Thanks. IMO the most fun is in the CAPTCHA, which turns on its head the Turing test.

But it goes even further than their greCAPTCHA and it solves their consumed time problem.

Indeed, in their proposal they have a human bottleneck, but in the june 2026 proposal is suggested that one could use an AI to generate the results without the knowledge of the submitted article.

If the AI can generate a pretty close result, with the article fed gradually as a prompt, then reject.

And even further, that it might be not even a need to publish anymore.

Just use the article for training and make a public database with some numbers about the successful researcher, where we see an influence score (how many times an idea from an accepted article are used by other accepted articles), a publication score (how many articles the author had).

I wonder if the reality will be more or less surprising, my bet is on "more".