Hacker News Reader

_zoltan_on Ask HN: How to deal with shipping AI-written code you don't understand?

so the workers understand cross worker topology? Do they always just see individual tiles? what if there is a dependency - eg I'm streaming data and want to compute something later on an earlier computation? How would that get distributed?

_zoltan_on Ask HN: How to deal with shipping AI-written code you don't understand?

sounds very cool - is this basically "C" support for cuTile? :)

_zoltan_on Intel D270 GPU AI Accelerator–Up to 480 GB LPDDR5X

In 2 years once it fails and Intel shuts it down completely you mean :)

_zoltan_on QGIS enacts moratorium on AI assisted contributions

That's not entirely true: my work might get zero traction because nobody will know about it.

_zoltan_on Passkeys are a trap

hmm, I'm sure with a model that is not restricted, one might be able to extract these locally.

_zoltan_on 'Jonathan' is the oldest land animal on Earth

Cold, no sun AND no alcohol? Man that must suck. :)

_zoltan_on Sharing AI progress in mathematics

As Terry Tao said, LLMs are not outsmarting us, they are out remembering us.

I'm fairly sure your understanding is not fully accurate.

_zoltan_on Agents don't need memory, they need documentation

But reading the doc if it's maintained llmwiki style is much cheaper than reading the code on a large codebase.

_zoltan_on Run Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s

vast.ai?

_zoltan_on The Forgetful CPU (Linux on M4)

reaching the goal is what matters. I never understood people who don't get satisfaction from getting there. I don't care that AI writes the code.

_zoltan_on Fakecloud: Local AWS cloud emulator for integration tests

why does it matter that it's a vibe coded website?

_zoltan_on AI coding has made CI a bottleneck, so we reworked ours to keep up

I actually look at the tests first. But for this to work, you have to have good hygiene. If the test makes sense, fails without the PR but passed with the PR, I'm almost happy.

_zoltan_on AX – Google’s Open Agentic Orchestrator

billions? who is running BILLIONS of agents?

tens, hundreds, maybe a couple thousand at a time? absolutely.

_zoltan_on How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip

I optimize a lot of CUDA and it got really, really good at it.

_zoltan_on PS5 Linux lead quits: "a bunch of noobs using LLMs" that "they don't understand"

you might be OK with calling everybody using LLMs a bunch of noobs. I disagree. this has nothing to do with my own usage.

_zoltan_on Building a Linux GPU Driver for the M4 Mac Mini in One Month

Anti LLM is not justified, it's crazy. They are digging their own grave.

_zoltan_on Pion, an agent designed to run any company autonomously

I don't need to go that far, living in Switzerland. I don't know why you'd assume I'm in the US, or somehow that I regard Kraft as great cheese.

_zoltan_on Pion, an agent designed to run any company autonomously

for your point a: writing code by hand puts you ahead of the people who don't write code by hand, eg, the people using LLMs to generate code? that doesn't make much sense assuming both of those personas are senior SWEs.

_zoltan_on Pion, an agent designed to run any company autonomously

buy another sub.

_zoltan_on Pion, an agent designed to run any company autonomously

Yes, it is a waste of time right now not to use AI to write code.

It's the same category of "artisan code" like "artisan cheese": you can do it if you have a ton of free time as a hobby but other than a hobby? Not worth it.

_zoltan_on Don't be the out of touch Kung Fu master

Eh, take your fav painter then.

_zoltan_on Don't be the out of touch Kung Fu master

that's a dumb comparison. are you equating writing code to painting like Michaelangelo?

_zoltan_on Don't be the out of touch Kung Fu master

coding for the sake of coding is, well, dumb.

_zoltan_on Don't be the out of touch Kung Fu master

> Focusing on data structures, architecture and algorithms is what competency in programming has looked like since forever

yes, being able to bang out leetcode hard from head on an interview on paper matters so much... oh wait. it does not.

in the real world, outside of interview questions, it hardly matters. when it comes there, I'll optimizite it, but getting the first system up and running in a way I want is way more fun & challenging. need a lock free ringbuffer? I'll look up what's the latest way to build it. or just ask my AI.

_zoltan_on Don't be the out of touch Kung Fu master

a lot of people don't understand that the joy of making something is not coding but actually being able to build anything I want. AI just made it 50x better, as exactly as you said, we don't need to deal with the useless junk (what's the API syntax? what's the library syntax? oh it changed in the last 3 versions, let me look up the new API, this and that...).

it's great.

_zoltan_on A misalignment of AI in mathematics

Welcome to the new world. I, for one, welcome our new math overlords. :)

_zoltan_on 118M Queries per Second on Neki

> The benchmark was very simple. A single-shard point select, one row fetched per-query by primary key. No writes, joins, or cross-shard queries. The workload that each shard receives is isolated, in that there are no single queries that span multiple shards.

I mean... What's the point of this "benchmark"?

_zoltan_on How the Chorleywood Bread Process transformed British bread

https://www.youtube.com/watch?v=n0U8RdRdFDU

do this. thank me later. (yes, it takes multiple days!). as good as when we hop over to France.

_zoltan_on Tell HN: OpenAI keeps re-enabling the 'allow training' setting

hahah :) sorry, typo... in hungarian both v and w are the same sound spoken so if I don't concentrate, easy to mix up.