🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Don't Undervalue Ideas: The common narrative states that ideas are cheap, but the reality is that most people have never taken the time or done the work to actually know how to value them. Good ideas always seem obvious in retrospect, but are often the result of years and years

I wrote some thoughts on designing and building stuff, from doing it for 1½ decade. Happy holidays! ryolu.notion.site/how-to-make-so…

A few reasons why I'm really excited about this project!

Read the full deep-dive on why visual reasoning is the next frontier in AI: arcturus-labs.com/blog/2025/03/3… And follow for more.

I think many people don't use custom Cursor rules because they don't know where to begin. I'm experimenting with a set of "metarules" which I can use in chat to generate custom rules for any project

How should we update AI memory of user intent as intent evolves? 🤖💭 How can we auto-update memory docs like Cursor rules as a user interacts with an AI agent over time? Check out our pre-print: Semantic Commit: Helping users update intent specifications for AI memory at scale!

This should be a really fun talk, Shawn Simister and I have been recalling to ourselves how it all worked. What we realized is that we were doing an early form of LLM-as-judge back in early 2023. maven.com/p/da8264/how-e…

Our talk on GitHub Copilot evaluation starts in less than an hour. Don't miss it!

In 1st- and 2nd-gen AI coding tools you now see a proliferation of modes, toggles, and buttons. This feels off. AI UX done well should mean we don't have to point/click or get lost in a sea of rules or modalities. Sure, you may still have to type or articulate your intent, but

day 1 building the IDE of the future: + TS interpreter that controls the UI via scripts + REPL for the interpreter with that every action in the IDE will be: - scriptable - promptable - tab tab tab - ctrl-z & ctrl-y every workspace: - git ready - forkable - deployable

I did a fun thing. 🥳 - I vibe-coded a CLI tool that summarizes YouTube videos - recorded myself making the said tool - used said tool to summarize the video of me making said tool - and finally – dumped the summary into an LLM to draft a blog post LLMs all the way down 🤣

# better diff views for AI agents +1 to the quoted post. To work well with AI agents, we need better diff views! your ability to check the agent's work is the bottleneck. here are some ideas for how to do that, based on our research at Ink & Switch : Zoomed out diffs: show

I've been building my own version of this recently and it has really helped me focus on what I want to build instead of wasting time on scaffolding and environment set up

Uploaded this talk with Vincent D. Warmerdam "Build your own Eval tools with notebooks" Our hot take is notebooks are the most useful tool for evals. Also, Marimo'a reactive, app mode, etc seem useful in certain scenarios. More links in reply youtu.be/aqKUwPKBkB0