The verification screen is fine. I guess the bigger issue, long-term, is that with "under attack" mode, all bots, including well-behaving, get blocked; correct me if I'm wrong. Being Google-searchable is a big benefit; having allowed AI models to be trained with the great wealth of information has benefited everyone - and the AI providers offer that information through their AIs and own their server cost, so it's one-time, relatively small crawl cost for the site. I'm under the impression that Dave sees this information crawling for search purposes exactly a positive thing. Not everyone agrees, of course, and some sites want to prevent bots completely.