I’ve done two podcast episodes on large language models. I recently spun up and installed a large language model locally on my computer and built a project around it. I’m an avid user of echoes from Amazon. I have had Grammarly kick in to spellcheck for me. I tell my phone to add an appointment to my schedule all the time.

And I have a dire warning for you.

A recent MIT study proved something that I have been thinking for quite some time. Using.EEG (brain scans.) to provide data, the study showed that those who relied on ChatGPT for ideas and help with papers tended to not even to remember most of what they wrote. A lot of them didn’t even properly remember the titles. The bottom line was that they’re deeper learning wasn’t being engaged and when it wasn’t being engaged, it meant they weren’t using it. And it was getting weaker.

AI, the firming paradox and the end of the world

What science fiction and many people tell us is that AI will become sentient, judge mankind, find them wanting and wipe them out. And that will be it for us. But I believe it will happen in a very different way and is already happening.

But we can stop it.

Consider what I am suggesting, reflect on your own use, and read the MIT study for yourself.

Just don’t use an LLM to summarize it for you.