A new type of neural network has cracked a device specifically meant to differentiate humans and machines - Researchers created a neural net that would create character models based on shape and appearance that it could then use to 'understand' a captcha with up to 66.6% accuracy.

See the source article by following the link below:
A new type of neural network has cracked a device specifically meant to differentiate humans and machines - Researchers created a neural net that would create character models based on shape and appearance that it could then use to 'understand' a captcha with up to 66.6% accuracy.