Traffic matrices (TMs) capture network-wide origin-destination demand and are central to traffic engineering, yet accurate whole-matrix forecasting remains challenging when prediction must be performed under the…
Article URL: https://xkcd.com/3266/large/ Comments URL: https://news.ycombinator.com/item?id=48777832 Points: 31 # Comments: 3
Article URL: https://inpreparation.substack.com/p/opinion-i-was-not-allowed-to-type Comments URL: https://news.ycombinator.com/item?id=48777728 Points: 34 # Comments: 16
RT Sean Smithhttp://x.com/i/article/2073093772817453056
In a previous post, I explain why the universe is probably not stable, but nevertheless unlikely to be intentionally destroyable even in the limit of advanced…
RT dexdamn go find the AIE day 1 livestream and watch the interview with @GergelyOrosz and @Sirupsen it’s super fun and some great technical content on…
RT Shlok KhemaniSome notes from @aiDotEngineer world fair:> the energy was incredible. it's magical to have a large group of smart, hungry, technical, and driven people…
RT Pratim BhosaleHad a long and eventful week at @aiDotEngineer It’s time to go back to Paris and ship. Taking back some great conversations, ideas and…
RT Victoria X LinThe video of my Stanford CS25 guest lecture, From Language Models to Native Multimodal Intelligence, is now online.I discussed how the core ideas…
The team at @vercel recently released the Eve agent framework, so we built a template that integrates LiteParse with it🦙 The template provides a set of…
Sent out my sponsors-only monthly newsletter for June, which means the May newsletter is now available here https://github.com/simonw/monthly-newsletter-archive/blob/main/2026-05-may.md - you can sponsor me through GitHub to…
https://huggingface.co/bartowski/DeepSeek-V4-Flash-GGUF This model is in MXFP4 and as such has only been provided in MXFP4 format! Original model: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash But on original's page none of tensors…
I have an intel n100 mini pc that's on 24/7 running proxmox. I want to use llama.cpp server with gemma 4 E2B for small tasks. Should…
RT Han Xiao ✈️ ICML 2026still many industry executives don't get the value of top-tier AI conf like ICML/ICLR/NeurIPS, they'd rather sit on some summit roundtable…
Anthropic is trying to block Chinese companies like ByteDance and Ant Financial from accessing Claude Code, but they're getting around the restrictions through VPNs and overseas…
Just a datapoint I wanted to share.Qwen 27b, at q6kxl, with multi-token prediction, on a 4090+3090 system, using lcpp, puts out 50-90 tokens/s decode and 1500-2200…
Article URL: https://www.fide.com/fide-ethics-disciplinary-commission-issues-a-decision-in-case-involving-gm-vladimir-kramnik/ Comments URL: https://news.ycombinator.com/item?id=48777266 Points: 20 # Comments: 5
"At least this felt ambitious until fable chewed through each of my requests like it was nothing..."Aaron Becker: @simonw I made this 3D video infographic explaining…
Article URL: https://github.com/kerlenton/mcpsnoop Comments URL: https://news.ycombinator.com/item?id=48777144 Points: 9 # Comments: 3
Epoch AI reports a sharp rise in security vulnerability reports. In June 2026, 21 organizations reported about 1,500 high-severity and critical CVEs, more than 3.5 times…
Every time I read Sebastian’s books, I understand LLMs better. I can’t recommend his build an LLM / reasoning model from scratch series enough. Reading the…
Article URL: https://www.theguardian.com/us-news/2026/jul/03/datacenter-recall-elections Comments URL: https://news.ycombinator.com/item?id=48776780 Points: 38 # Comments: 13
Hi, I am supposed to prepare for SLM and its software part for an on campus internship, i've worked with local models like ollama generally,in my…
In a study covering seven benchmarks, the UK's AI Security Institute shows that standard AI evaluations systematically underestimate agent capabilities by capping the compute budget. On…