Veganaise
June 2026 — The gist: Depending on how you think about ethics in design, you should consider either 1. Being AI vegan, or 2. When using AI, first seek human knowledge, avoid vibe coding, and fall back to a copy/paste workflow. Both paths will help you keep and grow your hard-earned skills.
I promised myself a year ago to only publicly share practical and brief thoughts regarding AI. (Thank you, friends and family, who listened to everything else).
I finally have a tidbit I hope is useful.
I find myself confused about the right/good/best way to go about what I've always aimed to do: make well designed digital products for diverse users. I remembered Mike Monteiro's point about the impossibility of a well-designed gun—“If a thing is designed to kill you, it is, by definition, bad design”. Scott Berkun critiqued—“This sounds powerful but it makes little sense because it pretends design and ethics are the same and they're not.”
Design + ethics, or design vs. ethics? I think however you lean may also help guide your AI path and stop second guessing the following:
- Do Google AI snippets count as using gen AI?
- Does copy/pasting code count as vibe coding?
- How much do I need to understand the design or code my AI tools build?
- Should I abstain from AI, period?
- What’s going to happen to my hard-earned skills?
My friend David Godshall and the team at Terremoto recently staked their AI position, and perhaps tongue-in-cheekily described it as “AI vegan”.
They lean design + ethics. If you do too, I think you should avoid any use of AI altogether. It'll be trickier than being vegan-vegan because AI is hidden everywhere (including soon Siri). But you'll surely get better and set a great example for like minded peers. (I'd love to learn from you).
For example, adding “-ai” at the end of your Google search currently filters out AI snippets. You can also turn off Apple Inteligence on your iPhone and Mac.
In your workplace, being AI vegan may not even be allowed. I think that's a shame and I'm sorry.
You might lean design vs. ethics. In other words, good design is entirely separate from what's morally right or wrong. In that case, I still think you should use AI as little as possible because it poses a long term threat to your own skills.
After a year of observation in the corporate environment and earnest attempts at using Cursor and Claude Code myself, my best advice is to keep AI generated anything as far away from your work as possible. Consider only asking AI basic questions and then doing the rest yourself.
For example, getting a code snippet from Google AI search results still pushes you to think of the concept yourself and copy/pasting the code into the correct place in your source file. You've still got an understanding of that bit of code; a gift you get to keep.
But if you ask Cursor to generate a thing that works (vibe code), even if you go back and study it, that process pales in comparison to how much you would've learned naturally if you'd just used Ask mode only.
I can't make a coherent argument that AI use sharpens skills more than learning from a human. I can only think that AI could fill voids where humans or documentation are lacking (e.g. CloudKitJS). Grow and keep your hard earned skills by learning from humans as much as possible. Email me any time.
I'll add too: solving bugs build skills. AI tries very hard to eliminate them on your behalf.
To recap: Consider being AI vegan, or alternatively, seeking human learning as much as possible and embrace a copy/paste workflow. You'll keep your skills far better than offloading your thinking (and mistakes).