Hacker News: airstrike

New comment by airstrike in "Patterns and problems in emerging multi-agent systems"

airstrike — Sun, 16 Aug 2026 12:31:03 +0000

I agree at a high level, but this then poses an even harder problem: choosing not to learn from something.

Humans, for example, can hear some advice, judge it to be unhelpful and dismiss it. LLMs can't learn let alone choose not to

New comment by airstrike in "Patterns and problems in emerging multi-agent systems"

airstrike — Sun, 16 Aug 2026 12:27:15 +0000

Comments should get more thoughtful and substantive, not less, as a topic gets more divisive

https://news.ycombinator.com/newsguidelines.html

New comment by airstrike in "Brazilian election filter in X For You timeline"

airstrike — Sat, 15 Aug 2026 17:29:40 +0000

I'm not sure what you mean, so it's hard to agree someone would have laughed at it.

It is almost a perfect definition of sovereignty—to shape reality such that it fits what the state deems appropriate.

This by no means imply it's perfectly effective, but that's besides the point. We're debating what the state ought to do, not whether it succeeds.

New comment by airstrike in "Brazilian election filter in X For You timeline"

airstrike — Sat, 15 Aug 2026 15:16:53 +0000

They didn't ban recommendations because they thought it was a "good idea".

They banned it to have reality comply with Brazilian electoral law which is written so that every candidate has the same amount of airtime.

It is not the job of the judiciary to determine whether a law should or shouldn't exist, but to make sure all laws are correctly enforced. Remembering this fact isn't optional.

Said differently, arguments about such decisions should revolve around whether the law says what the courts say it does or whether it was applied correctly—not whether the law should exist.

New comment by airstrike in "k-Coloring is Faster than Computing the Chromatic Number"

airstrike — Sun, 09 Aug 2026 15:24:30 +0000

And yet, if models could indeed reason, that would not be an issue. You're arguing my point.

New comment by airstrike in "k-Coloring is Faster than Computing the Chromatic Number"

airstrike — Sun, 09 Aug 2026 04:47:15 +0000

You're picking on the specifics of the carwash problem rather than acknowledging countless other similar problems exist. The original one was strawberry, but that got RL'd away

New comment by airstrike in "k-Coloring is Faster than Computing the Chromatic Number"

airstrike — Sat, 08 Aug 2026 14:55:15 +0000

The real problem is that human mistakes generally occur more frequently given the difficulty of the task, whereas LLM mistakes are somewhat random, like the carwash problem, because LLMs cannot truly reason.

I'd rather have a human on my team for whom I can reasonably surmise what tasks they're good at than have a robot who randomly gets shit wrong.

New comment by airstrike in "NASA figured out how to keep its Voyager 2 probe running for another year"

airstrike — Sat, 08 Aug 2026 12:08:19 +0000

This is literally triggering, and I don't say that lightly.

New comment by airstrike in "Learn how chips are made with this Rollercoaster Tycoon-inspired animation"

airstrike — Fri, 07 Aug 2026 03:36:39 +0000

Here's a video in case anyone else is also curious about what the crystal puller part looks like:

https://youtu.be/zVq8o9mgugc?si=YouOXIbJEEsDJS4z&t=135

New comment by airstrike in "Software development with AI is starting to feel like cooking steak"

airstrike — Fri, 07 Aug 2026 01:24:45 +0000

Pretty hard to be criticizing someone severely if I'm not even speaking to them specifically, don't you think? You're making a mountain out of a mole hill.

I'm entitled to my opinion, which I self-described as a bad hot take.

Sounds like you're the one specifically reprimanding me here? Oops...

New comment by airstrike in "Software development with AI is starting to feel like cooking steak"

airstrike — Fri, 07 Aug 2026 00:01:04 +0000

You should really read the dictionary definition of the word "chastise".

https://ahdictionary.com/word/search.html?q=chastise

New comment by airstrike in "Software development with AI is starting to feel like cooking steak"

airstrike — Thu, 06 Aug 2026 22:30:19 +0000

I'm not chastising anyone, I have no power over them

New comment by airstrike in "Branchless Rust: Making a Filter 4x Faster by Removing an If"

airstrike — Thu, 06 Aug 2026 19:56:16 +0000

The annoyance is real, and honestly? It's worth sitting with it.

New comment by airstrike in "Pareto Front"

airstrike — Thu, 06 Aug 2026 19:22:05 +0000

Ok, I'll make it!

New comment by airstrike in "Software development with AI is starting to feel like cooking steak"

airstrike — Thu, 06 Aug 2026 18:21:47 +0000

I agree, it's cheating.

My bad hot take is if you're using a meat thermometer and reverse sear to cook steak, you're like a kid on training wheels.

New comment by airstrike in "Pareto Front"

airstrike — Thu, 06 Aug 2026 16:41:54 +0000

Yeah, those features are awesome! Sadly support for them is quite lacking in applications...

Tabular numbers are awesome!

New comment by airstrike in "Pareto Front"

airstrike — Thu, 06 Aug 2026 12:05:06 +0000

Inter with optional open type features turned on?

https://rsms.me/inter/

New comment by airstrike in "When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation"

airstrike — Tue, 04 Aug 2026 23:09:33 +0000

It turns out being orders of magnitude faster at searching with the aid of a strong verifier is a great way to generate proofs

New comment by airstrike in "When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation"

airstrike — Tue, 04 Aug 2026 23:06:55 +0000

Shallow dismissals like this one are against HN guidelines because they make for very poor discourse.

We all learned more from the prior comment than we did from this one.

New comment by airstrike in "Waymo in Dallas"

airstrike — Tue, 04 Aug 2026 20:02:39 +0000

I love Waymo. The only problem I have with it is that it further siphons wages from local economies into Silicon Valley, beyond what Uber/Lyft already did.