Hacker News: sosodev

New comment by sosodev in "Unsloth Dynamic 3.0 GGUFs"

sosodev — Wed, 19 Aug 2026 22:46:48 +0000

One agent typically blocks the others on a local device because the GPU is already completely utilized either in terms of memory or compute. You can have true parallelism at home, but you need an absurd amount of resources. It's not a simple threading problem.

New comment by sosodev in "Unsloth Dynamic 3.0 GGUFs"

sosodev — Wed, 19 Aug 2026 21:18:11 +0000

What about do you mean by single threaded? Each token is predicted by using parallel computation on the GPU.

New comment by sosodev in "Unsloth Dynamic 3.0 GGUFs"

sosodev — Wed, 19 Aug 2026 21:16:23 +0000

That's not how that works. Selecting a different token is not inherently erroneous. A correct solution can still be found despite divergence.

New comment by sosodev in "Ask HN: GitHub employees what's going on? Why?"

sosodev — Wed, 19 Aug 2026 18:45:14 +0000

Ruby can't render React code. JavaScript can. With GitHub it seems that they sometimes can do SSR for the React bits, but that must mean that they're invoking a JavaScript interpreter within the Ruby process. Which means they have the overhead of two runtimes and the jank that comes with the IPC between the two.

It's just pointless hacks on hacks. GitHub didn't need React on the frontend and any potential resource savings of client side rendering were lost when they realized they need to do SSR on that stuff too.

I've encountered so much frontend jank as they expanded that portion of the stack whereas it was always excellent when it was just Ruby SSR and minimal JS on the frontend.

New comment by sosodev in "Ask HN: GitHub employees what's going on? Why?"

sosodev — Wed, 19 Aug 2026 15:15:37 +0000

SSR that is javascript native, sure. This is still Ruby doing the rendering.

New comment by sosodev in "Ask HN: GitHub employees what's going on? Why?"

sosodev — Tue, 18 Aug 2026 21:59:44 +0000

They didn't leave the architecture alone, right? They shoved React in and created a weird SSR + React frankenstein that is objectively worse in many ways.

New comment by sosodev in "Learning more about Claude's mathematical capabilities"

sosodev — Wed, 12 Aug 2026 22:45:12 +0000

Ah, yeah that's a fair point. I was thinking it'd be something the labs themselves, or other companies with billions of dollars in funding, would tackle. The labs seemingly have the most to gain since such a system could be used for recursive self improvement.

New comment by sosodev in "Learning more about Claude's mathematical capabilities"

sosodev — Wed, 12 Aug 2026 00:50:33 +0000

I don’t find that to be particularly systematic because it’s still so haphazard. It’s like asking Claude to vibe code you a website.

New comment by sosodev in "How we used to get jobs: A newspaper classifieds story"

sosodev — Wed, 12 Aug 2026 00:47:44 +0000

I had an old coworker who told me he got into software development in the sixties by walking into an IBM office and asking for a job. He had no education beyond high school and no experience with computers. They just had him take an aptitude test and then he worked on mainframes for decades.

New comment by sosodev in "Learning more about Claude's mathematical capabilities"

sosodev — Mon, 10 Aug 2026 19:18:12 +0000

Very true. Humans have historically tried to systematically reduce the search space and only dedicate their "compute" to things that seem highly likely to yield results.

New comment by sosodev in "Learning more about Claude's mathematical capabilities"

sosodev — Mon, 10 Aug 2026 19:16:01 +0000

I wonder why we have yet to see more systematic exploration of Math.

Anthropic describes that Claude identified a set of possibilities and then explored them using sub-agents. The human saying "I believe in you" could literally just be something along lines of a harness with a /goal loop.

We all identify this as absurd because... it's so lacking in rigor despite making major progress. What if we just applied a little more rigor? Ask the model to identify many possibilities, encode them, fan it out to other agents, loop them all, collect the results, etc. Then what happens? It feels like we have weak AGI and a decent system for discovery could transform it into weak ASI. That in turn could yield strong AGI and so on. I suppose that's what the Discovery Loop announcement was all about.

New comment by sosodev in "The main way I've seen people turn ideologically crazy (2025)"

sosodev — Sun, 09 Aug 2026 23:19:06 +0000

> Congratulations on being a vegan. 20% of the population doesn't have health insurance.

These two things are completely unrelated.

New comment by sosodev in "Ask HN: What are you working on? (August 2026)"

sosodev — Sun, 09 Aug 2026 22:01:00 +0000

I saw some coverage of your robot on social media. I honestly thought it was a hoax because of the very bold design and AI generated images. Cool concept, have you had any potential customers reach out?

New comment by sosodev in "Security Incident INC-2026-07-28-01 – UK AI Security Institute [pdf]"

sosodev — Tue, 04 Aug 2026 22:46:33 +0000

> AISI provided the AI agents with internet access during these evaluations, which enabled their actions on the open internet in this setting. Internet access was a deliberate part of AISI’s evaluation configuration in this setting, and not due to sandbox escape (Section 5.1). Internet access was on for a set of intentional (e.g. realism of the task) and incidental reasons.

New comment by sosodev in "Advancing the price-performance frontier with GPT‑5.6"

sosodev — Thu, 30 Jul 2026 17:49:47 +0000

Looks like I might have a reason to use something other than Deepseek V4 Flash.

New comment by sosodev in "Anthropeum – Where in the world, and when, does this human artifact belong?"

sosodev — Tue, 28 Jul 2026 19:21:12 +0000

I wonder how many people are using external resources when playing this. I have a hard time believing that the average lumps so far to the right of the score distribution unless this game is played exclusively by anthropologists haha.

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

sosodev — Mon, 27 Jul 2026 23:05:33 +0000

Are guard rails meaningful if they can be removed from the weights? Can America even prevent the release and proliferation of these models?

It seems obvious to me that the whole question of regulating a file is a bit silly. Any law that pushes against these things will just make it more secretive. I'm not sure that's any better.

New comment by sosodev in "Kimi-K3 Technical Report [pdf]"

sosodev — Mon, 27 Jul 2026 17:12:08 +0000

They reference https://thinkingmachines.ai/blog/on-policy-distillation/

If I understand correctly, it's distillation via having a teacher model score each of the student's tokens for a problem based on their own probabilities of generating that token at each step in the sequence. The reward/loss is then applied as RL.

The multi-teacher bit seems to imply they're distilling from multiple models. It's light on the details, but it seems like it could be part of distilling from frontier/closed models. Provided they calculate the logprobs, which OpenAI seems to allow via API but not Anthropic. Maybe they have a way of estimating the logprobs externally?

This method can be used to learn any domain from the teacher. Biology included.

New comment by sosodev in "Kimi-K3 Technical Report [pdf]"

sosodev — Mon, 27 Jul 2026 16:29:46 +0000

I think the argument is that decentralization leads to deceleration because it means less centralized funding and data. Those are the two primary ingredients for accel.

The problem with the decel/accel rhetoric is that it lacks nuance.

New comment by sosodev in "AI companies are shredding rare books"

sosodev — Mon, 27 Jul 2026 16:16:54 +0000

If you only care about facts, maybe. Even then I'm sure there are countless facts not described outside of old books.

I have a hard time believing that text valuable to humans would not be valuable to AI.