/reasoning variant and a default non-thinking variant. I would expect the reasoning variants (o3 or “GPT5 Thinking”, Gemini DeepThink, Claude with Extended Thinking, etc) to do better at this. I think there is also some chance that in their reasoning traces they may display something you might see as closer to world modelling. In particular, you might find them explicitly tracking positions of pieces and checking validity.

New comment by imenani in "LLMs aren't world models"

imenani — Sun, 10 Aug 2025 20:17:12 +0000

As far as I can tell they don’t say which LLM they used which is kind of a shame as there is a huge range of capabilities even in newly released LLMs (e.g. reasoning vs not).

New comment by imenani in "Does RL Incentivize Reasoning in LLMs Beyond the Base Model?"

imenani — Wed, 23 Apr 2025 19:24:20 +0000

They fix the temperature at T=0.6 for all k for all models, even though their own Figure 10 shows that RL model benefits from higher temperatures. I would buy the overall claim much more if they swept of temperature parameter for each k and model like they did in the Codex paper [1].

[1] https://arxiv.org/abs/2107.03374