

Marc Hatton
1.9K posts

@marchattonhere
🎙️ 20 min guest demos of LLM use cases :: B2B product hobbyist :: Long suffering @arsenal fan :: Dad



Meet the new Stitch, your vibe design partner. Here are 5 major upgrades to help you create, iterate and collaborate: 🎨 AI-Native Canvas 🧠 Smarter Design Agent 🎙️ Voice ⚡️ Instant Prototypes 📐 Design Systems and DESIGN.md Rolling out now. Details and product walkthrough video in 🧵

Cursor, Codex and Claude Code are all single-player. Your whole team builds alone and no one knows what anyone else decided. But building product is a team sport. AI should be too. The conversations, decisions, specs and builds. All of it, together, with your whole team. Launching soon → somehow.sh


late night pet project the other night forking @karpathy 's autoresearch -- basically something interesting here where it's like a swarm of ralphs each with a compass that eventually leads you to somewhere interesting robc.substack.com/p/welcome-to-t… github.com/kousun12/darwi…









Mole 1.29 is live. The Mac cleaning tool that can free up tens of GBs in one go. github.com/tw93/Mole What’s new: · destructive commands now support --dry-run / -n for full preview before execution · mo status and mo analyze now support --json, with auto JSON mode in non TTY pipelines · smarter cleanup for Xcode DeviceSupport, plus safer sudo handling in Automator and Shortcuts This release makes Mole much better for automation, scripting, and safer cleanup. If Mole helps, I’d love to hear what hidden junk or safe cleanup targets you want next.



Hate to be this guy, but is anyone else finding that GPT-5.4 feels noticeably dumber today?

I feel like gpt-5.4 writes way too much stuff out compared to other models. Not a huge fan :(