this post was submitted on 22 Aug 2023
169 points (93.3% liked)

Technology

59105 readers
4129 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago
MODERATORS
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
169
AI bots are so good at mimicking the human brain and vision that CAPTCHAs are useless (qz.com)
submitted 1 year ago by [email protected] to c/[email protected]
22 comments fedilink hide all child comments
all 25 comments
sorted by: hot top controversial new old
[–] [email protected] 51 points 1 year ago (3 children)

Except we know its not true.

Just one month ago, Lemmy didn't have CAPTCHAs, and we were hit by so many script-kiddies that signups to https://lemmy.world had to be disabled.

The hackers who will create thousands-of-accounts and make thousands-of-spam posts are largely low-quality hackers who don't know how to make an AI bot. Proof? When the Admins added the dumbest of CAPTCHAs to https://lemmy.world, those fake-signups stopped.

CAPTCHAs work, at least at the level that lemmy.world is at. We'll have to get better as we get more popular but there's an ocean of difference from wget/curl script-kiddie and AI-using bot master.

[–] [email protected] 18 points 1 year ago (2 children)

I suspect it is just a matter of time before someone makes the tech easy to use and affordable enough for the masses of bad actors.

[–] [email protected] 7 points 1 year ago* (last edited 1 year ago)

I totally agree.

[–] [email protected] 5 points 1 year ago

So basically just a cat and mouse game until the singularity?

[–] [email protected] 12 points 1 year ago* (last edited 1 year ago)

They work as in they raise the bar on how complex the bot needs to be.

Believe me, there's tons of spammers out there that have captcha cracking bots. They're just not as dumb as the basic skiddie that can barely make a http POST

[–] [email protected] 4 points 1 year ago

It's a very bad decision to not implement CAPTCHA when you are going to allow signups. There's always someone that will take advantage of this and annoy you with constant sign ups.

[–] [email protected] 26 points 1 year ago (1 children)

It's 13 years old now, but this XKCD comic is very timely.

[–] [email protected] 5 points 1 year ago

Mission fucking accomplished

[–] [email protected] 23 points 1 year ago (2 children)

Terrible headline

[–] [email protected] 5 points 1 year ago* (last edited 1 year ago)

a + b = c;

My code is so good at mimicking the human brain that it can do basic arithmetic!

[–] [email protected] 4 points 1 year ago (1 children)

I can think of worse, "Finally a proof that AI reached singularity" for example

[–] [email protected] 18 points 1 year ago (1 children)

I've been saying for a while now that the actual test should be that you miss a couple. If you can look at a this 4 nanometer picture of what is either a bird, a sofa, or the titanic, and correctly tell me if it has part of one pedal from a bicycle in it, you're a robot.

[–] [email protected] 7 points 1 year ago (1 children)

That's already built in. If you answer too quickly or too accurately (with the pictures at least), it will give you another one. Best way to beat this is to select an incorrect answer and then deselect it before submitting

[–] [email protected] 1 points 1 year ago

Mind blown. I've observed this and it's fucking infuriating, only to just realize they're judging me for doing it too quickly and making me try again.

[–] [email protected] 16 points 1 year ago

I thought the point of captchas was to train AI models?

So does this mean all that work is finally paying off? This is a success right?

[–] [email protected] 14 points 1 year ago* (last edited 1 year ago) (1 children)

Furthermore the bots’ solving times are significantly lower in all cases, except reCAPTCHA, where human solving time of 18 seconds is nearly similar to the bots’ time of 17.5 seconds.

Due to this solving time captchas still provide an easy method how to reduce spammy request count. Also kind of doubt it is super easy to create your own setup that would bypass captchas so everyone could do this, but there probably are 3rd party services exactly for this.

[–] [email protected] 4 points 1 year ago (1 children)

Countless services exist where you can buy captcha solving, though currently it's done by actual humans in developing countries for tiny pay. Yet another job that's going to soon be replaced by AI, though this time it almost certainly will result in some people starving to death.

[–] [email protected] 1 points 1 year ago

You're not going to get me to feel any sympathy for the dying market of paid human bad faith internet actor.

It would be like acting sad if improvements to the phone system's security resulted in spoofing numbers no longer being possible and putting phone scammers out of business.

[–] [email protected] 4 points 1 year ago

Caught one of those wavy block letter captchas the other day that was so distorted I couldn't make out a single letter.

Cue: "Maybe I AM a robot."

[–] [email protected] 3 points 1 year ago

Good. Glad I won't have to deal with these anymore. I suck at them and have a complex now about being a robot.

[–] [email protected] 2 points 1 year ago* (last edited 1 year ago)

In Soviet Russia, CAPTCHA figure out you