stared

14,022 karmajoined September 2, 2012On HN

A curious being, doctor of sorcery. Posts, projects & resume at: https://p.migdal.pl/

Now: benchmarking AI at: https://quesma.com/blog/

Previously: co-founder & CTO at https://quantumflytrap.com/

Recent submissions

Knowledge vs. wisdom: asking AI "What mushroom is that?"quesma.com
2stared2ddiscuss
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses quesma.com
284stared3d140 comments
I threw a ton of various logic puzzles at Astra and it solved of them twitter.com
1stared5ddiscuss
GPT-6 Astra has autonomously completed Portal twitter.com
8stared5ddiscuss
Does your landing page look generated?slop-detect.com
3stared5d3 comments
A connectomics milestone: Mapping the complete male fruit fly brain research.google
2stared5ddiscuss
Show HN: Genetic Distance Map p.migdal.pl
2stared6d1 comment
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses quesma.com
3stared7d1 comment
Beef production in Brazil has been the largest driver of global deforestation ourworldindata.org
5stared9ddiscuss
Mushroom hunting with LLMs: what can go wrong?quesma.com
56stared9d75 comments
Witcher 3's DLC Won't Include Russian Voice Acting thegamer.com
5stared9d3 comments
Forget the Pelican, It's Weevil-Time Benchmaxxing-Proof SVG and Vision Test reddit.com
1stared10ddiscuss
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses quesma.com
3stared14d2 comments
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses quesma.com
15stared16d2 comments
If Eastern Europeans reported the world like the world reports Eastern Europe badbaltic.com
4stared18ddiscuss
Gemini 3.7 Flash, Grok 4.6, GLM-5.3 and DeepSeek V4 Pro joined the frontier quesma.com
2stared19ddiscuss
NanoGPT Speedrun Frontier primeintellect.ai
140stared20d39 comments
Gemini 3.7 Flash, Grok 4.6, GLM-5.3 and DeepSeek V4 Pro joined the frontier quesma.com
5stared22ddiscuss
Pander Score: How much do AI models mirror what users believe?sophronresearch.org
2stared23ddiscuss
Show HN: Claude Code and Codex usage screen for TRMNL X e-ink github.com/stared
3stared1modiscuss
AI assistant hacks gym website in first known Australian autonomous cyber attack abc.net.au
78stared1mo60 comments
Quantization hurts knowledge nonlinearly – Qwen3.6 27B case study quesma.com
2stared1modiscuss
Kimi K3 Is Open, Opus 5 Is Good, DeepSeek V4 Flash Is Cheap: LLMs on Baba Is You quesma.com
1stared1mo1 comment
Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?quesma.com
2stared1mo1 comment
Pomegranate – Sci-Fi Short Film [video]youtube.com
2stared1modiscuss
Benchmarking Kimi K3, Opus 5, Grok 4.5, and Gemini 3.6 Flash on Baba Is You quesma.com
4stared1modiscuss
Do Qwen 3.6 27B quantizations break the pelican?quesma.com
9stared1modiscuss

All submissions and comments on Hacker News.