In this episode of Plain English, Derek Thompson and Cal Newport spend over 40 minutes discussing the effects of using AI and its detrimental effects on cognitive fitness.

Towards the end of the episode, Thompson feels the need to inject some positivity about AI to counter Newport’s more sober analysis. He says this:

It does some things spectacularly well. If I need a 10,000 word essay about Hoot Smalley tariffs in the early 1930s to understand the Great Depression tariff policy from several different angles, I can whip it up in 5 minutes and that’s pretty extraordinary compared to how long certain kinds of research would have taken me to go the library to track down a book and everything else.

This was exactly what they were talking about as a poor use of AI as it will diminish cognitive fitness. His understanding of the Hoot Smalley tariffs will be limited to the output he receives from the chatbot.

AI does produce BS spectacularly well. Unfortunately, it is unreliable when it comes to factual information. He may feel his understanding of this piece of history will improve but how will he know if the essay is accurate?

Scott Hanselman had a great quote about this approach towards using AI on the Fork Around and Find Out podcast.

AI is going to make the incurious less curious.

Derek Thompson knows that the output of LLMs is unreliable yet he still struggles to let go of this fantasy. This mindset values quantity over quality. It is performative. It gives the illusion of effort. It is careless.