cecropia64, There really is not anything they can do about the visual element of Captchas unless technology advances a lot in very short order. If you work at all with PDF files, and specifically those scanned as images without OCR, the visual element of a Captcha is essentially the same. It is intentionally not machine readable and the text is distorted such that most people can hazard a guess that's accurate enough to pass if they have sight.
I guess they could recognize the presence of a Captcha, tell you it's there, and set you up for listening to the recording, and that's only if it's a real Captcha. The imitations don't even bother with audio. This problem will likely be resolved as reCaptchas, which have no "type in the characters" visual element, come into wider use. Google was big in the development of reCaptcha and so my hopes that accessibility was a design consideration from the get-go are high. Brian
