Hacker News: maxignol

New comment by maxignol in "Kimi-K3 on HuggingFace"

maxignol — Mon, 27 Jul 2026 10:16:25 +0000

I dont quite understand why GGUF is better optimized. Are the performances better for the same amount of VRAM ?

New comment by maxignol in "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"

maxignol — Wed, 22 Jul 2026 15:51:44 +0000

> I’d suspect the harness to massively affect token use and optimisation

Yes they do according to databricks -> https://www.databricks.com/blog/benchmarking-coding-agents-d...

That's why I find comparing models on benchmarks only gives the tendency. We should be comparing model x harness to have accurate metrics.

New comment by maxignol in "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"

maxignol — Wed, 22 Jul 2026 15:44:27 +0000

What is your harness with every model ?

New comment by maxignol in "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"

maxignol — Wed, 22 Jul 2026 15:43:08 +0000

> my use cases stop aligning to swebench pro around 50% accuracy, and more closely align with DeepSWE.

What do you mean by that ? If the model is higher than 50% on swebench pro then it tends to drift from what you like it to do, like DeepSWE benchmarks ?

New comment by maxignol in "Qwen 3.8"

maxignol — Sun, 19 Jul 2026 10:36:32 +0000

Shouldn’t we fear they start doing only close source like most us labs once they catch up in market shares ?

New comment by maxignol in "Kimi K3: Open Frontier Intelligence"

maxignol — Fri, 17 Jul 2026 18:06:56 +0000

Well, I had not heard of RLM before, just read the paper, thank you for introducing me to your lazy version !

New comment by maxignol in "Kimi K3: Open Frontier Intelligence"

maxignol — Fri, 17 Jul 2026 09:12:03 +0000

I’ve never seen the Id approach before, that’s a good idea ! Though I was wondering how do you manage to keep costs low within the 7 agents ?

New comment by maxignol in "Inkling: Our Open-Weights Model"

maxignol — Thu, 16 Jul 2026 07:35:45 +0000

Giving the accuracy-token graph and not the accuracy-cost graph, thus we cannot easily compare costs with other models, is not a way to gain my trust

New comment by maxignol in "Show HN: Getting GLM 5.2 running on my slow computer"

maxignol — Fri, 10 Jul 2026 09:21:22 +0000

I’m truly impressed by your work ! I don’t know if this is planned for near future, but how about adding energy efficiency benchmarks ? Because running locally is a great feeling, but the electricity bill should not be forgotten

New comment by maxignol in "Re: I'm Begging You to Leave Your AI Note-Taker at Home"

maxignol — Tue, 07 Jul 2026 20:47:20 +0000

Well, had a meeting with a VC that literally brought his note taking 3rd person to the meeting, did not even say hello. Tbh, I still find it creepier than an AI transcriber.

New comment by maxignol in "GPT-5.5 Codex reasoning-token clustering may be leading to degraded performance"

maxignol — Sun, 05 Jul 2026 07:39:57 +0000

This seems really bad…

New comment by maxignol in "Mouse: Precision Editing Tools for AI Coding Agents"

maxignol — Sun, 05 Jul 2026 07:31:50 +0000

I guess the technology used here must be ground-breaking lol

New comment by maxignol in "Jamesob's guide to running SOTA LLMs locally"

maxignol — Fri, 03 Jul 2026 20:02:03 +0000

Did not seem to find how much tokens per second he achieved with this setup ?

New comment by maxignol in "Why Switzerland has 25 gbit internet and America doesn't"

maxignol — Fri, 03 Jul 2026 06:41:02 +0000

Thus the prices in switzerland are higher than anywhere else. I don’t know about you, but I’d have no use of 25gbits/s

New comment by maxignol in "HackerRank open sourced its ATS. My resume scored 90/100. Oh wait 74. No – 88"

maxignol — Mon, 29 Jun 2026 12:59:03 +0000

Lol next time I’ll just apply with 4 accounts and maybe get in once.

New comment by maxignol in "HackerRank open sourced its ATS. My resume scored 90/100. Oh wait 74. No – 88"

maxignol — Mon, 29 Jun 2026 07:23:42 +0000

Are many people using HackerRank ATS ?

New comment by maxignol in "GLM 5.2 beats Claude in our benchmarks"

maxignol — Mon, 29 Jun 2026 07:16:54 +0000

Have you tried opencode go ?

New comment by maxignol in "GLM 5.2 beats Claude in our benchmarks"

maxignol — Mon, 29 Jun 2026 07:15:25 +0000

Would you recommand some ressources about how multiple neural engines are used in data centers ?

New comment by maxignol in "Apple raises prices of MacBooks, iPads"

maxignol — Thu, 25 Jun 2026 17:52:03 +0000

I guess it was inevitable. Is it only RAM related ?

New comment by maxignol in "Show HN: I made Google Trends for Hacker News by indexing 18 years of comments"

maxignol — Thu, 25 Jun 2026 17:47:30 +0000

Funny one x) Though I ain’t sure if even more data is useful on hackernews