Hacker News: sudb

New comment by sudb in "Desert Ant Labs: local, fast models that run on device"

sudb — Wed, 09 Sep 2026 14:49:45 +0000

It is now also reasonably straightforward if you have access to frontier LLMs, a recent-ish mac and a recent-ish iPhone to point them at the job of porting a given model to run on the ANE - it's a reasonably easy task to hill-climb at this point!

New comment by sudb in "Desert Ant Labs: local, fast models that run on device"

sudb — Wed, 09 Sep 2026 13:16:52 +0000

The underlying model (DFNet3) is not particularly great but it is very small and fast - imo the best commercially usable denoising model is MossFormer2 (no affiliation - it's just excellent) with one drawback in that it can't remove reverb.

Nvidia's RE-USE model can do what MossFormer2 does _and_ can remove reverb, but it is non-commercial licensed.

New comment by sudb in "qm – Multiplayer agent harness for work"

sudb — Sat, 01 Aug 2026 00:18:52 +0000

This was literally in YC's Request For Startups for Fall 2026: https://www.ycombinator.com/rfs#multiplayer-ai

New comment by sudb in "Everyone is building LLM routers, we deprecated ours"

sudb — Sat, 01 Aug 2026 00:17:32 +0000

To add: I doubt frontier labs will build routers - they are not financially incentivized to optimize token usage (though in the short term they may be incentivized by constrained GPU capacity to reduce load).

New comment by sudb in "Situational Awareness down 67% in July in AI stock rout"

sudb — Fri, 31 Jul 2026 23:40:17 +0000

Obviously not, no need to patronize. Did you read it? It seems to be honest about the numbers but it's pretty skillfully written to reassure.

New comment by sudb in "Situational Awareness down 67% in July in AI stock rout"

sudb — Fri, 31 Jul 2026 17:12:15 +0000

I found a tweet purporting to show the letter that Leopold sent to his LPs - it looks pretty thoughtful and doesn't sound as bad as some news sources seem to be portraying.

https://x.com/tbpn/status/2083226453509030285

New comment by sudb in "The most official water costs $120k a gallon"

sudb — Fri, 31 Jul 2026 16:56:15 +0000

Reminds of me of the "most expensive peanut butter" at ~$2.44/g - sold by NIST, I assume for some sort of peanut butter factory process calibration: https://shop.nist.gov/ccrz__ProductDetails?sku=2387

New comment by sudb in "Banning AI will not make it go away"

sudb — Wed, 29 Jul 2026 00:14:28 +0000

Absolutely agreed - though the usual claim for coding agents in particular is that they are already very capable and fast, replacing and speeding up software development hugely (anecdotally reasonably true for me and my colleagues, though again depends on domain and complexity).

New comment by sudb in "Banning AI will not make it go away"

sudb — Tue, 28 Jul 2026 23:22:43 +0000

I was going to say that this surely depends on the type of work being done with AI but then a small dive into your HN profile shows you're working on SWE agents at Google?

I am struggling to reconcile your viewpoint with this but may have found the wrong profile!

New comment by sudb in "Show HN: FeyNoBg – Automatic background removal model and training library"

sudb — Mon, 27 Jul 2026 18:45:25 +0000

Of course, would be happy to!

New comment by sudb in "Google's own system admits the back end never deletes – video proof included"

sudb — Mon, 27 Jul 2026 18:33:24 +0000

I don't know that you can trust Google's AI Overview to have accurate knowledge about Google AI Studio internals

New comment by sudb in "Show HN: FeyNoBg – Automatic background removal model and training library"

sudb — Mon, 27 Jul 2026 18:31:22 +0000

very cool! I've been having a fun time porting non-LLM models to iOS/Android (recently had some success with making Bonsai's image generation model run on Android) - have you attempted to run any of the workflows on a mobile device at all? Mind if I try?

New comment by sudb in "Canceling "Hey""

sudb — Mon, 27 Jul 2026 18:27:57 +0000

I think some of the trouble here is that the application of the law cannot be guaranteed to be free of discrimination or bias, regardless of the letter of the law itself?

New comment by sudb in "Canceling "Hey""

sudb — Mon, 27 Jul 2026 18:25:29 +0000

I am a native speaker, I agree with this assessment.

New comment by sudb in "Canceling "Hey""

sudb — Mon, 27 Jul 2026 18:22:22 +0000

You only need to find an email platform run by people you agree with more, no? I'd vouch for proton mail here.

New comment by sudb in "I wanted a clock that never needed setting. Things escalated"

sudb — Mon, 27 Jul 2026 14:05:07 +0000

I mentioned this specific model because it also has Tough Solar which should mean much longer between battery replacements!

New comment by sudb in "Teaching Kids Forth"

sudb — Mon, 27 Jul 2026 00:36:40 +0000

I doubly enjoyed reading this because I feel like I'm reasonably sure the author never ran this through an LLM because of the "cirriculum" typo.

New comment by sudb in "I wanted a clock that never needed setting. Things escalated"

sudb — Mon, 27 Jul 2026 00:32:54 +0000

This reminded of this other clock project: https://mitxela.com/projects/precision_clock_mk_iv

Also: I reckon the closest best-value watch alternative that ticks most(?) of the author's boxes is the Casio GW-M5610-1ER, so long as you are inside a radio-controlled time area.

New comment by sudb in "Inflect-Micro-v2: complete voice in 9.36M parameters"

sudb — Sun, 26 Jul 2026 16:39:13 +0000

this is extremely encouraging for individuals/small companies being able to train pareto-frontier TTS models (specifically compute required to run vs quality of model output)

New comment by sudb in "Agent swarms and the new model economics"

sudb — Tue, 21 Jul 2026 01:40:45 +0000

I don't think Gas Town was actually all that much ahead of the curve - see Cursor's swarm post from January this year, mentioned at the very top of this blog post.

I also don't know that beads is a very strong example that Yegge knows what he's doing - it's famously been described as (pseudo)malware.