FTL, 클라우드를 위한 새로운 운영체제
프로세스, 파일시스템, TCP/IP 같은 OS 기능을 컨테이너별 사용자 공간 라이브러리로 구현 해, 운영체제를 애플리케이션처럼 수정하고 디버깅할 수 있게 하는 클라우드용 OS 컨테이너마다 서로 다른 OS 기능과 버전 을 사용할 수 있으며, 머신 전체를 재부팅하지 않고...
TUMN 인사이트 · AI와 IT를 이해하는 실용 정보
AI와 IT의 변화를 이해하고, 실제 서비스에 적용하는 데 필요한 정보를 모읍니다.
CogSend - 여러 소셜 플랫폼에 글을 예약 발행하는 셀프 호스팅 도구
Mastodon, Bluesky, LinkedIn, Threads, X에 글을 한 번 작성하고 플랫폼별로 수정해 즉시 게시하거나 예약 발행 하는 오픈소스 도구 별도 서버 대신 자신의 Cloudflare 계정에서 운영 하며, 게시물과 계정 인증 정보는 D1 데이터베이스에, 이미지는 R2에 보관 ...
OpenAI safety leader quits, warning AI company's culture is 'broken'
Comments
The Agent Said It Was Done. The Database Disagreed.
The Agent Said It Was Done. The Database Disagreed.
We're going to need default hard budget caps on pretty much everything
<p>Here's a product feature which the world is going to need a whole lot more of over the coming months and years: <strong>default hard budget caps</strong>. I'm talking about the feature of pay-by-usage services and APIs that lets you say "after $X/month, cut this thing off and
September sponsors-only newsletter
<p>I just sent the September edition of my <a href="https://github.com/sponsors/simonw/">sponsors-only monthly newsletter</a>. If you are a sponsor (or start a sponsorship now) you can <a href="https://github.com/simonw-private/monthly/blob/main/2026-09-september.md">access it he
Claude와 Claude Code에서 Opus 5.5를 최대한 활용하는 방법
Opus 5.5는 긴 다단계 작업을 이어가는 능력이 개선돼, 작업 전체와 완료 조건, 멈춰서 질문할 조건 을 한 번에 전달하는 방식이 적합함 모델이 응답 전에 스스로 추론하므로 “신중하게 생각하라”는 상투적인 지시는 생략 할 수 있음. Claude Code에서는 eff...
나는 OpenAI의 조직 문화가 망가져 퇴사했다
OpenAI의 주요 출시마다 안전 보고서 작성을 이끌던 직원이 퇴사 했으며, AI 기업의 안전 문제는 개별 규칙이나 새 법률을 넘어 조직 문화 차원에서 살펴야 한다고 봄 더 큰 AI 시스템을 만드는 데 성공한 강한 자신감 과 끊임없이 전력 질주하는 업...
업스테이지, GPU 한 장에서 실행하는 Solar Mini 4 발표
정보 조회, 문서 분류, 도구 호출처럼 반복되는 에이전트 작업을 적은 연산과 비용으로 처리 하는 언어 모델 총 350억 파라미터 중 토큰당 30억만 활성화 하는 MoE 구조이며, 양자화 모델은 H100 80GB 한 장 에서 배포 가능 Artificial Ana...
Ringg’s AI agents resolve up to 65% of customer calls with OpenAI
Using GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.
How invideo improves color grading 3x with GPT‑6 Astra
With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.
Sam Altman’s remarks at the United Nations Security Council
OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.
Two years of OpenAI Academy
Marking two years of OpenAI Academy and bringing AI skills to even more communities.
OpenAI extends cyber access to Ukraine for civilian defense
OpenAI is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.
Introducing MentalHealthBench
MentalHealthBench is an expert-informed benchmark for evaluating helpful and safe AI responses across realistic mental health conversations.
Harvey turns legal context into stronger drafts with GPT-6 Astra
GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.
Multiplayer AI: Why your team (and its agents) need a group chat
Ryan chats with the GM of Slack, Rob Seaman, about how their new Code Channels feature is bringing multiplayer AI to your team chats.
Bringing Private Processing to Meta AI Glasses
We believe glasses are the best form factor for having AI help throughout your day. They can understand your personal context better than other kinds of devices and keep you present without picking up a mobile phone. Most of the time, glasses are helping you see well, protecting
Shadow roots, explained with live examples
<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/shadow-roots">Shadow roots, explained with live examples</a></p> <p>Prompt to Fable 5.1 Medium:</p> <blockquote> <p><code>Build an artifact to explain shadow roots in CSS with interactive examples</code></p> </blo
Validate GPU Cluster Readiness Before AI Workloads Land
A GPU cluster can pass every health check and still fail to run an AI workload. Even when every GPU, network link, and pod reports healthy, a 512-GPU training...
Gemini 3.8 TTS Playground
<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/gemini-tts-playground">Gemini 3.8 TTS Playground</a></p> <p>Google <a href="https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-text-to-speech/">released two new Gemini text-to-speec
We just shipped support for the ugliest part of HTTP: Vary
<p><a href="https://news.ycombinator.com/item?id=49823195#49823961">My comment</a> on <a href="https://news.ycombinator.com/item?id=49823195">We just shipped support for the ugliest part of HTTP: Vary</a> — Hacker News.</p><p>I've been wanting this from Cloudflare <em>for y
Introducing NV-Reason-CT Open 3D CT VLM for Radiologist Chain-of-Thought Reasoning
Radiology AI has made remarkable strides in detecting abnormalities across chest X-rays, pathology slides, and 2D scans. Yet one of the most clinically rich and...
Manage Kubernetes Node Fleets with NodeWright
Kubernetes manages what runs on your nodes. Managing the nodes themselves is the challenge: kernel settings, system packages, storage layouts, security agents,...
How SWE-Serve Exposes the Gap Between Local Tests and Live Serving
An AI coding agent’s patch can pass tests yet fail when the server loads a real model and handles requests. Evaluating changes to inference-serving software...
commit-rewriter 0.2
<p><strong>Release:</strong> <a href="https://github.com/simonw/commit-rewriter/releases/tag/0.2">commit-rewriter 0.2</a></p> <blockquote> <p>Support for branches other than the default branch. Use <code>uvx commit-rewriter --branch other</code> to run against another branch. <a
Northern Gannet, Great Blue Heron, California Brown Pelican
<p><img src="https://static.inaturalist.org/photos/740453177/large.jpg" alt="Northern Gannet"></p><p><img src="https://static.inaturalist.org/photos/740453341/large.jpg" alt="Great Blue Heron"></p><p><img src="https://static.inaturalist.org/photos/740454506/large.jpg" alt="Califo
Efficient MoE Training for Biological Foundation Models
As language models grow, scaling dense architectures becomes increasingly expensive. In a dense transformer, every token passes through every layer, so adding...
Note on 24th September 2026
<p>The more time I spend working with coding agents, the more convinced I am that they make software engineering even harder.</p> <p>We can do amazing things with them, but unlocking their full potential requires extraordinary discipline and knowledge.</p> <p>Tags: <a href="https
Accelerating vision-language models with LFM2.5-VL-DSpark
Accelerating vision-language models with LFM2.5-VL-DSpark