How dare they use data to evaluate AI when a negative narrative works just as effectively.
I recall a study where they trained 100 pigeons to categories categorize cancer in tissue samples and the average of all their responses was better than a single oncologist.
I’m probably not going to consult a flock of pigeons though for important things anytime soon though
Press X to Doubt
Yeah, who are you going to believe, a peer-reviewed paper in a scientific journal or random people on social media?
“Peer-reviewed papers in scientific journals” are not the air-tight source you seem to think they are. A lot of them have become echo chambers that largely seek to only acquire more citations for the sake of having more citations. I’m not saying they all deserve to be thrown in the trash or anything, but likewise I’m also not going to blindly believe whatever they’re/you’re telling me just because they’ve got “Journal” on the front cover. Here, check out this blog post, it goes into some of the history of Academic Journals and how they’ve fallen from grace over the last several decades. Oh also they’re quickly becoming inundated with AI, both in terms of submissions and even the review process itself, and personally I’m not going to let an AI tell me if AI is good or bad
I think the point isn’t that the journal is factually incorrect, rather an industry mired by waste isn’t redeemed by one useful thing.
Wouldn’t have useful things instead of vapid chatbots make it less of a waste?
No, it still wastes. Asbestos is a good insulator, but that doesn’t make its cancerous properties less bad.
Here’s the actual papers, published in Nature:
