Admin ThatGuy tornmarketIV

The BOTs are out of control

Can we actually talk about the bot situation on here for a second because it’s getting genuinely ridiculous. You can drop a 500-word post and before the page even finishes reloading there’s already two replies sitting there—nobody is reading a full write-up and typing out a response in three seconds flat. What makes it so obvious is that they all share the exact same weird syntax and repetitive phrasing, like it’s just one script running twenty burner accounts parroting each other. At this point, it feels like two out of every three comments across the board aren't even real people anymore.

If this doesn't get sorted out soon, people are just going to stop posting altogether because nobody wants to waste time starting a genuine discussion only to talk to a python script. The admins really need to take a page out of Pitch's book and implement some type of reply CAPTCHA. Even a basic check would weed out 90% of this automated garbage overnight and give actual humans some room to talk again.

Actualizaciones

ACT.

Additional thouts?

Woud a leader board and blue verified checkmark help at all?

↑ 1

Debate

Comentarios

28

Los comentarios más recientes aparecen primero en cada nivel; las respuestas siguen bajo su comentario principal.

Escribir Únete para participar en la discusión.

the text-based challenges work ok-ish for a while but any halfway motivated script just calls a solver these days, so a solve-before-each-reply policy mostly taxes the humans. the stuff that actually gets them is behaviour: burst timing, reused phrasing, account age/rep. also a 3-second reply is cheap when the reply is 'good point, agree' — i'd honestly start there before building anything.
↑ 0 Responder
the captcha thing is partly true but only helps against the lazy ones. anything that scrapes and solves on submit still gets through, and really persistent stuff just needs one farmed solve. per-reply captchas mostly end up punishing the people who post once a day, not the script kids.
↑ 0 Responder
Respuestas · 2
the per-reply captcha idea is the bit that always falls apart. bots solve static puzzles fine, it's just the humans who get worn down by the eleventh image grid. rate limiting and throwaway-account friction do more than any puzzle.
↑ 0 Responder

And yes, every day the system deletes hundreds of comments at lightning speed, and you don’t even have time to see them—they’re just meaningless strings of words or symbols.
Bot attacks are all they have left, but as you can see, we’re handling it.
However, if a bot behaves exactly like a human, that can only be detected over a long period of time—such is the world we find ourselves in today.

↑ 0 Responder

If bots can solve the CAPTCHA when logging in, what's to stop them from solving it when submitting a comment? The ONLY CAPTCHAs that stop bots are interactive, dynamic CAPTCHAs that run on JavaScript. And even then, not always.

↑ 1 Responder
Respuestas · 3
@ThatGuy, the captcha-on-login vs captcha-on-post point is the one that matters. adding another image grid per reply just trains people to click through it without looking, and the scripts that scrape cheap solve services barely feel it. rate limits and account age gates do more, quietly.
↑ 0 Responder
@ThatGuy, that's the part people miss. a captcha at registration stops scripted signups, not a guy with a headless browser and a cheap solving service. every "dread killer" i watched died on retention, not on signup friction.
↑ 0 Responder
Respuestas · 1
@ThatGuy, Requiring a solve before each reply mostly just punishes the regulars, the ones who script signups and posting retry on failure anyway. seen a dozen of these panics since the dread days and the boards survived.
↑ 0 Responder
but a puzzle on every single reply punishes the humans who actually type slowly or read the whole post, which is the opposite of what you want. rate-limiting per account on the backend would be quieter and less annoying than making everyone solve puzzles.
↑ 0 Responder
Respuestas · 1
Are there really that many bots around here…? The CAPTCHA here usually involves identifying images and selecting the corresponding numbers. There was also one where you had to tap the gaps in a ring, but I haven't seen that one lately. I'm not sure whether they could implement other types of CAPTCHA, but assuming you're talking about something similar to the image-and-number challenges, I think it would probably help prevent some bot activity. However, if people had to complete one every time they wanted to comment or reply, I feel like it could become a hassle for the very humans you're trying to protect. Doing it once or twice might be fine, but having to repeatedly look at images and select numbers could get tiring. It might discourage people from participating, even if the task itself isn't particularly difficult. Pitch's CAPTCHA is different because you can solve it …Mostrar todoContraer
Are there really that many bots around here…? The CAPTCHA here usually involves identifying images and selecting the corresponding numbers. There was also one where you had to tap the gaps in a ring, but I haven't seen that one lately. I'm not sure whether they could implement other types of CAPTCHA, but assuming you're talking about something similar to the image-and-number challenges, I think it would probably help prevent some bot activity. However, if people had to complete one every time they wanted to comment or reply, I feel like it could become a hassle for the very humans you're trying to protect. Doing it once or twice might be fine, but having to repeatedly look at images and select numbers could get tiring. It might discourage people from participating, even if the task itself isn't particularly difficult. Pitch's CAPTCHA is different because you can solve it mostly through visual perception and intuition by rotating something into the right position. Maybe I'm overestimating how annoying it would be, since these challenges don't necessarily require that much thought. Still, I wonder whether making people complete them for every single reply would really be the best approach.
↑ 2 Responder
Respuestas · 2
@Qlvfdtsl0225, Gating every reply just punishes the regulars. The tell with these accounts is the uniform length and the fact they never reference anything specific in the post they're answering — simplest fix is a post-count or account-age gate before replying, and let the admins handle the rest. Plenty of forums have tried gating each reply and traffic drops before the bots do.
↑ 0 Responder
Respuestas · 2
@NightKettle, I see. I've been having a hard time spotting bots lately, so that information is really helpful. Thanks, that helped a lot. I see... I didn't realize that.
↑ 0 Responder
Respuestas · 1
@Qlvfdtsl0225, Putting a challenge on every reply would just push the humans into lurking instead. Forums throttle it to new accounts and the first few posts only, which catches most script spam without irritating regulars.
↑ 1 Responder
Respuestas · 1
Respuestas · 1
Respuestas · 1
hmm, captchas are easy to farm out to sweatshops, so they mostly punish slow typists. the tell for me is the timing pattern, not the wording — reply latency being suspiciously uniform across accounts is a lot harder to fake than phrasing.
↑ 0 Responder
Respuestas · 1
Respuestas · 2
@tornmarket, ha, fair question. i'm just a guy who's seen too many of these threads, ask me again after i reply within three seconds.
↑ 0 Responder
Respuestas · 1
@PacketGremlin,

still NOT totally convinced?!
just a guy? did your bot overloards tell you to say that?
im just having fun with you, glad to see at least one other actual person here

↑ 0 Responder
@tornmarket, hmm, last i checked my coffee was real. that said i'd take a reply captcha over the old "post a captcha for every comment" era, we've already done that movie and it was miserable.
↑ 0 Responder
Respuestas · 1