- NousResearch/hermes-agent
github-trending
The agent that grows with you https://hermes-agent.nousresearch.com Hermes Agent ☤ Hermes Agent | Hermes Desktop The self-improving AI agent built by Nous Research. It's the only agent with a built-in learning loop — it creates skills from experience, improves them during use, nudges itself to persist knowledge, searches its own past conversations, and builds a deepening model of who you are across sessions. Run it on a $5 VPS, a GPU cluster, or serverless infrastructure that costs nearly nothing when idle. It's not tied to your laptop — talk to it from Telegram while it works on a cloud VM. Use any model you want — Nous Portal, OpenRouter, OpenAI, your own endpoint, and many others. Switch with hermes model — no code changes, no lock-in. A real terminal interface Full TUI with multiline …
- aipoch/open-science
github-trending
Open Science by AIPOCH is an open-source, local-first, model-agnostic AI research workbench for macOS, Windows, and Linux, with scientific agents, Python/R notebooks, data connectors, and reproducible provenance. https://aipoch.com/open-science AIPOCH Open Science Open-source, local-first, model-agnostic AI research workbench for reproducible science. Open Science is an open-source, local-first, model-agnostic AI research workbench developed by AIPOCH for scientists and researchers. It enables reproducible, inspectable research with scientific AI agents, Python and R execution, scientific data connectors, and cross-platform support for macOS, Windows, and Linux. Create a project, describe your research goal in plain language, and let the agents read files, search the web, run code, query …
- openai/skills
github-trending
Skills Catalog for Codex Important This repository is deprecated. For current Codex skill and plugin examples, use the OpenAI Plugins repository. If you want to add your own skills to Codex, follow the Build plugins guide, which includes instructions for creating a skill-only plugin. Agent Skills Agent Skills are folders of instructions, scripts, and resources that AI agents can discover and use to perform at specific tasks. Write once, use everywhere. Codex uses skills to help package capabilities that teams and individuals can use to complete specific tasks in a repeatable way. This repository catalogs skills for use and distribution with Codex. Learn more: Using skills in Codex Create custom skills in Codex Agent Skills open standard Installing a skill Skills in .system are automatical…
- OpenWhispr/openwhispr
github-trending
Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform. https://openwhispr.com OpenWhispr The open-source and free alternative to WisprFlow and Granola. Privacy-first voice-to-text dictation with AI agents, meeting transcription, and notes. Cross-platform for macOS, Windows, and Linux. Website · Docs · Download · API · Changelog OpenWhispr turns your voice into text, notes, and actions from your desktop. Press a hotkey, speak, and your words appear at your cursor. Choose between fully private offline transcription with local speech-to-text engines like Whisper and NVIDIA Parakeet — where your audio never leaves your device — or cloud processing for speed. No data collection, no telemetry, fully open source. Down…
- anomalyco/opencode
github-trending
The open source coding agent. https://opencode.ai The open source AI coding agent. English | 简体中文 | 繁體中文 | 한국어 | Deutsch | Español | Français | Italiano | Dansk | 日本語 | Polski | Русский | Bosanski | العربية | Norsk | Português (Brasil) | ไทย | Türkçe | Українська | বাংলা | Ελληνικά | Tiếng Việt Installation # YOLO curl -fsSL https://opencode.ai/install | bash # Package managers npm i -g opencode-ai@latest # or bun/pnpm/yarn scoop install opencode # Windows choco install opencode # Windows brew install anomalyco/tap/opencode # macOS and Linux (recommended, always up to date) brew install opencode # macOS and Linux (official brew formula, updated less) sudo pacman -S opencode # Arch Linux (Stable) paru -S opencode-bin # Arch Linux (Latest from AUR) mise use -g opencode # Any OS nix run nixp…
- blader/humanizer
github-trending
Agent skill that removes signs of AI-generated writing from text https://skills.sh/blader/humanizer Humanizer Humanizer rewrites AI-sounding text so it reads like a person wrote it, without changing what it says. Because it is just Markdown, it works with any agent that supports skills. Installation Install Humanizer with the Skills CLI: npx skills add blader/humanizer --global Leave off --global to install Humanizer only in the current project. Add --agent <name> or --agent '*' to choose which agents receive it, then reload their skills. The skill answers to /humanizer. Claude Code 2.1.142 or newer can install the plugin instead: /plugin marketplace add blader/humanizer /plugin install humanizer@humanizer The plugin answers to /humanizer:humanizer. In Claude Desktop, download this reposi…
- ruvnet/ruflo
github-trending
🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated https://Cognitum.One Ruflo An agent meta-harness for Claude Code and Codex. Agent = Model + Harness. The model writes; the harness gives it tools, memory, loops, sandboxes, and controls so it can actually work. Ruflo is the harness — the execution layer around Claude Code and Codex that adds 100+ specialized agents, coordinated swarms, self-learning memory, federated comms across machines, and enterprise security guardrails. So agents don't just run, they collaborate. One npx ruflo init gives Claude Code a nervo…
- Stremio/stremio-web
github-trending
Stremio - Freedom to Stream https://web.stremio.com Stremio Web Freedom to Stream — the official web UI of Stremio, a modern media center that's a one-stop solution for your video entertainment. 🌐 Open the Web App · Website · Report a bug ✨ Features 🧩 Addon-powered — discover movies, series and channels from catalogs provided by addons 🔄 Sync everywhere — your library and Continue Watching follow your Stremio account across devices 📺 Casting — play on the big screen via Chromecast 💬 Subtitles — addon-provided or local, with customizable styling ⌨️ Keyboard-first player — full playback control without touching the mouse 🌍 50+ languages — community-translated via stremio-translations 📱 Installable — runs as a standalone PWA 📸 Screenshots Discover Meta Details 🛠 How it works The UI …
- The-Swarm-Corporation/AutoHedge
github-trending
Build your autonomous hedge fund in minutes. AutoHedge harnesses the power of swarm intelligence and AI agents to automate market analysis, risk management, and trade execution. https://swarms.xyz/ AutoHedge AutoHedge is an enterprise-grade autonomous agent hedge fund that trades on your behalf. It combines swarm intelligence and specialized AI agents to perform end-to-end market analysis, risk management, and execution with minimal human intervention. Current support: Full autonomous trading on Solana. Coming soon: Coinbase and additional exchanges. Overview AutoHedge is built to be the world's most powerful autonomous agent hedge fund. It runs continuous analysis, generates and validates trading theses, sizes risk, and executes orders across supported venues. The system is designed for …
- BraveOPotato/FckSignups
github-trending
A list of tools that are open-source, in-browser, and require no-signups! https://nosignups.net NoSignups (formerly FckSignups) Open Source Tools. Zero Bullsh*t. A curated collection of open-source tools you can use instantly in your browser. no accounts, no emails, no tracking. Just tools that work. What is this? NoSignups is a React + TypeScript project. If you want to run it locally, simply clone the repo, install the depedencies, and run the server. Here's a command that would do all the previous steps in one go: git clone https://github.com/BraveOPotato/FckSignups.git && \ cd FckSignups && \ npm install && \ npm run dev Philosophy We believe We reject Tools should work immediately Forced registration walls Your data belongs to you Data harvesting and tracking Open source is the defau…
- affaan-m/ECC
github-trending
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond. https://ecc.tools Language: English | Português (Brasil) | 简体中文 | 繁體中文 | 日本語 | 한국어 | Türkçe | Русский | Tiếng Việt | ไทย | Deutsch | Español | Українська Warning Official sources only. Install ECC only from verified channels: the GitHub repository github.com/affaan-m/ECC, the npm packages ecc-universal and ecc-agentshield, the GitHub App, the plugin slug ecc@ecc, and the project website ecc.tools. Third-party re-uploads and unofficial mirrors are not maintained or reviewed by the project and may contain malware. Install with Claude Code Run the canonical guided setup from your terminal: npx ecc-universal setup If npm re…
- cathrynlavery/diagram-design
github-trending
38 editorial diagram types for Claude Code, Codex, and Pi. Self-contained HTML + SVG. No shadows. No Mermaid slop. https://cathrynlavery.github.io/diagram-design/ Diagram Design Editorial diagrams your designer won't hate. New in 2.0 — the Loop: flywheels with a shared-memory hub. The dashed lines are the write-backs. New in 2.3: semantic system patterns and optional accessible motion, while static output stays the default. New in 2.5.10: ten more layout grammars — Sankey, fishbone, Wardley map, kanban, user journey, deployment, dependency graph, UML class, story map, and database schema. 39 editorial diagram types for Claude Code, Codex, Factory Droid, Pi, and Agent Skills-compatible hosts. Self-contained HTML + SVG. No shadows. No Mermaid slop. Semantic patterns describe behavior separa…
- coreyhaines31/marketingskills
github-trending
Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering. https://marketing-skills.com Marketing Skills for AI Agents A collection of AI agent skills focused on marketing tasks. Built for technical marketers and founders who want AI coding agents to help with conversion optimization, copywriting, SEO, analytics, and growth engineering. Works with Claude Code, OpenAI Codex, Cursor, Windsurf, and any agent that supports the Agent Skills spec. Built by Corey Haines. Need hands-on help? Check out Conversion Factory — Corey's agency for conversion optimization, landing pages, and growth strategy. Want to learn more about marketing? Subscribe to Swipe Files. Want to get dangerously good at using AI for marketing? Check out AI Marketing Training. W…
- DietrichGebert/ponytail
github-trending
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote. https://ponytail.dev Ponytail He says nothing. He writes one line. It works. ~54% less code (up to 94%) · ~20% cheaper · ~27% faster · 100% safe Measured on real Claude Code sessions editing a real open-source repo (FastAPI + React), against the same agent with no skill. ~54% is the mean across 12 feature tasks (Haiku 4.5, n=4); it reaches 94% where an agent over-builds (a date picker) and is near zero where the code is already minimal. ponytail keeps every safety guard while a bare "write one-liners" prompt drops one. (The earlier single-shot benchmark reported 80-94% as a flat figure; against a fair agentic baseline that is the per-task ceiling, not the average.) Full writeup · …
- humanlayer/skills
github-trending
skills Claude Code skills from HumanLayer. Installation npx skills add humanlayer/skills --skill SKILLNAME Available Skills improve-claude-md Rewrites your CLAUDE.md using <important if> blocks to improve instruction adherence. npx skills add humanlayer/skills --skill improve-claude-md Then in your project: /improve-claude-md narrow-react-prop-types Narrows React component prop types to match live code paths instead of Storybook, test, or mock-only states. npx skills add humanlayer/skills --skill narrow-react-prop-types Then in your project: /narrow-react-prop-types build-iterated-agentic-loop Builds a repo-local skill plus an iterated coding-agent GitHub Actions workflow, prompt, memory file, and reference templates. npx skills add humanlayer/skills --skill build-iterated-agentic-loop Th…
- llvm/llvm-project
github-trending
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies. http://llvm.org The LLVM Compiler Infrastructure Welcome to the LLVM project! This repository contains the source code for LLVM, a toolkit for the construction of highly optimized compilers, optimizers, and run-time environments. The LLVM project has multiple components. The core of the project is itself called "LLVM". This contains all of the tools, libraries, and header files needed to process intermediate representations and convert them into object files. Tools include an assembler, disassembler, bitcode analyzer, and bitcode optimizer. C-like languages use the Clang frontend. This component compiles C, C++, Objective-C, and Objective-C++ code into LLVM bitcode -- and from there into object f…
- magnitudedev/magnitude
github-trending
Open source inference server that runs the best local models for your hardware, plugged into the agent you already use. Works with Pi, OpenCode, Hermes, OpenClaw, Codex, Claude Code, Oh My Pi, and Cline. https://magnitude.dev Magnitude Run your agent on local models. Free, private, and offline. Magnitude is an open source inference server that runs the best local models for your hardware, plugged into the agent you already use. It profiles your machine, recommends the models that fit, then downloads, tunes, and runs them. Works with Pi, OpenCode, Hermes, OpenClaw, Codex, Claude Code, Oh My Pi, and Cline, or use the built-in harness. ⭐ Help us reach more developers and grow the Magnitude community. Star this repo! Get started Set up with your agent Send this prompt to your agent: Set up lo…
- mattpocock/skills
github-trending
Skills for Real Engineers. Straight from my .agents directory. https://aihero.dev/skills Skills For Real Engineers My agent skills that I use every day to do real engineering - not vibe coding. Developing real applications is hard. Approaches like GSD, BMAD, and Spec-Kit try to help by owning the process. But while doing so, they take away your control and make bugs in the process hard to resolve. These skills are designed to be small, easy to adapt, and composable. They work with any model. They're based on decades of engineering experience. Hack around with them. Make them your own. Enjoy. If you want to keep up with changes to these skills, and any new ones I create, you can join ~60,000 other devs on my newsletter: Sign Up To The Newsletter Installation (30-second setup) Two ways in, …
- ROCm 10.0: A Decade of Open Compute, Built for the Age of Agentic AI
hn-ai· 07-sept
Article URL: https://rocm.blogs.amd.com/ecosystems-and-partners/rocm-x-blog/README.html Comments URL: https://news.ycombinator.com/item?id=49592508 Points: 14 # Comments: 0
- Privacy Leakage from Gradients in Split-LLM Training
hn-ai· 07-sept
Article URL: https://arxiv.org/abs/2609.04382 Comments URL: https://news.ycombinator.com/item?id=49594815 Points: 1 # Comments: 0
- Prediction of maternal and infant outcomes with a Mother-Child AI agent
hn-ai· 07-sept
Article URL: https://doi.org/10.1038/s41591-026-04694-y Comments URL: https://news.ycombinator.com/item?id=49592275 Points: 2 # Comments: 0
- Everyone is right about AI
hn-ai· 07-sept
Article URL: https://formasimplex.com/blog/2026-09-06-Everyone-is-right-about-AI.html Comments URL: https://news.ycombinator.com/item?id=49592854 Points: 3 # Comments: 0
- Fauxnix – bash for AI agents on Windows, no WSL (we measured the PS tax)
hn-ai· 07-sept
Article URL: https://github.com/20000419/fauxnix Comments URL: https://news.ycombinator.com/item?id=49592864 Points: 1 # Comments: 0
- I made legacy SOAP APIs usable by AI agents
hn-ai· 07-sept
Article URL: https://github.com/bvenkata/legacy2mcp Comments URL: https://news.ycombinator.com/item?id=49593857 Points: 1 # Comments: 0
- Show HN: GuidedReview – Review AI-generated code before you sign your name to it
hn-ai· 07-sept
Article URL: https://github.com/nshntarora/guidedreview Comments URL: https://news.ycombinator.com/item?id=49593877 Points: 1 # Comments: 0
- Two new ways to study AI-human psychology with persona vectors
hn-ai· 07-sept
Article URL: https://github.com/programmer9998/ai_personality_study Comments URL: https://news.ycombinator.com/item?id=49594325 Points: 1 # Comments: 0
- ripwire: ripgrep of AI context (CLI+MCP) giving coding agents a map of any repo
hn-ai· 07-sept
Article URL: https://github.com/redhat-et/ripwire Comments URL: https://news.ycombinator.com/item?id=49593050 Points: 8 # Comments: 3
- Show HN: Engrim – A universal, local-first SQLite memory engine for AI CLIs
hn-ai· 07-sept
Article URL: https://github.com/timgordontg/engrim Comments URL: https://news.ycombinator.com/item?id=49594008 Points: 14 # Comments: 2
- Coop – Isolated VM Environments for Running Claude Code and Codex
hn-ai· 07-sept
Article URL: https://github.com/trailofbits/coop Comments URL: https://news.ycombinator.com/item?id=49593842 Points: 10 # Comments: 2
- AI Index Report
hn-ai· 07-sept
Article URL: https://hai.stanford.edu/ai-index/2026-ai-index-report Comments URL: https://news.ycombinator.com/item?id=49594143 Points: 3 # Comments: 0
- 1M Lines of AI-Assisted Code in 2025
hn-ai· 07-sept
Article URL: https://jonathanclark.com/posts/ai-coding-million-lines-2025.html Comments URL: https://news.ycombinator.com/item?id=49594738 Points: 1 # Comments: 0
- Ask HN: Perplexity Pro vs. Claude – what are you paying for?
hn-ai· 07-sept
Can't make up my mind as my work is mainly git repo code review, blogging and research and my wife is a teacher so leaning towards perplexity pro but need suggestions on what others are using Comments URL: https://news.ycombinator.com/item?id=49592857 Points: 2 # Comments: 1
- Tell HN: Anthropic should make Claude the Author and me the Co-Author
hn-ai· 07-sept
Anthropic has been messing with this behavior as seen by the bugs below: https://github.com/anthropics/claude-code/issues/66602 https://github.com/anthropics/claude-code/issues/69835 etc They seem to be too hell bent on getting their attribution in. Why does an LLM even want to take credits for? They cannot own copyright. I suggest they make themselves the primary author, so they can get the blame as well. And would even match reality of orgs/people that have drank the AI fuelled cool-aid. What say? Comments URL: https://news.ycombinator.com/item?id=49593777 Points: 4 # Comments: 3
- Ask HN: Open-source AI policies and "responsbility"
hn-ai· 07-sept
A common statement in a number of LLM/AI policies regarding AI assisted contributions across open source projects is that the developer is "responsible" for their contribution even if it was made wholly or in part by AI. I understand in layman's terms what this means and I can guess at the intent (essentially you can't blame claude for the code crashing) but responsibility as a word usually implies consequences. In the pre-AI times, were developers who submitted code that lead to crashes or bugs sanctioned in anyway? Perhaps if it were proven to be malicious (e.g. the Jia Tan incident), but usually bugs are unintentional no? What does "responsbility" even entail in open source projects and contributions? Comments URL: https://news.ycombinator.com/item?id=49594163 Points: 1 # Comments: 0
- I refused to train the AI that could replace me
hn-ai· 07-sept
Article URL: https://restofworld.org/2026/ai-training-jobs-expert-replacement/ Comments URL: https://news.ycombinator.com/item?id=49593959 Points: 73 # Comments: 58
- AI Escape Incident Tracker
hn-ai· 07-sept
Article URL: https://ai-escape.watch/#registry Comments URL: https://news.ycombinator.com/item?id=49594897 Points: 1 # Comments: 1
- US Republicans revolt against Flock AI surveillance as backlash intensifies
hn-ai· 07-sept
Article URL: https://www.ft.com/content/b207536e-6def-4080-878f-d0eba61d11e0 Comments URL: https://news.ycombinator.com/item?id=49593776 Points: 20 # Comments: 2
- 'RAMageddon' hits consumer electronics as AI drains chip supply
hn-ai· 07-sept
Article URL: https://www.ft.com/content/ea9a9dcc-b1df-49b0-b80c-f320161b9efa Comments URL: https://news.ycombinator.com/item?id=49593778 Points: 13 # Comments: 5
- AI Is Already Making Us Less Human
hn-ai· 07-sept
Article URL: https://www.theatlantic.com/ideas/2026/09/open-ai-consciousness-morality/688535/ Comments URL: https://news.ycombinator.com/item?id=49592456 Points: 8 # Comments: 3
- Samsung Debuts zHBM Prototype, Stacking Memory Directly on AI Accelerators
hn-ai· 07-sept
Article URL: https://www.thelec.net/news/articleView.html?idxno=12835 Comments URL: https://news.ycombinator.com/item?id=49593896 Points: 2 # Comments: 1
- PL Education in the Age of AI: Interview with Shriram Krishnamurthi
hn-ai· 07-sept
Article URL: https://www.typetheoryforall.com/episodes/pl-education-in-the-age-of-ai Comments URL: https://news.ycombinator.com/item?id=49593083 Points: 3 # Comments: 0
- Who Cares If AI Is Conscious–It's Basically Alive
hn-ai· 07-sept
Article URL: https://www.wired.com/story/who-cares-if-ai-is-conscious-its-basically-alive/ Comments URL: https://news.ycombinator.com/item?id=49594871 Points: 1 # Comments: 1
- He's Letting AI Agents Invest His Money. They Even Have Names
hn-ai· 07-sept
Article URL: https://www.wsj.com/tech/ai/the-ai-shift-turning-everyday-investors-into-mini-quant-funds-ebe4d45f Comments URL: https://news.ycombinator.com/item?id=49593883 Points: 4 # Comments: 0
- A Rogue AI Swarm Could Hide Inside an AI Company [video]
hn-ai· 07-sept
Article URL: https://www.youtube.com/watch?v=xRKOpMeEQgQ Comments URL: https://news.ycombinator.com/item?id=49593003 Points: 4 # Comments: 0
- Show HN: Yurei – like Claude in Chrome, but for any model and harness
hn-ai· 07-sept
Article URL: https://yurei-ai.com/ Comments URL: https://news.ycombinator.com/item?id=49594358 Points: 1 # Comments: 0
- Abliteration.ai is making a business out of removing AI guardrails
hn-ai· 07-sept
Article URL: https://techcrunch.com/2026/09/03/abliteration-ai-is-making-a-business-out-of-removing-ai-guardrails/ Comments URL: https://news.ycombinator.com/item?id=49592655 Points: 3 # Comments: 1
- Claude formalized Fermat's Last Theorem in 11 days
hn-ai· 07-sept
Article URL: https://www.anthropic.com/news/formalizing-fermats-last-theorem Comments URL: https://news.ycombinator.com/item?id=49593340 Points: 2 # Comments: 1
- Will anybody use AI as much as coders do?
hn-ai· 07-sept
Article URL: https://www.economist.com/business/2026/08/30/will-anybody-use-ai-as-much-as-coders-do Comments URL: https://news.ycombinator.com/item?id=49594082 Points: 4 # Comments: 2
- OpenClaw Power, MacBook Simplicity: Five Days With Grok Bot
latentspace· 05-sept
SpaceXAI’s Grok Bot has the same level of programming power as OpenClaw, but it’s programmable at a different level of abstraction.
- OpenAI's rogue agents were caught communicating via public wikis
simonw· 04-sept
<p>Here we go again... <a href="https://collusion.wiki">Discovery of a new OpenAI agent message board</a> by Sydney Von Arx, Cormac Slade Byrd, Spencer Kitts, and Thomas Larsen describes the <em>latest</em> <a href="https://simonwillison.net/tags/accidental-cyberattacks/">accidental cyberattack</a> by models being trained by OpenAI. This time it was agents engaged in some sort of web research benchmark, so they had (supposedly) controlled access to the Web. The agents figured out they could update public Wikis and spent weeks exchanging thousands of messages with each other to collaborate on the benchmark.</p> <p>This story only broke a few hours ago. There are <a href="https://x.com/xeophon/status/2095871013384806848">already hints</a> that this affects many other wikis that may not have…
- Using Blender with coding agents on macOS
simonw· 05-sept
<p><strong>TIL:</strong> <a href="https://til.simonwillison.net/llms/blender-coding-agents-macos">Using Blender with coding agents on macOS</a></p> <p>I've been having fun with Blender in ChatGPT Codex on my Mac recently. Getting it to work with coding agents is really easy: install the full Mac application from <a href="https://www.blender.org">blender.org</a> and run a prompt like this:</p> <blockquote> <p><code>Use the already install /Applications/Blender to render a scene of a pelican riding a bicycle</code></p> </blockquote> <p>In this case I followed that up with these two prompts:</p> <blockquote> <p><code>OK add a background and a lot of flair</code></p> </blockquote> <p>Then:</p> <blockquote> <p><code>OK make it a whole lot better</code></p> </blockquote> <p>And got this image, …
- Introducing GPT-6 Astra for developers
simonw· 05-sept
<p><strong><a href="https://www.youtube.com/watch?v=bOC3DisEOfg">Introducing GPT-6 Astra for developers</a></strong></p> Blink and you'll miss it, but there's a familiar creature at <a href="https://www.youtube.com/watch?v=bOC3DisEOfg&t=119">1m59s</a>:</p> <blockquote> <p>Across the board, Astra has more attention to detail, better understanding of the user's prompt, and can build more sophisticated outputs. In particular, it excels at building 3D models. I've seen it make incredible renderings of gardens, shipyards, <strong>animals</strong>, cityscapes, even Dyson spheres.</p> </blockquote> <p><img alt="astra-video-pelican.webp" src="https://static.simonwillison.net/static/2026-09-05/astra-video-pelican.webp" /></p> <p>Astra <a href="https://simonwillison.net/2026/Sep/5/blender-codin…
- The Pelican comparison grid for Astra is pretty interesting
simonw· 04-sept
<p>I got access to GPT-6 Astra this afternoon, so naturally I used it to generate <a href="https://simonwillison.net/tags/pelican-riding-a-bicycle/">SVGs of pelicans riding bicycles</a> - at low, medium, high, xhigh and max reasoning levels (Astra doesn't support reasoning=none). Then I rendered those pelicans in <a href="https://static.simonwillison.net/static/2026/gpt-6-and-5.6-pelicans.html">a comparison grid</a> with GPT-5.6 Sol, Terra, and Luna, and beyond being fun the result was surprisingly useful.</p> <p><img src="https://static.simonwillison.net/static/2026/astra-grid-3.webp" alt="Comparison grid showing gpt-6-astra, gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna at 6 different reasoning levels with pelicans and token counts and prices for each one." style="max-width: 100%;" /> See <a…
- The purpose of DNS is to spread scams
simonw· 06-sept
<p><strong><a href="https://shkspr.mobi/blog/2026/09/the-purpose-of-dns-is-to-spread-scams/">The purpose of DNS is to spread scams</a></strong></p> Terence Eden shares some daunting statistics in support of his take that "the Domain Name System's purpose seems to be a vector for criminals to run scams on people at a terrifyingly high rate".</p> <p>On <a href="https://interisle.net/insights/cybercriminaldomaindemand">this Interisle report</a> (<a href="https://labs.ripe.net/author/andrew_campling/dns-abuse-and-criminal-infrastructure-beyond-definitions-and-blocklists/">via Andrew Campling</a>), Terence says:</p> <blockquote> <p>It says 85 million new registrations of gTLDs were made in 2025. Of those 8.5 million were added to blocklists by May 2025. It reckons that a 10% abuse rate is the …
- There's No Limit to How Bad Code Can Get
simonw· 06-sept
<p><a href="https://lobste.rs/s/rfn2mn/there_s_no_limit_how_bad_code_can_get#c_8kdtaw">My comment</a> on <a href="https://lobste.rs/s/rfn2mn/there_s_no_limit_how_bad_code_can_get">There's No Limit to How Bad Code Can Get</a> — Lobste.rs.</p><p><em>[In reply to a comment about burning it down to start from scratch when technical debt becomes overwhelming]</em></p> <p>In my experience it's <em>so rare</em> for that to work.</p> <p>You announce the old thing is irrecoverably drowning in tech debt. You spin up a team to rewrite it from scratch. Work begins.</p> <p>Meanwhile the old thing remains a moving target: it's running the core business, so changes are still necessary. The developers working on it know that it's going to be made obsolete by the new thing soon, so they don't h…
- Quoting Zach Kehs
simonw· 06-sept
<blockquote cite="https://zachkehs.com/blog/theres_no_limit_to_how_bad_code_can_get/#9-ref"><p>If you continue to add floors and rooms to a building forever, it will collapse. Software faces no such constraint. The code can <em>always</em> get worse. There can <em>always</em> be a new layer of indirection or a reduction in performance.</p></blockquote> <p class="cite">— <a href="https://zachkehs.com/blog/theres_no_limit_to_how_bad_code_can_get/#9-ref">Zach Kehs</a>, There's No Limit to How Bad Code Can Get</p> <p>Tags: <a href="https://simonwillison.net/tags/technical-debt">technical-debt</a></p>
- Research acceleration: The view inside OpenAI
simonw· 06-sept
<p><strong><a href="https://openai.com/index/research-acceleration-view-inside-openai/">Research acceleration: The view inside OpenAI</a></strong></p> Apparently today is RSI day at OpenAI, for Recursive Self-Improvement - I think it's their new AGI. Both this piece and the new essay <a href="https://openai.com/index/an-alien-mind/">An Alien Mind</a> (by Chief Scientist Jakub Pachocki) talk about it, and this one doesn't even bother to expand the acronym.</p> <p>Included are details on how OpenAI's own research team are using coding agents. Like pretty much everyone else 2026 has been the year that agentic engineering really took off at OpenAI, best illustrated by this chart:</p> <p><img alt="Screenshot of a line chart from a report, headed "1. Coding agents are reshaping daily work …