It doesn't matter how "hard" you work. Or how wrinkly your brain is. Nobody gives a shit if you're "actually smart". People only care about results.

If the "result" of using LLMs is that you write better emails, then, no, you're not getting dumber. You're getting better at communication.

If the "result" of using LLMs is that you back your arguments with appeals to authority, i.e. "I'm right because GPT said so" (a logical fallacy), then maybe you are getting dumber.

"Computers are a bicycle for the mind" -- Steve Jobs

If the result of using a bike is traveling faster, and you use that result to go to more places and get more cardio, then a bike is making you healthier.

If the result of using a bike is traveling faster, and you use that result to avoid cardio and "waste" any time that you've saved while traveling, then a bike may be making you sicker.

One of the earliest critiques of new information technology comes from a myth written by Plato:

"This invention [written language] will produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them." — Plato, Phaedrus

Large Language Models are part of the next technological evolution of human language.

Spoken words enable people to transmit information in real-time.

Written words enable people to transmit information into the future, without relying on human memory.

LLMs enable transmission of information into the future, without relying on human memory, AND without requiring another human to assimilate the information verbatim.

The result of written language may have "weakened" our ability to memorize. Just like LLMs may weaken our ability to assimilate information. But the result of writing means people can "record" more information than they could without. And the result of LLMs means we can assimilate more information than ever before.