Hacker News Reader

eyegoron Qualcomm Cloud AI 100 Ultra – 150W AI Accelerator Card with 128GB at 548GB/s

600w is way off the vf curve though... Tiny bit of tuning and you'll get ~80-90% of the performance at ~150w nominal. At least that's what I got out of some pro 6000s when I spent 15 mins to underclock it.

eyegoron Qualcomm Cloud AI 100 Ultra – 150W AI Accelerator Card with 128GB at 548GB/s

Holy cow the v100s go for like $1k on ebay now... They were selling for $100 or less a couple years ago.

eyegoron Anthropic's internal AI use, at retail prices, now exceeds its employee pay

I'm impressed you have a workflow that can keep 16 separate agents fed without the whole thing collapsing into nonsense. Even running full agent teams I can't think of ways to do that without being intentionally inefficient

eyegoron MiniMax M3.1 Flash

Any chance you could try qwen3.8 flash next? That is currently the best local model for actual coding tasks in my + many coworkers experience. It will run reasonably fast if you have 64+ gb ram and use strata or a strata fork (vram not super important with this method).

https://github.com/Niko1221/Strata

eyegoron LFM2.5-24B-A2B: Built for fast on-device agents

Oh you should be running flash next with strata. Just go to stratas github page [0] and it will explain everything. It runs quite fast on 16+64, you should be able to get 50 tok/s or so. It's a 180b model and it feels a lot smarter than most of the other open weight options. As far as knowing "what will fit" in general that's hard to know these days. The moe models are hard to predict and run wildly different in different engines (llama.cpp, vllm, strata likes, etc) but they tend to be the only good options around 16gb vram. Dense models like qwen3.8 27b are only good if they're entirely in vram. Moe models you need the active count + a few experts (although it will run faster the more experts you can fit).

[0] https://github.com/Niko1221/Strata

eyegoron LFM2.5-24B-A2B: Built for fast on-device agents

Don't waste time on bonsai unless you have like 16gb or less ram (ram, not vram). Although I'm pretty sure you could still run this model well on 16+8, and qwen3.6 35b-a3b runs well on ~32gb combined ram at q4. Just offload most of the experts to cpu/ram if you have limited vram budget. With the moe models it doesn't hurt performance too much.

But bonsai is basically trash on real agentic coding work. I've tried both bonsai models, all the quants, messed with the hyperparameters, tried to fine tune (didn't succeed), and it just kept looping or failing to finish tasks. Qwen3.6 35b-a3b is my current go to for agentic work for any systems under 24gb vram, although honestly I'm not sure the dense qwen3.8 27b is meaningfully better. The main difference I see is it does a lot more testing attempts and burns a ton of tokens + compactions, so every task it takes 1.2-2x as long as 35b-a3b does. But the quality of the results is comparable.

eyegoron Gemini 4 Argon

I'm surprised no one's complaining about it randomly switching to mandarin. It seems like no matter what I do once every few days it'll go back to mandarin mid thinking trace and then stays in mandarin until I remind it to use English.

eyegoron OLED burn-in test: 30-month update

This one isn't really a "burnout test" though, that's just what they call it to get clicks. They're just using an oled monitor as their primary for daily work. This represents a pretty normal use case, it isn't a torture test like what RTings does. I don't see a world where I'd buy a monitor and not use it for static content the majority of the time.

eyegoron Backblaze drive stats for Q2 2026

Spinning rust is still the only cost effective answer to redundant 10+ tb storage. If you have 4+ disks, use raid 6 or raid 10 and you get decent fault tolerance too. Raid 10 also increases write performance but it's a bit less robust imo and more expensive per gb.

As for speed it's mainly tied to rpm and the market decided no one wants higher failure rates / lower mtbf for marginal speed gains.

If you need speed and capacity, throw an ssd or two in front of a raid disk array for caching with primocache or whatever.

eyegoron Goose:experimental lang 1.16x faster than C++ and 1.12x than safe Rust, mem safe

Might as well go back to ye olden days and put 100% of your memory in a giant preallocated block and instead of stack locals you just use the block. No allocation cost at runtime, cheese benchmarks by making the super arena.

eyegoron GitHub RCE Vulnerability: CVE-2026-3854 Breakdown

On prem beats the heck out of github post Microsoft though... At least you know how to get it working again when someone breaks it. These days with github you expect a weekly 500, a rainbow unicorn error, build failures due to unavailable errors, etc. Last I checked the third party tracker github services were barely pushing one 9 of reliability.

eyegoron OrangePi 6 Plus Review

"Chinese repos" is a very charitable interpretation of the Google drive links they used to distribute the os. It seemed like it was on the free plan too, it often didn't work because it tripped the maximum downloads per month limit.

eyegoron Spotlight on pdfly, the Swiss Army knife for PDF files

As nice as it looks, I have a lot of trouble believing the "we have magic money, it's free because that's good for business" logic.

    PDFgear is free of charge, and we don’t generate income through any hidden means. We Do NOT misuse or sell user data and we Do Not display ads. Here’s how we keep operations running: 
    We’ve secured investment to cover operational costs, including team expenses and technology like the ChatGPT API. We’re also experienced in optimizing technology usage to manage costs more effectively.

eyegoron OpenZL: An open source format-aware compression framework

Plans for language bindings? Should be trivial to whip up simpler ones like python or dotnet but I didn't see any official bindings yet.

eyegoron Zig builds are getting faster

> even a 1 minute compile time is dwarfed by the time it takes to write and reason about code, run tests, work with version control, etc.

You are far from the embedded world if you think 1 minute here or there is long. I have been involved with many projects that take hours to build, usually caused by hardware generation (fpga hdl builds) or poor cross compiling support (custom/complex toolchain requirements). These days I can keep most of the custom shenanigans in the 1hr ballpark by throwing more compute at a very heavy emulator (to fully emulate the architecture) but that's still pretty painful. One day I'll find a way to use the zig toolchain for cross compiles but it gets thrown off by some of the c macro or custom resource embedding nonsense.

Edit: missed some context on lazy first read so ignore the snark above.

eyegoron The Deletion of Docker.io/Bitnami

I've used them as a quick way to get rootless configured base images. Not sure if official repos provide those now, but it used to be a big hassle to get things like postgres images running without root in their containers. Although I often had to read through their dockerfiles to figure out the uid setup, where configs live, etc because they were not consistent between the various bitnami images.

eyegoron The Best Line Length

If you really want the 80 col experience, I think that was the fortran 77 default. You'd even get compiler errors if you try to exceed it. Of course there were flags to increase the line length limit but don't tell your colleagues.

eyegoron .NET 10 Preview 6 brings JIT improvements, one-shot tool execution

F# has been a third class citizen for a long time... Last I heard the entire f# team was ~10 people. Pretty sure "find references" still doesn't work across c# and f# (if you call a c# method from f# or vise versa). That also means symbol renames don't work correctly.

eyegoron Opencode: AI coding agent, built for the terminal

Side note, if you're a lazygit fan, consider using gitui as an alternative. Feature wise they're pretty similar but gitui is much faster and I find it easier to use.

https://github.com/gitui-org/gitui

eyegoron uv: An extremely fast Python package and project manager, written in Rust

I wish setup.py was actually on the way out, but sadly, it's the only straightforward way to handle packages that use cython or interop. In these cases, libraries use setup.py to compile the dll/so/dylib at install time. Naturally this is a bit of nightmare fuel since installing gets arbitrary code execution privileges but there's no real standard for privileges in python package installs.

eyegoron Python can run Mojo now

I guess they abandoned the python superset idea? I followed them for a bit when they first publicly launched and they said "don't worry, we'll be a real python superset soon" and the biggest omission was no support for classes. A few years later, it looks to be missing the same set of python features but added a lot of their own custom language features.

eyegoron End of 10: Upgrade your old Windows 10 computer to Linux

Is this advice insane or am I missing something

> to fix your busted drive, just nuke the boot sector and send it

> bash

> dd if=/dev/zero of=/dev/xxx bs=512 count=1 conv=notrunc

eyegoron Procolored printer drivers contained malware

That costs money to maintain, even if it's just a few bucks a month. I've seen plenty of Chinese companies using mega/gdrive/etc just because it's free. I used to think it was just cheapness, but depending on the company it can be a huge hassle to set up recurring small bill items. At my current company for example, it's much easier to pay $5-10k once than pay $5/mo.

eyegoron The 12-bit rainbow palette

https://cran.r-project.org/web/packages/khroma/vignettes/tol...

The tol palettes are the best looking colorblind friendly palettes to me. Most of the others get complaints from non colorblind users about looking bad/desaturated.

eyegoron I genuinely don't understand why some people are still bullish about LLMs

It sounds like you're trying to use these llms as oracles, which is going to cause you a lot of frustration. I've found almost all of them now excel at imitating a junior dev or a drunk PhD student. For example the other day I was looking at acoustic sensor data and I ran it down the trail of "what are some ways to look for repeating patterns like xyz" and 10 minutes later I had a mostly working proof of concept for a 2nd order spectrogram that reasonably dealt with spectral leakage and a half working mel spectrum fingerprint idea. Those are all things I was thinking about myself, so I was able to guide it to a mostly working prototype in very little time. But doing it myself from zero would've taken at least a couple of hours.

But truthfully 90% of work related programming is not problem solving, it's implementing business logic. And dealing with poor, ever changing customer specs. Which an llm will not help with.

eyegoron Murder Mystery: GCC builds failing after sbuild refactoring

I'll just pitch in as someone who's worked with several 40+ year old codebases and 3+ year old perl, the perl code is significantly harder to maintain. The language lends itself to unreadable levels of terseness and abuse of regex. Unless you use perl every day, the amount of time trying to understand the syntax vs the code logic is skewed too heavily towards obscure syntax. Even the oldest fortran/c is much easier to work with.

Except maybe arithmetic gotos designed to minimize the number of functions. Those are straight evil and I'm glad no modern language supports them.

eyegoron Htmy – Async, pure-Python rendering engine

In the real world, for web things, people use django or fastapi. I'd suggest picking a project with lots of stackoverflow questions and poking around their docs to see which makes you the most comfortable. Personally I tend to favor litestar these days since it has good docs and issues don't sit around for years waiting on one dude to merge prs (fastapi) and it's a lot nicer than django (and I hate django docs).

Flask/quart are painful to work with due to horrible documentation in my experience, but they're popular too. Quart is just an async rewrite of flask by the same owners.

Litestar has a half baked comparison chart here: https://docs.litestar.dev/latest/

eyegoron Noise-Canceling Single-Layer Woven Silk and Cotton Fabric

For papers sometimes you need to hit a page count to make your sponsor/advisor/conference happy. I've been told "this is a great paper but can you pad it out to 12pgs?", maybe that happened here as well.

eyegoron Is Python That Slow?

Nodejs/js runtimes in general get a lot of development effort to make the runtimes fast from Google et al. It's the default web language so there's a ton of effort put into optimizing the runtime. Python on the other hand is mostly a hacker/data science language that interops well with c, so there's not much incentive to make the base runtime fast. The rare times a company cares about python interpreter speed, they've built their own runtime for python instead.

eyegoron A Student's Guide to Writing with ChatGPT

I don't honestly think anyone can remember bash array syntax if they take a 2 week break. It's the kind of arcane nonsense that LLMs are perfect for. The only downside is if the fancy autocomplete model messes it up, we're gonna be in bad shape when Steve retires cause half the internet will be an ouroboros of ai generated garbage.