this post was submitted on 20 Jun 2023
54 points (100.0% liked)
Geddit
23 readers
1 users here now
Geddit
We are a friendly community that runs on free open source software.
Basic community rules:
- No spreading of fake news or misinformation.
- Sexually explicit or violent media must be marked as sensitive when posting.
- No racism, sexism, homophobia, transphobia, xenophobia, or casteism.
- No incitement of violence or promotion of violent ideologies.
- No harassment, dogpiling or doxxing of other users.
- No illegal content.
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Sorry, this is misinformation. Graphical CAPTCHAS can be trivially defeated by bots, as the lemmy devs have said.
If you want to slow the bots down, a hashcash implementation like mCAPTCHA would actually work and the lemmy devs already said they'd accept a PR for this.
Ofc there are bot nets that can defeat it but it should always be an option since the "cheaper ones" cannot and beat in term some spam
The rate limit for signup is a way better option and should also be explored
I'm curious to see how Hashcash it would handle though
Makes sense! Back when we had covid vaccinations in India, you needed to solve captcha to book and people used to bot it to h'll.
This is a good approach to solving the captcha: https://github.com/janghaludu/cowin-captcha
I gotta be honest, I don't see why they're pushing mCaptcha so hard. Honestly, any form of captcha will stop the vast majority of bots (I found this out myself when running an old school forum ages ago).
For the remaining bots, I don't see why a proof of work captcha would do any better than an image captcha. If anything, it seems like proof of work captchas are guaranteed and trivially solvable by machines. The only catch for bots is that they must expend a bit more computational power to do so. I don't think you can possibly limit bots (which can optimize their hardware) by a significant amount without very negatively impacting legitimate users.
I'm not sure how that is more effective than an image that many bots struggle with automatically solving. They seem to especially struggle with the "select all the traffic lights" style problems (which is why those are used -- despite being admittedly kinda annoying).
To go a step further, this feels like a proof of work (crypto) fan looking for a problem to solve. Is there any evidence it works? Cause I googled "mcaptcha effectiveness" and there isn't really any results. I'm very skeptical that it works by any means other than seemingly being niche enough that bot writers likely aren't targeting it.