Recent submissions

Comment · 15 days ago
so, can I program it over an API ?

Can I put it under grokbot's control?

Comment · 1 month ago
Okay, AI now can write a poem or compose a song.

But can an AI construct Terminal Bench 5.0 or GDPval 2027 ?

Without this, there is just no path towards RSI.

Comment · 2 months ago
easy! google "archwiki dd" scroll to 2.2
Comment · 2 months ago
will it run on 2x 5060Ti with 16GB each?
Comment · 2 months ago
Do they train on their own data?

I mean, when Meta's engineer is creating some new DINOv4 or Segment Anything, with all the scaffolding around it, do they train on that?

Comment · 3 months ago
I tried to use Tailwind in a hobby project couple years ago. I havn't updated my CSS skills since 2010 or 2015 -- I certainly missed grid layout and some other new features.

Quickly, I realized Tailwind is documented on top on CSS, no way around. This makes you keep in mind both.

No idea why is Tailwind even a thing.

Comment · 3 months ago
What harness do they use for testing?

Back in 2025 it was common to test models in a different harnesses.

I remember watching a guy on youtube, who was testing every new model in opencode, cline, codex, claude, etc.

Why did it come out of fashion ?

EDIT: ah, yeah. the point was that a harness would often affect results (task completion rate, I think) for more than 10%

Comment · 3 months ago
note that ARC-AGI-3 has its rules changed.

before today all the contestants were capped at $10k

Comment · 3 months ago
> Model API is not available in your region.

:(

Well, Vietnam is not in the list of restricted territories.

Anyway, what is "your region" ?

Is this where I am now, or is it where I activated my Oculus 2 five years ago ?

Comment · 6 months ago
This billing cycle I was billed $20 three times.

I contacted my bank and got a reply (from a human) that all three payments are valid.

Emails from Anthropic state that the first two payments failed, but the third went through. Fin says that my question will be elevated to a human being, but so far I was not contacted.

Comment · 6 months ago
This makes a good benchmark LLMs:

``` look at this paper: https://arxiv.org/pdf/2603.21852

now please produce 2x+y as a composition on EMLs ```

Opus(paid) - claimed that "2" is circular. Once I told it that ChatGPT have already done this, finished successfully.

ChatGPT(free) - did it from the first try.

Grok - produced estimation of the depth of the formula.

Gemini - success

Deepseek - Assumed some pre-existing knowledge on what EML is. Unable to fetch the pdf from the link, unable to consume pdf from "Attach file"

Kimi - produced long output, stopped and asked to upgrade

GLM - looks ok

Comment · 8 months ago
spatial or not, arc-agi is the only test that correlates to my impression with my coding requests
Comment · 8 months ago
why don't they publish at ARC-AGI ? too expensive?
Comment · 10 months ago
I've created Asterisk Codex Skill, but turns out there is ten seconds timeout for scripts

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton