Is using AI good or bad for us?
A super useful framework for thinking about this comes from Jake Hofman, Daniel Goldstein, and David Rothschild’s 2023 HBR article, A Sports Analogy for Understanding Different Ways to Use AI (author-hosted PDF):
- Steroids: Better performance now, but your skills get worse over time.
- Sneakers: Better performance now, without changing your underlying skills.
- Coach: Better skills over time, even without AI.
The same tool can be used in all three ways. So the question isn’t whether AI is good or bad for us. It’s when you want it to be steroids, sneakers, or a coach.
I use it to give me feedback on lectures. I also use it to speed up writing, and I’m still working out when that’s sneakers or steroids.
AI is lowering the effort many tasks require, while changing the return to the effort we still put in.
My hypothesis: people using AI as a coach are accumulating skills faster than those using it just to get things done. That gap may become a new source of inequality.
When do you find yourself using AI as a coach?