All good and great but I think it’s a bit misleading. The participants were asked to do LeetCode—which limits this study to the suggested effects of cannabis on arbitrary and abstract programming tasks.
Most developers mid level and upwards from majority of the companies I have held a role in are not spending the majority of their day coding, they're writing specs, doing meetings, providing guidance to their junior counterparts, assisting in support tickets on call, and if you had time, documentation. You were quite lucky to meet 30% time in any day/week/month/year spent writing code. The vast majority of developers were also not developing new, foundational code either.
So I disagree, its not misleading, its actually relevant, and arguably even more mentally stimulating than all the cruft that you actually do every day. At the least you're reusing and thinking about all the DSA knowledge everyone has lost touch with from decades ago.
It's orthogonal to your day job. The "cruft" you talk about is how we remove ambiguity and fill in details so that work isn't a series of free-floating gedanken experiments. If you're able to apply codified knowledge like LeetCode to a programming task it means someone else has already done the heavy lifting.
We "lost touch" with DSA knowledge because it turns out that it's easier to work things up from base principles using the scientific method than memorizing crap that worked for the vacuum tube generation. We don't *need* a reductive LUT approach to the process anymore.
The research performed in this study is relevant, and no one so far has demonstrated that it is "misleading", "bad or poor quality research", or really anything to demonstrate to me that it's results are questionable, going back to the parent thats what I address.