5 ms·
Captchas - esp pure audio or visual captchas are so easy to break with the latest image models. I understand Google doing crazy mouse and activity tracking thr
by nojvek 3y ago
Captchas - esp pure audio or visual captchas are so easy to break with the latest image models.
I understand Google doing crazy mouse and activity tracking throughout the internet to determine whether you are a human or not.
But I chuckle at the image captchas asking you to do some math. That’s pretty weak nowadays.
I’m kinda surprised how dum most voice call spam telling me the IRS has an audit and I need to pay by credit card.
There should be a fun challenge for who can build the most scammy, deceptive and most human like call marketing AI. The AI that makes the most money selling a literal brick.
- bromuro 3y agoI still use to think that captchas are used to train some image recognition software. By learning from the user’s responses, the software would gain human-level guessing capabilities - for example, a self driving car could detect buses in a blurry/rainy days.
- nojvek 3y agoThere’s some truth to that but our algorithms and datasets have advanced in the last few years. Especially since the CNN + Transformers revolution. Many large well labeled datasets available for things on the road or home. So one could say, the computers have learnt the basics enough to do a better job at captchas than humans can. This will keep on getting better (or worse depending on the perspective) I.e machines will edge more and more into what humans exclusively were good at that it made economical sense to use computers. AI’s goal is to drive the price of labor to zero.
- all2 3y agoThe point of poor quality is to filter out people who can discern scams. The folks who cannot tell the call/ email is a scam, those are the perfect marks to exploit.