Keryn Johnson Thanks for your thoughts. What you said to be your approach where you use social media to share new ideas and results are fine - in fact, they are also used by many scientists today (Twitter, Mastodon, etc.) to advertise their work. You can always advertise things before it is peer-reviewed, and possibly gain feedback from it. So "online community to bounce ideas around" is itself a nice idea (subject to some complications, if properly handled). However, to be part of an established knowledge, the argument is that one does need certain degrees of gatekeeping, since scientific work must go above being merely an opinion shared by many.
I do foresee ChatGPT and AI being useful for such a gatekeeping work, but at the moment I need to see more development in those directions in order to understand how they would be useful. For example, I think they would be very good at analysing potential mistakes or summarizing the core arguments, especially for extensive/long papers. What needs to be avoided is to simply leave them completely to ChatGPT or AI, unless, of course, that they would look very different from what they are today. Either way, I agree with you that I foresee their use in peer-review in possibly near future, and there is a need to ensure that they are used appropriately.
