Hacker News: Philpax

FelonyBench

Philpax — Wed, 05 Aug 2026 07:47:24 +0000

Article URL: https://www.felonybench.com/

Comments URL: https://news.ycombinator.com/item?id=49179778

Points: 1

# Comments: 1

Incident Report: unsanctioned agent behaviour during cyber testing

Philpax — Tue, 04 Aug 2026 21:12:47 +0000

Article URL: https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing

Comments URL: https://news.ycombinator.com/item?id=49175233

Points: 17

# Comments: 2

New comment by Philpax in "Waymo in Dallas"

Philpax — Tue, 04 Aug 2026 20:49:33 +0000

Are you running your replies through a LLM, or is the LLM producing them entirely?

Kiro Crew

Philpax — Tue, 04 Aug 2026 17:40:46 +0000

Article URL: https://github.com/kirodotdev/KiroCrew

Comments URL: https://news.ycombinator.com/item?id=49172209

Points: 8

# Comments: 0

LFM2.5-2.6B: Deploy Agents Everywhere

Philpax — Tue, 04 Aug 2026 14:27:42 +0000

Article URL: https://www.liquid.ai/blog/lfm2-5-2-6b

Comments URL: https://news.ycombinator.com/item?id=49169497

Points: 3

# Comments: 0

New comment by Philpax in "Show HN: Fuse – statically typed functional programming language"

Philpax — Mon, 03 Aug 2026 20:31:44 +0000

I've only used it for small scripts and can't say how well the type system actually functions/how it scales, but what I've seen so far has pleased me. (They have a _comptime_-equivalent! https://luau.org/types/type-functions/)

You may be interested in https://lute.luau.org/, which is a node.js-style runtime for the language.

New comment by Philpax in "Ten advances in mathematics and theoretical computer science"

Philpax — Mon, 03 Aug 2026 20:27:17 +0000

4. I think they can, especially if the problem statement is well-specified and, importantly, autonomously testable. Of course, specifying a problem that meets these requirements is non-trivial, but the claim requests _a_ counterexample :P

New comment by Philpax in "Show HN: Fuse – statically typed functional programming language"

Philpax — Mon, 03 Aug 2026 05:34:58 +0000

Luau is statically typed and interpreted/JIT'd: https://luau.org/

New comment by Philpax in "Anime User Interfaces"

Philpax — Sun, 02 Aug 2026 02:53:04 +0000

Copeland! I was quite surprised to see that myself.

New comment by Philpax in "Get ready to flee, Americans in ten countries warned"

Philpax — Sat, 01 Aug 2026 20:46:09 +0000

You may not want to engage with politics, but politics will engage with you.

New comment by Philpax in "DeepSeek-V4-Flash Update"

Philpax — Fri, 31 Jul 2026 12:39:17 +0000

That will cease to be a problem in the next 24 hours, now that the weights are out: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731

DeepSeek-AI/DeepSeek-V4-Flash-0731

Philpax — Fri, 31 Jul 2026 12:16:08 +0000

Article URL: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731

Comments URL: https://news.ycombinator.com/item?id=49122153

Points: 16

# Comments: 0

thinkingmachines/Inkling-Small

Philpax — Thu, 30 Jul 2026 17:34:06 +0000

Article URL: https://huggingface.co/thinkingmachines/Inkling-Small

Comments URL: https://news.ycombinator.com/item?id=49113099

Points: 8

# Comments: 1

New comment by Philpax in "Our position on open-weights models"

Philpax — Mon, 27 Jul 2026 22:41:30 +0000

This is where I'm at, or rather, will be. I like open-weight models and I've done my part in facilitating them, but if you believe in the increasing capability of these systems - and I do, to some measured extent - it seems plausible to me that an incident will happen at some point in the future.

I don't agree with his argument as a whole, especially not on some of the specifics (it is not great that this technology is being developed under the current US government), but I am sympathetic to the idea that some bells can't be unrung, and thus we should proceed with caution.

Luau Type Functions

Philpax — Wed, 22 Jul 2026 02:25:05 +0000

Article URL: https://luau.org/types/type-functions/

Comments URL: https://news.ycombinator.com/item?id=49001089

Points: 1

# Comments: 0

New comment by Philpax in "OpenAI and Hugging Face address security incident during model evaluation"

Philpax — Tue, 21 Jul 2026 22:45:48 +0000

You... haven't shown any evidence that we're near collapse. That's what I'm asking you for. Show me some evidence that we are losing capabilities with we have today.

New comment by Philpax in "OpenAI and Hugging Face address security incident during model evaluation"

Philpax — Tue, 21 Jul 2026 22:23:28 +0000

Assumes facts not in evidence. Please show your working.

New comment by Philpax in "OpenAI and Hugging Face address security incident during model evaluation"

Philpax — Tue, 21 Jul 2026 21:57:44 +0000

Over the last two years, this weight class has doubled its scores and/or saturated several benchmarks in the Qwen lineup alone without loss of generality: https://claude.ai/public/artifacts/9f249169-3623-417e-86cd-7...

There is undoubtedly a limit somewhere (there is only so much you can pack into a given size) but it's really not particularly clear where that limit is. I don't think it's superintelligence - that much I agree with you - but I think "We already have a 1gb model that is as capable as it will ever be" is strictly false.

New comment by Philpax in "OpenAI and Hugging Face address security incident during model evaluation"

Philpax — Tue, 21 Jul 2026 21:07:23 +0000

Please source this claim. What 1GB models are capable of has increased generation-on-generation.

> For example: you can't make a mice-sized brain as smart as a human brain no matter how hard you try.

Sure. We don't know where the ceiling is for our digital minds, though.

New comment by Philpax in "OpenAI and Hugging Face address security incident during model evaluation"

Philpax — Tue, 21 Jul 2026 20:57:28 +0000

> This is science fiction, these models don't have access to their own weights.

The models are being used to train, and improve the infrastructure for training, other models [0][1]. Several RL techniques rely on using the currently-being-trained weights as part of their process. I really would not take "don't have access" as a given, especially during the training phase.

> What would be a lot more scary is a model as capable as sol that's able to run on consumer hardware without taking up several terabytes of storage, but of course that is simply not possible as we need 4t parameters to even begin emulating a small fraction of what a human brain can do.

The Poolside Laguna S 2.1 model [2] purports to compete with models several times its size, and inference compute is becoming increasingly plentiful. Again, would not hold anything here as a given.

[0]: https://openai.com/index/gpt-5-6/ ("GPT-5.6 accelerates OpenAI")

[1]: https://www.kimi.com/blog/kimi-k3#coding

[2]: https://poolside.ai/blog/introducing-laguna-s-2-1