Hacker News: darknoon

New comment by darknoon in "Rewrite Bun in Rust has been merged"

darknoon — Thu, 14 May 2026 18:50:00 +0000

the `claude` binary is essentially a packed copy of bun + the js code, so this will replace the native runtime part of claude code.

New comment by darknoon in "You can beat the binary search"

darknoon — Thu, 30 Apr 2026 18:45:30 +0000

This relies on knowledge of the distribution, just querying in the middle of A = [1, 2, 4, 8, 16, ..., 2^(n-1)] is slower than binary search

New comment by darknoon in "SVG Path Editor"

darknoon — Wed, 28 Jan 2026 16:16:30 +0000

I haven't seen one that worked properly—can you list a couple examples? Some of the ones that say they're "AI" are just VTracer / Potrace and don't give nice control points.

New comment by darknoon in "ChatGPT Containers can now run bash, pip/npm install packages and download files"

darknoon — Tue, 27 Jan 2026 13:54:37 +0000

I think you'd find that it's far from "any human" who can do this without looking anything up. I have 15y of dev exp and couldn't do this from memory on the cli. Maybe in c, but less helpful to getting stuff done!

New comment by darknoon in "Furiosa: 3.5x efficiency over H100s"

darknoon — Thu, 15 Jan 2026 01:16:51 +0000

really weird graph where they're comparing to 3x H100 PCI-E which is a config I don't think anyone is using.

they're trying to compare at iso-power? I just want to see their box vs a box of 8 h100s b/c that's what people would buy instead, and they can divide tokens and watts if that's the pitch.

New comment by darknoon in "Sacrificing accessibility for not getting web scraped"

darknoon — Sun, 14 Dec 2025 18:24:05 +0000

Here's the problem, you're still going to get scraped and the LLM will understand it anyway. Maybe at best you'll get filtered out of the dataset b/c it's high perplexity text?

New comment by darknoon in "Helion: A high-level DSL for performant and portable ML kernels"

darknoon — Fri, 07 Nov 2025 16:05:11 +0000

The developers also gave a talk about Helion on GPU Mode: https://www.youtube.com/watch?v=1zKvCLuvUYc

New comment by darknoon in "Intel Foundry demonstrates first Arm-based chip on 18a node"

darknoon — Wed, 20 Aug 2025 14:51:16 +0000

Here's the thing, they've completely given up and started making their (inferior to AMD) CPUs on TSMC. For example, Arrow Lake is on TSMC N3B. So it's not getting amortized over anything at all and their valuation is going to 0.

New comment by darknoon in "GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models [pdf]"

darknoon — Tue, 12 Aug 2025 02:55:07 +0000

It's ok, somewhere between a qwen 2.5 VL and the frontier models (o3 / opus 4) on visual reasoning

New comment by darknoon in "Show HN: TokenDagger – A tokenizer faster than OpenAI's Tiktoken"

darknoon — Mon, 30 Jun 2025 15:54:19 +0000

In ML, often it does work to a degree even if it's not 100% correct. So getting it working at all is all about hacking b/c most ideas are bad and don't work. Then you'll find wins by incrementally correcting issues with the math / data / floating point precision / etc.

New comment by darknoon in "Cap: Lightweight, modern open-source CAPTCHA alternative using proof-of-work"

darknoon — Fri, 30 May 2025 19:32:17 +0000

If you were doing a lot of scraping, you could just solve this on a GPU in 1/10 or less of the time it takes a human's phone to do it. Generally you need a decent computer to render a webpage while scraping it these days, so I don't see what this is solving.

New comment by darknoon in "Physics of Language Models: Architecture Design and the Magic of Canon Layers"

darknoon — Thu, 15 May 2025 00:19:34 +0000

anyone know why they mix in the 3 previous tokens? could have just as easily done 5 or 2 right?

New comment by darknoon in "Show HN: NCompass Technologies – yet another AI Inference API, but hear us out"

darknoon — Mon, 16 Dec 2024 22:35:57 +0000

One vote for image inputs here. I would love a fine-tuned qwen-2-vl-72b on demand, but most of the solutions are "talk to us" level expensive. I'm assuming you beat the price or convenience of a replicate / modal solution?

New comment by darknoon in "OpenAI, Google and Anthropic are struggling to build more advanced AI"

darknoon — Thu, 14 Nov 2024 21:59:30 +0000

I think just reading the code wouldn't make you a good programmer, you'd need to "read" the anti-code, ie what doesn't work, by trial and error. Models overconfidence that their code will work often leads them to fail in practice.

New comment by darknoon in "Transfusion: Predict the next token and diffuse images with one multimodal model"

darknoon — Mon, 09 Sep 2024 22:17:00 +0000

this is somewhat similar, but diffusion transformers typically use a pre-trained text model as the text conditioning whereas, in this case it's integrated and trained together multimodally.

New comment by darknoon in "Serving AI from the Basement – 192GB of VRAM Setup"

darknoon — Mon, 09 Sep 2024 14:45:58 +0000

I tried this w/ AM5, but realized that despite there theoretically being enough lanes for dual x16 PCI-e 4.0 GPUs, I couldn't find any motherboards that are actually configured this way, since dual-GPU is dead in consumer for gaming.

New comment by darknoon in "Superconducting Microprocessors? Turns Out They're Ultra-Efficient (2021)"

darknoon — Wed, 31 Jul 2024 21:39:08 +0000

Relevant video about some of the history of superconducting computers: https://www.youtube.com/watch?v=14r2oMsAaE8

New comment by darknoon in "We created the first open source implementation of Meta's TestGen–LLM"

darknoon — Tue, 21 May 2024 12:15:06 +0000

Why does this webpage have auto-playing audio?

New comment by darknoon in "Boom announces successful flight of XB-1 demonstrator aircraft"

darknoon — Fri, 22 Mar 2024 21:20:00 +0000

No, it's connected to AirTrain, which is slow and unpredictable, which is then connected to either the A or the LIRR.

New comment by darknoon in "Mark Zuckerberg says Quest 3 'is the better product, period'"

darknoon — Wed, 14 Feb 2024 15:02:30 +0000

I think Vision Pro isn't really a product, it's for developers / early adopters to make apps that will then be available once a consumer version is ready.