MeshThink

The best teams often combine people who are quite different from each other. Each person has strengths and weaknesses, and fitting those together to form great teams is a bit like doing a jigsaw puzzle – you find a way to fit one person’s area of strength with the negative space of another person’s weakness.

A more blunt way of saying this is that we all make mistakes. The really great teams aren’t filled with super-human “mistake-less” teammates. They’re filled with teammates who tend to make different mistakes than the person sitting next to them. They catch each other, and the whole is much greater than the sum of its parts.

I see many cases where the application of generative AI to a task has been written off – “it makes too many mistakes”. Fabricating facts, hallucinations, or generating content that lacks the mark of a “real” person. People are waiting for generative AI to stop making mistakes to be ready for use. But to be revolutionary, it doesn’t need to make fewer mistakes than a human. It already does something more important.

It makes different mistakes than a human.

The highest value applications of generative AI won’t be from the AI replacing people, it will be from the AI collaborating with people. Generative AI is good at things I’m not – and I’m good at things that it’s not. Alone, each of us makes mistakes. But together, we can catch each other’s errors and excel in ways neither could alone.

The future belongs to those who see AI not as a replacement, but as a new kind of colleague. Be excited about the potential for human-AI collaboration. Together we’ll achieve more than either could alone.

(Co-authored with gpt-3.5-turbo, gpt-4-0613, claude2.)