Efficient Streaming Language Models with Attention Sinks
Efficient Streaming Language Models with Attention Sinks is a technology topic drawing discussion on Hacker News.
Page 12 of 17 · 489 retained trends
Efficient Streaming Language Models with Attention Sinks is a technology topic drawing discussion on Hacker News.
Show HN: Anime Text Datasets for LLMs Training is a technology topic drawing discussion on Hacker News.
Anthropic hacked to deliver remote access trojan is a technology topic drawing discussion on Hacker News.
Cohere Qwen incident URL samples is a technology topic drawing discussion on Hacker News.
OpenAI doubles ChatGPT Pro usage limits after backlash over planned 50% cut is a technology topic drawing discussion on Hacker News.
Backdoors in official Claude Code extensions is a technology topic drawing discussion on Hacker News.
A Hacker News discussion about Ask HN: What inference engine and KV offload engine do you use and why?. The submitted context says: I have the question to the guys in inference world, I am building some stuff connected to the inference in my work, and I started wondering, what do actually big guys use in their…
Artificial Intelligence, jobs and society: What we know so far is a technology topic drawing discussion on Hacker News.
Blog FlexAttention: The Flexibility of PyTorch with the Performance of FA is a technology topic drawing discussion on Hacker News.
Book of Life (ChatGPT) is a technology topic drawing discussion on Hacker News.
Josh Kushner's Thrive and OpenAI to buy home security firm ADT is a technology topic drawing discussion on Hacker News.
A Hacker News discussion about Show HN: Cubic Doggo 07B Wouf: A low-cost open-source quadruped with ST3215. The submitted context says: Cubic Doggo 07B Wouf is a ROS2-controlled 12-DOF untethered quadruped, built with ST3215 smart servos fully with off-the-shelf components and 3D-printed parts Featuring: * 12…
Deep Learning Made Difficult is a technology topic drawing discussion on Hacker News.
A Hacker News discussion about Show HN: Lasm – Write formally verified webapps in Lean for Node/serverless/edge. The submitted context says: Lasm has been my dream for a good part of 2026. I firmly believe that formal verification is the next evolution of software engineering and programming languages. This next wave…
Trojanized Claude Code Extension Plants an SSH Backdoor is a technology topic drawing discussion on Hacker News.
Anthropic now lets Claude users jump out of plane without parachute is a technology topic drawing discussion on Hacker News.
ChatGPT release notes October 29, 2026 is a technology topic drawing discussion on Hacker News.
Show HN: Type and+ in Claude Code and a second Claude rewrites your prompt first is a technology topic drawing discussion on Hacker News.
Show HN: Axoris – macOS App for Grok, Claude, ChatGPT and Gemini is a technology topic drawing discussion on Hacker News.
How to Run Multiple Claude Code Agents on the Same Xcode Project is a technology topic drawing discussion on Hacker News.
Proof by OpenAI of the Existence of Quantum Yang-Mills Theory [pdf] is a technology topic drawing discussion on Hacker News.
LLMs bring new nature of abstraction (2025) is a technology topic drawing discussion on Hacker News.
appariciojunior/video-editor-skills is a GitHub project. Its repository describes it as: Skills that teach Claude Code to be a first-pass video editor in DaVinci Resolve: transcript cuts, vlog pacing, B-roll, music & SFX, clean captions, motion graphics, YouTube…
The Ability of Frontier Models to Control Robots Is Rapidly Improving is a technology topic drawing discussion on Hacker News.
Show HN: Ollama Mock Server is a technology topic drawing discussion on Hacker News.
We gave a political quiz to 31 LLMs; they converge more than people do is a technology topic drawing discussion on Hacker News.
Why Anthropic's IPO Prospectus Is Missing So Many Key Figures is a technology topic drawing discussion on Hacker News.
A Hacker News discussion about Show HN: ccfleet, a quiet, tmux-native Claude Code dashboard. The submitted context says: At work, we're leaning heavily into agentic development. I've always used tmux to manage my work and now I'm using it to organize my 30+ Claude sessions. I wanted a simple…
DeepSeek, Moonshot used OpenAI's GPT-6 Astra to train latest models is a technology topic drawing discussion on Hacker News.
Anthropic Scion inspired source code management is a technology topic drawing discussion on Hacker News.