neomindryan
Joined 16 years ago
Karma
366
Submissions
47
Member since
Sep 1
Founder of https://neomindlabs.com
View on Hacker NewsRecent submissions
Comment · 9 days ago
As AI makes sophisticated cyberattacks cheaper to attempt, being overlooked becomes a less reliable security strategy.
Comment · 3 months ago
I think I was just following along with the previous post about running Gemma on a Xeon. Next I’m going to see which model can give the highest tokens/sec
Comment · 3 months ago
hey, I’m the author. That box has 384gb, but loading the model “only” uses about 80gb.
Comment · 3 months ago
Thank you for sharing!
Comment · 3 months ago
Thank you for sharing / linking!
Comment · 3 months ago
Author here, it looks like my original comment was flagged for some reason. The fix is open upstream as PR #2138 (https://github.com/ikawrakow/ik_llama.cpp/pull/2138)
Comment · 3 months ago
Author here. The short version: a viral post ran Gemma 4 on a 2016 Xeon; my Xeons are 2013, and the fork it used assumes AVX2, which Ivy Bridge doesn't have. The build failure was easy. The fun bug was the silent one: two MoE graph ops with no dispatch case on non-AVX2 builds, so every expert FFN output was uninitialized memory. Deterministic, NaN-free, fluent-looking multilingual gibberish.
The fix is open upstream as PR #2138 (https://github.com/ikawrakow/ik_llama.cpp/pull/2138), awaiting review. Fair warning on the AI angle: the patch was written by Claude at my direction. The post is explicit about which parts were me and which weren't. Happy to answer questions about either the bug or the workflow.
Comment · 4 months ago
As engines got more efficient, coal became cheaper to put to work, and so people found far more things to do with it. The same thing is happening with developer hours.
Comment · 7 months ago
We've been running Rails apps in production continuously since 2007. If you treat software as anything other than completely disposable, it's been a no-brainer for the entire 19+ years I've been paying attention (not despite its age, but because of it).
The premise that you get meaningful efficiencies from JavaScript on the back-end just because you have to use it on the front-end has been pretty thoroughly debunked at this point. Instead you mostly get a larger blast radius when the front-end ecosystem has its monthly identity crisis. OP's "stacks-du-jour" and programming language "flavour of the month" framing is exactly right. A shocking amount of web software architecture is just following fashion trends dressed up as technical decision-making.
Most of the churn in tech stack isn't driven by engineering requirements, it's driven by résumé optimization and Hacker News anxiety. Rails has quietly been powering serious businesses the whole time. Does anyone think NPM's 3.1 million packages enable more functionality than RubyGems' 190,000 packages?
Comment · 9 months ago
This looks interesting, but I'm stuck on step 4 of the web setup: where do I get agents to start with? Shouldn't there be a default one that can help me get other ones?
Comment · 11 months ago
FWIW I've had success with self-hosted [LinkDing](https://github.com/sissbruecker/linkding) and the firefox SingleFile plugin (so it archives what I'm seeing / gets around logins etc). LinkDing also links directly to Internet Archive for any URL.