I regularly browse HN through tor. As a tor user I'm used to having to solve CAPTCHAs everywhere I go, as there's always some assholes using tor for abuse.
In the past couple of days, the CAPTCHAs for HN got way harder, to the point that I can't solve them. It used to be two words where one was an easy control word you could skip, the other a semi-hard word that I'd get ~70% success rate on.
Now, both words are equally hard with additional distortion, and my success rate approaches 0%.
Are CAPTCHAs really still workable if this is what's needed to keep the bots out? If humans can't solve the problem, it's probably not a good metric to determine whether someone is a human or a bot.
I don't have any good solutions here, but this is very frustrating and I'd love to see some discussion about alternatives.