Wisdom for the past and future self
Welcome to fall, friends. Not you, Texas. Sorry that summer ends on Halloween now.
Let me tell you what’s in my head lately: wisdom I would impart upon my past self and ideas that might help my future self.
Things I would tell past versions of myself:
- The best proxy for success is: how much do you feel like you’re making something that hits the mark on feel, responsiveness, function, and form?
- When you can’t surprise and delight, can you at least avoid surprise and dread?
- When you’re in trouble, seek a novel solution rather than trying harder. That is, dig out and not up.
- Coding and writing and working with people are the most important. But finishing is a skillset you can always develop and will never regret improving.
Things that really stuck in my head after reading, in a great way:
AI as teleportation – software starts as technology, ends up as part of culture, eventually becoming invisible, but ever present.
Another example. One of the great joys of my life is having nerdy friends explain things to me. Now I can get explanations from AI with less friction, anytime, anywhere, with endless follow-up.
Even if the AI explanations are “better”, there’s a social cost. I can try to mindfully nudge myself to still ask people questions, but now it requires more effort.
What I think about when I think about Claude Code – easy, like Sunday morning. (Not quite, but wouldn’t it be lovely?)
No, spending a morning coding with Claude Code is different.
You just loop one minute composing a thoughtful paragraph to the agent, and three minutes waiting, gazing out the window contemplating the gentle breeze on the leaves, the distant hum of traffic, the slow steady unrelenting approach of that which comes for us all. … Writing that “thoughtful paragraph”…
The trick with Claude Code is to give it large, but not too large, extremely well defined problems.
Mark Wunsch, The Greatest of Ease – coding agents as enabling collaboration and stepping out of your comfort zone.
More broadly, I think AI is flattening the landscape between technical stacks. The gaps that once felt like cliffs — the difference between writing Ruby, or Swift, or TypeScript — now feel like small inclines. With an ever-present assistant, the cost of switching is lower, and the fear of getting stuck is gone. If the stack turns out not to be the right fit, it’s less daunting to revise.
The takeaway: if you’re building today, give yourself permission to step outside your comfort zone. Pick up a stack you’ve never touched. AI can absorb the drag of getting started. What’s left are the parts that really matter — the architecture, the modeling, the craft of building.
Writing Mac and iOS Apps Shouldn’t Be So Difficult – in a quite different time in computing history it was feasible to build a desktop (native!) application with a core in C and the UI/behaviors in a (bespoke) scripting language. Even when computers weren’t yet Quite Fast. If the lines between the widgets, database, and data structures were drawn in the right places, this made for a pretty good developer experience. I.e., the iteration was quick and users liked the results!
We do seem pretty far away from these days now, eh? The discussion on Lobsters is pretty excellent, should you want an even more belt-and-suspenders perspective.
I hope you find one or more of these helpful, if not insightful!