We compared different LLMs on IMO 2026 [R]
There are a few reasons why problems from International Mathematical Olympiad function as a good benchmark for LLMs: - The problems are new, not included in…
Latest
The AI coding tutor paradox grows as educators scramble to rethink how they test real skills
An ACM survey of 763 computer science educators from 49 countries shows that 68 percent have already changed their exams because of AI, shifting toward oral…
I implemented the YOLO26n model inference from scratch using ARM64 Assembly Language (No framework) [P]
This was my Bachelor's Final Project: implementing YOLO26n inference completely from scratch using ARM64 Assembly Language and C, without relying on existing inference frameworks. The goal…
A year later, Fable builds me the Cezanne city builder game. The AI came up with the idea of an impressionist city builder where you paint with gestur…
A year later, Fable builds me the Cezanne city builder game. The AI came up with the idea of an impressionist city builder where you paint…
RT hardmaru: 今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️
RT hardmaru今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️Denny’s: @JensenHuang @nvidia
RT hardmaru: 今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️
RT hardmaru今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️Denny’s: @JensenHuang @nvidia
今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️
今夜はデニーズのドリンクバーでオープンモデルの未来について語り合うしかない ☕️Denny’s: @JensenHuang @nvidia
AI Morning Brief — 26 July 2026
Overnight in AI: an open-weights letter left Anthropic as the lone major holdout, OpenAI's Hugging Face breach deepened, and Claude Opus 5 launched.
GrapheneOS protections against data extraction from locked devices
Article URL: https://discuss.grapheneos.org/d/40700-grapheneos-protections-against-data-extraction-from-locked-devices Comments URL: https://news.ycombinator.com/item?id=49055169 Points: 3 # Comments: 0
Simple Dutch books versus Sleeping Beauty halfers
The Sleeping Beauty problem is a famous philosophical puzzle. Described by Elga (2000): Some researchers are going to put you to sleep. During the two days…
Any use cases for RTX PRO 4500?
At its price point, PRO 4500 doesn’t offer as much raw performance due to its lower power draw at 300W. The 5090 can perform up to…
GLM 5.2 and ik_llama.ccp
Running GLM-5.2 (the new glm-dsa arch), Unsloth UD-Q4_K_XL, on a 4-socket Xeon E7-8880 v4 box with 1TB RAM and a single RTX 3060 12GB. ik_llama.cpp, experts…
RT Jeffery Mead: Elon’s not wrong, much of the mainstream media is hated. Also, the left has ran so far left that they now call normal things far rig…
RT Jeffery MeadElon’s not wrong, much of the mainstream media is hated.Also, the left has ran so far left that they now call normal things far…
b10133
ui: fix context gauge card regressions and land at the conversation e…
sqlite-utils 3.39.1
Release: sqlite-utils 3.39.1 I back-ported a fix for table.delete_where() that shipped in version 4. Tags: sqlite-utils
true fact
true factTitanium: Codex is sooooo good! Damn!
W4ME Station – a WASM-4 runtime for Java ME phones
Article URL: https://github.com/mulfyx/w4me-station Comments URL: https://news.ycombinator.com/item?id=49054697 Points: 12 # Comments: 2
RT Yun-Ta Tsai: Many engineers in my teams find the power of /voice in Grok Build. They keep brain-dumping their ideas. Combining it with /dream, /rec…
RT Yun-Ta TsaiMany engineers in my teams find the power of /voice in Grok Build. They keep brain-dumping their ideas. Combining it with /dream, /recap and…
An OpenAI model left notes about how to evade containment; we need more details
The OpenAI AI attack on Hugging Face wasn’t the first loss of control incident at OpenAI, Reuters recently reported, and perhaps not even the most concerning.In…
RT Aakash Gupta: The Matrix took its name from a 1984 novel written on a manual typewriter by a man who didn't own a computer. 42 years later, that no…
RT Aakash GuptaThe Matrix took its name from a 1984 novel written on a manual typewriter by a man who didn't own a computer. 42 years…
How do you optimise and test your model configs/params? (blindly using GPT-5.6 to help me test/measure/optimise)
Current setup: headless machine for LLMs llama.cpp in router mode models settings configured via models.ini documentation in a repo which AI manicures It seems to be…
Understanding GPU Inference Workloads [D]
Hey everyone, I have been looking into how people source compute for their Inference workloads (and in general). I wanted to understand some specific pain points…
/dream
/dreamMorgan: I've been using Grok Build since the first 24-hours it was made available. It is still my favorite agentic coding cli, and there are so…
M2 Ultra 64gb vs m1 ultra 128gb
Trying to weigh if I should buy a $3000 m1 ultra at 128gb when I currently already have an M2 Ultra albeit at 64gb ram. I…