Discussion about this post

User's avatar
Juan Mora's avatar

This might be the most useful piece in the series, because it answers the question the Jobs articles leave open. Roles are legible. Anyone can read the five and recognize themselves in one. Whether you can actually hold it is a different question entirely, and this is where that gets answered.

The chicken farming study is the part I keep turning over. The AI wasn't wrong. It gave reasonable advice that didn't fit. The people who benefited already had the judgment to know which suggestions applied to their situation, which means AI widens the gap between people rather than closing it. The tool amplifies whatever discernment you bring to it, including none.

I'd argue empathy is the highest form of that discernment. Not the soft version, but the version that lets you model a situation you're not currently standing in. It's what makes a conversation worth more than a search query. Someone who asks a real question and actually listens to the answer walks away with context no retrieval system could have produced, because the useful part was never written down anywhere.

On agency, and the part I think people get wrong: potential isn't a possession. There's a line I return to often, from a bodybuilder named Mike Mentzer: "Potential is only the expression of a possibility, something that can be assessed accurately only in retrospect. In other words, you'll never know how good you might have become, unless you try." Which means agency isn't about believing in your potential. It's the willingness to convert it into something checkable, and to accept the answer if it comes back lower than you hoped.

Taste is the same shape. It can't be prompted into existence because it's the residue of choices you actually made and lived with. AI can generate a thousand options. Knowing which one is right requires having been wrong before.

Ready for more?