The danger of cognitive surrender
How much should managers let bots do the thinking
Login to view embedded media
Here's a link to the paper for you nerds who actually want to read analyze it:
And for the cool tomato in gumbo type kids, here's a perplexity AI summary of the paper:
I'm glad that this is being written about. Not lying, as close as I am to this industry and for as much as I use AI, both in coding and writing, I absolutely think very differently than I did prior.
Some of it is in a bad way as this would suggest... it's easy to get lazy. I find when I'm coding / vibe coding I'm really lazy like that. If it doesn't run, it's way easier to fix in AI than it ever was when you had to bug trace.
Written word, however, I find that the benefits are that I spend less time thinking about how I'm going to say something and far more time thinking about the facts and information I'm trying to convey. Not infrequently I pass my AI-co-authored writing through a secondary model to validate and verify (e.g. work in Claude and validate with Gemini). Using a secondary model expressly for critique and a devil's advocate opinion is hugely important.
Ahhh... Fuck... who cares... this will all be over soon enough with impending retirement.