LLMs Are Not a Default Execution Engine
Article URL: https://unmeshed.io/blog/using-ai-wisely-starts-before-the-first-prompt Comments URL: https://news.ycombinator.com/item?id=48812489 Points: 4 # Comments: 1
Communities 6878 stories
Local Ai Build – Part 2
Hey all. I have an older post here with these stacked / Air cooled. Just wanted to update the group here with how the build sits…
Strata – An app that talks me out of dying outdoors
Article URL: https://strata.highloop.co/ Comments URL: https://news.ycombinator.com/item?id=48812282 Points: 3 # Comments: 0
How to sequence your own DNA at home
Article URL: https://bradleywoolf.com/links-1/sequencing-my-own-dna-at-home Comments URL: https://news.ycombinator.com/item?id=48812156 Points: 16 # Comments: 2
Experiments With Fabel's Fiction
In keeping with my tradition and given I will lose access to Fabel in a couple days, I have asked fable to read all my organically…
Late to the party but… Holy MTP
Just ran Qwen 3.6 27B using MTP for the first time. Doubled my t/s. Wow. That is all. I'm going to go look for abliterated MTP…
Show HN: InstantVideos.org – short documentaries in ~30 seconds
Hiya! So I've been playing around with having Claude make videos for a bit now even had some success posting the results to TikTok (and setup…
Can we find whether models have been backdoored?
This is the first post in a two-part sequence regarding the state of defending from data poisoning attacks. We describe some methods for determining whether a…
Info dump: Some Important Models for Health and Fitness
This is a synthesis of many facts I've learned over the last few years, mainly about metabolism and exercise,[1] that have helped me become much healthier,…
NSA and IETF: Fairness
Article URL: https://blog.cr.yp.to/20260706-fairness.html Comments URL: https://news.ycombinator.com/item?id=48811887 Points: 5 # Comments: 0
Claude Code as a Claude Coach
Exercise is hard but it's even harder if you have to use your brain and muscles at the same time. I wish a personal trainer would…
Ternlight – 7 MB embedding model that runs in browser (WASM)
Article URL: https://ternlight-demo.vercel.app/ Comments URL: https://news.ycombinator.com/item?id=48811644 Points: 22 # Comments: 1
Are there any interesting local versions of the OpenRouter "Fusion" or Sakana Fugu methods, being worked on, as of yet? I’m curious if these setups will enable us to use, say, a panel of various ~30B-ish local models together to get outputs closer to GLM quality without needing as much memory.
Given how much of a boost in output quality these methods seem to enable when it comes to the big cloud models of having several smaller,…
Be less open-minded: Against uncalibrated open-mindedness
People sometimes tell me to be more open-minded. That it’s always a positive trait, all else being equal.I disagree.I think what they mean is that you…
Fable turned remarkable into Tom Riddle's diary from Harry Potter
Article URL: https://github.com/MaximeRivest/Riddle Comments URL: https://news.ycombinator.com/item?id=48811591 Points: 29 # Comments: 6
Vessel
Article URL: https://claude.ai/public/artifacts/2c647671-3f94-4d22-b3c1-f2b5a0e17b6e Comments URL: https://news.ycombinator.com/item?id=48811494 Points: 10 # Comments: 1
OpenSSH 10.4/10.4p1 Released
Article URL: https://www.openssh.org/txt/release-10.4 Comments URL: https://news.ycombinator.com/item?id=48811373 Points: 7 # Comments: 3
Acronym Fatigue Series Introduction: why I'm wary of acronyms
Article URL: https://devz.cl/posts/acryonym-fatigue-series-why-i-m-wary-of-engineering-acronyms/ Comments URL: https://news.ycombinator.com/item?id=48811339 Points: 4 # Comments: 0
So what’s the best software/kernel/config combo for SM120 FP8 training?
I’m having a hellluva time trying to fine-tune an FP8 LLM model (both FFT and LORA). I’ve been using VLLM and they sure don’t make it…
Poly/ML – A Standard ML Implementation
Article URL: https://github.com/polyml/polyml Comments URL: https://news.ycombinator.com/item?id=48811317 Points: 7 # Comments: 1
How little exercise can you get away with?
Article URL: https://www.economist.com/science-and-technology/2026/07/03/how-little-exercise-can-you-get-away-with Comments URL: https://news.ycombinator.com/item?id=48811147 Points: 25 # Comments: 8
Full Writeup of the Windows GDID
Article URL: https://github.com/SmtimesIWndr/gdid-reversal Comments URL: https://news.ycombinator.com/item?id=48811081 Points: 8 # Comments: 3
Titan's Resources and Their Utilization
Article URL: https://arxiv.org/abs/2606.06608 Comments URL: https://news.ycombinator.com/item?id=48810841 Points: 4 # Comments: 0
Local models + big context = slow. How are you orchestrating "map-reduce" style agent workflows?
I tried running local models (qwen3.6*, ds4 flash, gemma4*, etc) on my mbp pro m5 with 128Gb of unified memory and concluded the bottleneck is context…