Clearspace (#2 on yesterday’s leaderboard) is a screen time manager that conditions access to selected apps on “proof of motion” — squats, pushups or steps. It reminds me a bit of how legendary mathematician John Conway (allegedly) made his computer login a difficult math problem to keep his mind sharp. More generally, I’m excited to see how advanced computer vision will enable all sorts of new technological possibilities.
Cursor, the AI-powered code editor, is a game-changing tool. Period. GitHub Copilot introduced us to the concept of an AI editor, but Cursor takes it to a new level. Cursor’s recently added composer(CMD + I) and tab autocomplete tools are now indispensable to my workflow. They take some time to fully grasp, but are incredibly powerful once you get a feel for them. Though o1 is an option, I still run it on Claude 3.5 Sonnet — it’s fast and gives mostly good answers.

You A/B test product features - why not test your body? Pulse lets you run experiments on your sleep, activity, and vitals. Wondering if late-night calls kill your recovery? Or if morning cardio sharpens your 2PM focus? Track inputs, measure outputs, and find your edge. It’s the insight engine for founders and operators who treat themselves like a system.
5,000+ high-performers are already running their own experiments. Start testing what actually works for you: pulse.site
LlamaCoder is the latest web app that lets you generate entire apps from natural language prompts. As prompt-to-create gets more powerful, I’d argue that it will actually become more important to understand how to modify the output. Prompting AI to create an app will get you started, but unless you can actually understand and tweak the code, you’ll be stuck with a mediocre product. Gen AI is a complement to knowledge, not a substitute for it.