I actually think that research is happening it's just not popular on Bsky

The strongest claims in the press release are direct quotes from an author. I don't appreciate this shuffling off that academics like to do where they write a paper, do press for it, then act angry that people read what they said and wrote.

"we show that AI interaction can result in significant impairments in independent performance and persistence – capacities that are foundational to life-long learning" significant impairment in the context of claiming you're doing a cognitive study! The word impairment has a clinical meaning

In fact "erosion" is used in paper: "no prior work has examined how AI use erodes cognitive processes and undermines autonomous thought. Our work addresses this gap directly, providing cognitive evidence for a “gradual disempowerment” hypothesis."

Paper: "raising the question of what prolonged AI use does to human capability over time." "we find that AI assistance impairs independent performance and reduces persistence" Impairs, capability.....pretty much matches press release language "erodes and ability" imho

Obviously as I have made clear in my threads I don't agree with the argument of this paper (and in fact wrote the same when the preprint popped up the first time), but in terms of the press release basically doing the job of sharing what researchers are arguing? It is.....?

I've read this study about five times and I can't see what the press release is supposed to be misrepresenting. It quotes the authors multiple times

Love it

Is us all doing interactive online experiments for absolutely no reason but the love of the game the cognitive gym I heard we needed

Genuinely very impressed that so many of you have taken this and some of you are doing bonus questions?????? CRITICAL THINKING BONANZA over here

Something can be sensitive to effort without being a good measure OF effort. You have a lot of other things to prove on that road if that's the story.

Low math scores can tell you something got in the way of good performance. But there are lots of answers to that question, and they're not all cognitive ability answers. They can be test condition answers, anxiety answers (also drains your working memory), not knowing what the test wants

thank you!

Because so many people have poor math self-efficacy, it can actually be an area that's highly responsive to psychological intervention (when you have nowhere to go but up...). Many studies in psychology use math problems as a mild acute stressor!

I find it somewhat disheartening to see the rise of math quizzes being treated as pure and simple measures of cognitive effort, cognitive ability, and "good thinking" with AI. It's easy to implement a math quiz and easy to grade it. That doesn't mean it's a good method for all questions.

But it's not magically objective. Math performance in a constrained test is a kind of performance; performance is subject to many effects. It's shaky when measured with inaccurate assumptions about people, or uneven test conditions.

Using math quizzes to attempt to estimate cognitive ability or aptitude is nothing new. I have a copy of the IBM programmer test which did so for "programming aptitude" (kind of a made up idea at the time really) for instance. Math is often chosen because it feels very objective to us.