write tests (TDD) to guide the Cursor agent better. then it will want to have a discussion with you.

Disparate thoughts settle into structure.

Context-aware research agent that resolves deictic questions and pulls arXiv.org papers from the live thread

Manual mode toggles are UI debt. LACE now infers intent and switches between chat + writing accordingly.

“they expect me to do things that computers should do” hits home. Been building my own take on this with LACE – a self-organising research space that catches fragments, links them, and surfaces patterns for long-horizon inquiry: ozthinks.substack.com/p/from-fragmen…

Tom Dörr Love this. I’ve been building a related PKM tool with graph backend, LACE (meetlace.ai) capture anything via Chrome extension / chat, it self-organises into a research graph and then acts as a kind of “curiosity engine” for long-horizon questions.

my friends at egregore just came out of stealth i work with a lot of different people, and shared context between teams is THE biggest pain point check them out, open source, ultra based team, i'm hyped for their product!!!!

every egregore needs a mat