Write software @buzzsprout.
Ruby on Rails believer.
Foster care advocate.
Interested in leadership and design.
Next week is where dreams and money go to die.
@rywestlake@jasonfried Their nav is behind the basecamp logo at the top. They also put nav like items at the very bottom and corner. Its more spread out than a typical app.
This makes me think Im doing it all wrong. Instead I wonder if I should keep a local copy of fizzy and tell AI "refactor this controller so that it looks like one from fizzy". Now AI could have a much better context than even I might be able to communicate.
"I avoid prompting at all costs."
That's @jameygannon, a brand designer and AI creative director who has spent thousands of hours in @midjourney building a system that lets her create consistent, beautiful brand imagery with AI -- without complex prompting and srefs.
In this ep you'll see:
- how @Pinterest mood boards communicate your aesthetic to Midjourney better than text
- how to build a personalization code so Midjourney learns YOUR taste specifically
- her @floraai flow for fixing and delivering final assets to clients
This is by far the prettiest episode of How I AI, and has already changed how we make thumbnails for the podcast.
Thanks to our awesome sponsors
๐ซ @TrustVanta - Automate compliance and simplify security
๐ @Lovable - Build apps by simply chatting with AI
Watch now: https://t.co/sHp36q2hpP
@hamids@RJScaringe@buyholdrant@dustinalper How was the video made? Was there a lot of editing?
Just curious how people are creating professional videos these days, especially if they are easy to create.
How are people maintaining AI code?
It can be tempting to let it write a bunch of code and just have it figure it out. Especially when it builds so fast.
For an MVP, sure. But for production code you want to maintain and understand, I have to review it all.
Are you maintaining it caring about the details or just letting it run? I canโt imagine just letting it run for anything real and long term.
Your AI can write 500 lines of code in a minute. Your team can review maybe 50 lines with full attention. The new bottleneck is not generation. It is verification. Staff accordingly.
Stop teaching your juniors to write code. Start teaching them to read it. AI writes plenty. Humans who can spot what's wrong will run circles around those who can only prompt.
The teams that are actually shipping with AI agents share one trait: they made their environments disposable. If an agent can break things beyond repair, you haven't set up the right guardrails. Confidence to let it run comes from knowing you can roll back.
Convention over configuration was always about developer happiness. Now it's about AI effectiveness. Rails apps have smaller context windows, clearer patterns, and less ambiguity. The framework you picked for productivity ten years ago is now the best choice for AI-assisted development.
Your job as an engineering leader is not to protect your team from AI. Your job is to make sure they are the ones directing it.
The managers who survive will be the ones who taught their teams to think, not just code.