Hacker News: catgary

New comment by catgary in "Now is the best time to be a duct tape engineer"

catgary — Thu, 04 Jun 2026 16:49:57 +0000

I think you should just focus on the road because most of us are just trying to get home safely to our families. Some of us are even biking beside the road on a lightly-protected bike lane.

New comment by catgary in "The Old Guard: Confronting America's Gerontocratic Crisis"

catgary — Thu, 07 May 2026 02:46:52 +0000

Yeah, because they aren’t in need of a social safety net.

New comment by catgary in "China blocks Meta's acquisition of AI startup Manus"

catgary — Mon, 27 Apr 2026 19:49:57 +0000

I think you need to give some concrete examples, considering the US happily let its companies offshore a lot of work to China over the years, and Chinese funds own large chunks of American companies.

New comment by catgary in "Taking on CUDA with ROCm: 'One Step After Another'"

catgary — Mon, 13 Apr 2026 02:32:38 +0000

I’ve found they’re quite good when you’re higher in the compiler stack, where it’s essentially a game of translating MLIR dialects.

New comment by catgary in "Some uncomfortable truths about AI coding agents"

catgary — Wed, 01 Apr 2026 14:33:53 +0000

Because it doesn’t? Thats why the field of formal methods exists.

New comment by catgary in "Some uncomfortable truths about AI coding agents"

catgary — Sun, 29 Mar 2026 23:28:33 +0000

Because the code is the actual thing, tests can only show that the code fails in certain cases, they don’t actually prove the code is correct.

New comment by catgary in "Some uncomfortable truths about AI coding agents"

catgary — Sat, 28 Mar 2026 21:19:43 +0000

Generally speaking, I try to ensure that the LLM is using core abstractions throughout the codebase in a consistent manner. This makes it easier for me to review any changes it makes.

New comment by catgary in "Some uncomfortable truths about AI coding agents"

catgary — Fri, 27 Mar 2026 21:22:53 +0000

And even then - I still read the code it generates, and if I see a better way of doing something I just step in, write a partial solution, and then sketch out how the complete solution should work.

New comment by catgary in "Telnyx package compromised on PyPI"

catgary — Fri, 27 Mar 2026 21:00:26 +0000

This line of thought is honestly a bit silly - uv is just a package manager that actually does its job for resolving dependencies. You’re talking about a completely orthogonal problem.

New comment by catgary in "Iran launched unsuccessful attack on UK's Diego Garcia"

catgary — Sat, 21 Mar 2026 19:56:00 +0000

I think there are 5-7 thousand confirmed deaths by the UN, and medical reports in Iran estimated there could be 20,000+ casualties.

New comment by catgary in "The Misuses of the University"

catgary — Thu, 26 Feb 2026 02:56:50 +0000

I see the value of the students, it just seems like an odd thing for a government to subsidize via NIH/NSF funding. We don’t really have anything analogous to that in Canada and it just seems awfully weird that it exists in the US without the “it’s older than the country” excuse that Oxford/Cambridge have.

New comment by catgary in "Adoption of EVs tied to real-world reductions in air pollution: study"

catgary — Sun, 25 Jan 2026 03:46:30 +0000

EVs should do much better on brake dust thanks to regenerative braking, no?

New comment by catgary in "New Kindle feature uses AI to answer questions about books"

catgary — Fri, 12 Dec 2025 21:37:17 +0000

Luckily, the LLM has the text to cite, it can be passed in at inference time, which is legally distinct from training on the data.

New comment by catgary in "New Kindle feature uses AI to answer questions about books"

catgary — Fri, 12 Dec 2025 21:08:10 +0000

Because text analysis is substantially easier than video analysis?

New comment by catgary in "New Kindle feature uses AI to answer questions about books"

catgary — Fri, 12 Dec 2025 20:55:04 +0000

Sure, in the sense that any belief about the law isn’t cut and dried until a judge has explicitly dismissed it in the court of law.

New comment by catgary in "New Kindle feature uses AI to answer questions about books"

catgary — Fri, 12 Dec 2025 20:47:21 +0000

Are you implying that an LLM needs to be trained on a specific piece of text to answer questions about it?

New comment by catgary in "New Kindle feature uses AI to answer questions about books"

catgary — Fri, 12 Dec 2025 20:41:04 +0000

You don’t need any rights to execute the feature. The user owns the book. The app lets the user feed the book into an LLM, as is absolutely their right, and asks questions.

New comment by catgary in "SIMA 2: An agent that plays, reasons, and learns with you in virtual 3D worlds"

catgary — Thu, 13 Nov 2025 20:54:20 +0000

I work on a much easier problem (physics-based character animation) after spending a few years in motion planning, and I haven’t really seen anything to suggest that the problem is going to be solved any time soon by collecting more data.

New comment by catgary in "When models manipulate manifolds: The geometry of a counting task"

catgary — Mon, 03 Nov 2025 14:27:16 +0000

I think this is an interesting direction, but I think that step 2 of this would be to formulate some conjectures about the geometry of other LLMs, or testable hypotheses about how information flows wrt character counting. Even checking some intermediate training weights of Haiku would be interesting, so they’d still be working off of the same architecture.

The biology metaphor they make is interesting, because I think a biologist would be the first to tell you that you need more than one datapoint.

New comment by catgary in "Roadmap for improving the type checker"

catgary — Fri, 31 Oct 2025 09:45:30 +0000

Did he get an actual type theorist for that part of the project?