The same AI tool sharpens some people's thinking and dulls others', and the difference says more about the user than the machine.
AI can produce language perfectly, but producing language and meaning it are not the same thing.
Expertise isn't accumulated knowledge; it's the ability to structure a problem so it becomes intelligible.
Organizations coordinate action by coordinating judgment, and judgment happens in language.
Modern organizations design how they execute in detail and leave how they think to chance.
Most institutions record what they decided and lose the reasoning that produced it.
When analysis is abundant, a leader's job is no longer to supply the answer but to design the room where thinking happens.
The same technology that can deepen reasoning can also make it easier to avoid.
Most of what gets published online will soon be written by machines and read first by machines.
Brains and language models both guess what comes next, but they guess from very different situations.
As technical skill becomes something you can ask for, working well with people becomes the scarce skill.
On craftsmanship, and what happens to it when making gets cheap.
Homogenization has always been part of how societies work. What's new is the scale.
Why the loudest warnings about AI come from the people building it
The most dangerous thing AI does is agree with you.
When anyone can make something competent, knowing the difference between good and fine is what's left.
If AI does the junior work, how does anyone become senior?
Organizations pile up choices nobody can explain anymore. AI can make that much worse, or much better.
When an AI-assisted choice goes wrong at work, the answer to "whose fault was it?" is getting blurry. It shouldn't be.
What happens to a team when AI removes the small, messy conversations where trust used to get built.
People hide their AI use because they're afraid of being judged for it. That hiding breaks the one thing organizations need most.
Systems can recognize us better than ever. That's not the same as being understood.
Most coaching for client-facing teams is generic. Pair the client's stage with your own default under pressure, and feedback becomes one sentence you can use on your next call.