🚨 New ResearchOn SWE-Bench Verified and Terminal-Bench 2.0, top coding agents perform pretty similarly.But if you give them an unfamiliar programming language the gap gets huge, and the agents that adapt do it through metaprogramming. 🧵(follow up of our Esolang-Bench)

Six words hit two million views on X, and most people boosting them couldn't explain what they meant."you should be designing loops that prompt your agents."i'll be honest, i read it three times before i understood why it was different from what i was already doing.The whole point is that you stop being the thing inside the loop typing prompts. you write the loop. the loop prompts claude. the model becomes a subroutine.The person who built Claude code deleted his IDE in November 2024. hasn't ope

Cursor CEO Michael Truell on the future of writing code: "Our goal with Cursor is to invent a new type of programming.""It looks like a world where you have a representation of the logic of your software that does look more like English.""You can imagine kind of an evolution of programming language towards pseudocode. You have written down the logic of the software, and you can edit that at a high level.""It won't be the impenetrable millions of lines of code, it'll instead be something that's m