Regarding signups/spam

[Charon]

Just be aware I've changed the registration settings in the hope of making it easier for humans and harder for bots. If this works, I won't have to manually delete as many spambots and humans won't have to worry about being accidentally deleted if their username or email looks sus.
However, it's possible some spambots might get through. If they do, don't worry, I'll clean up and revert back to the previous system.

[Charon]

Six days with no spambot signups. I've temporarily disabled admin confirmation of new signups, so humans should be able to start straight away. If any bots manage to get through I'll tidy up and revert back, but for now I think I've bamboozled them. :P

[Grodech]

Until they find a new way around them....

Good work, though.

[Charon]

I think it might work for a while, because from a bot's point of view, it isn't obvious what I'm referring to in the question and isn't something they can guess. And they can't dictionary attack it because they get timed out after a few wrong attempts. Whereas for a human, it's pretty bleeding obvious.
:D

[Grodech]

So of course, it now seems that AI can be used to bypass that now. So good of them to state that in their signature.

[Charon]

Interesting. I'm sure I can come up with something. For now it will still be blocking the vast majority.
It's also possible that was a manual signup. XD

Edit: For a little context, if you're interested. I currently work with AI and it's nowhere near as clever as people think it is. I'm confident it can't solve the current question, though it can make an educated guess if it's able to pull context. I have a backup question that is far more difficult both to solve and to guess and a third which is harder still but also trickier for humans. I'm saving those until we see a significant influx past the current method.

Well, it's possible the word was guessed and is now logged somewhere so we're on question two. It's trickier so hopefully it won't prove too hard for humans.

[Charon]

Just an update to this.
There's an annoying thing now where rather than using AI to defeat the countermeasures, companies will pay people in poor countries to simply go from site to site answering them and making lists of the answers that can then be sold on to the people with the bots.
We had one such manual signup yesterday, meaning the current question is technically solved although I've yet to see any further bots bypassing the question.
In anticipation, I've turned on first post moderation. Simply put, the bots can sign up but not post.
So, you won't see any spam on the forum, but you may notice new members that are obviously bots. I'll delete them in due course, no need to report them, but it's easier for me to do this in bulk, so depending on how many (or few) there are It might not happen straight away. Once I've assessed it and it becomes clear they are bypassing en masse I'll attempt a new question, though again, it only takes one human to bypass it so there's a limit on how secure these can be. Blows my mind that in this day and age, spamming forums is still profitable enough to waste money on people or AI in this manner. XD You'd think with so few forums and lower memberships of generally older, savvier members, there'd be no money to be made and yet here we are.