But good to know that some things in life do stay constant.

Just wanted to highlight that at least with endocrine data you won't need to wait long to see impact of synthetic participants in the most recent research...

Any chance for pity link?

Saying these somehow contribute massively to reviewer fatigue is just silly, and esp. if you'd try to make an argument that due to inherent conflicts they're actually not reviewed well (and if they are, than presumably they aren't piss?).

I don't buy this connection. You looked at 10 years of data, found 126,000 papers that fit your definition (which is anyway arguable). So that is over some, let's conservatively say, 20 million papers published in that period, so maybe 0,5%?

Just don't kill the messenger when you should've killed your darlings first.

Though I would like to know response rate for this survey, which is just about the only number that is never given in the original report...

It almost makes me want to cry but at this point I'm afraid anything I do will be seen as replaceable by AI, and I don't want to be told by the likes of Altman that I'm not allowed because we have ChatGPT for that. So I guess I'll internalise 7/7

Or perhaps because of the ludicrous timelines which would not make sense anywhere even if they haven't just spend a year first gutting federal workforce and then sending remainder for furlough for months... 6/

Is it - and I find this one quite interesting - because of the call to also rely on proprietary datasets (I look forward to seeing how you get your hands on those)? 5/

Is it because of the demand to secure (which presumably means build) more planet destroying computational facilities to fuel this pipe dream? 4/