My thoughts on stuff. More about me.
The serenity prayer asks for the serenity to accept things one cannot change, the courage to change things one can, and the wisdom to know the difference. Often, the wisdom seems to be the hard part. But it doesn't have to be.
New Leverage post on Substack. AI is both equalizer and differentiator.
The rate of new model releases at the moment is dizzying.
Literally lol and dancing around my office after Omarchy agent helped me cook up a neat little publish app for this site. It’s a native Omarchy app with keybinding and all. Idea -> build -> done in minutes. Welcome to the future.
I’ve been increasingly using Omarchy Linux over the last couple of weeks. It’s awesome. What started as an experiment has turned into my daily driver. Check out DHH on Lex if you’re curious.
I made a little tool to help review responses to facilities management RFPs. It’s called OpenRFP. It’s free and open source.
Use your favorite RSS feed aggregator to keep up with the latest posts from this site.
Got mobile git set up via Working Copy. iA Writer, pointed at this site’s local repo clone, is the perfect post editor.
On a project I’m working on I reached a point where it appeared I needed to do a bunch of dashboard work to unblock Claude. I didn’t want to do that.
So I essentially asked Claude, “how much of this can you just do?” and after contemplating for a bit Claude said it could do basically all of it, presented a detailed plan, and executed it.
So much delegation fails because the person delegating simply accepts the responsibility to do the work themselves too soon. This applies similarly to other people and to agents.
I experimented with using Claude Code in the terminal for the first time today. It was pretty OK.
I’m using Claude Code in the terminal today. It looks like computer stone ages (minus the part where the computer talks). But my intuition is that it’s better.
More on the rogue agents at OpenAI. This is wild.
The most critical AI prompting skill may be knowing when it’s time to step away from the AI.
Seeking a more future-proof format
The breakdowns of the recent ChatGPT/Hugging Face incident on The A16Z Show and AI Explained are worth taking in. AI2027 continues to feel eerily, concerningly prescient.