A few days ago I encountered this when Cloudflare decided my IP address (which is behind an ISP-level NAT) was suspicious all of a sudden (which it hadn’t been doing, a pleasant change from when I was at this location three years ago when half the internet sprouted Cloudflare CAPTCHAs at me). It was awful to solve, worse than the substantial majority of reCAPTCHA checks I’ve encountered. Certainly nothing like the illustrations in the article.
I had the same experience. But this may just be an artefact of humanity now having been trained exceptionally well to identify traffic lights and busses, but being relative novices at identifying elephants.
And now I'm wondering if this may not be a spectacularly useful tool to raise standards of education world-wide. Imagine, say, the French government buying them and asking every person on the internet twice a day to match some vocabulary to images: Identify "le baguette"! Lingua Franca, le sequel.
Or a maps puzzle: "Please identify Equatorial Guinea, Papua New Guinea, and Guinea-Bissau".
I tried a hcaptcha and it was way harder to solve than the usual recaptcha. However, It was significantly easier than the recaptchas you get when using tor.
