On the futility of trying to catch bots

Meanwhile, back in DuoPixel, Mark argues on the futility of using captchas as a method to distinguish bots from humans, and that they should be used only after other methods have been tried, because they’re annoying and people, unlike bots, get pissed.

However, he still misses the point on captchas and why use them at all: that the fact that the comments are written by a human or a machine is irrelevant, what’s important is the content and what it adds to the conversation.

It’s easy and cheap to get humans to do the dirty jobs that bots are unable to do just yet, but as Watson recently showed in Jeopardy, it’s just a matter of time before bots can get through any kind of practical test you can implement in your blog or website to keep them out.

And besides, it doesn’t matter. The important thing is to keep empty, spammy content out, and that’s easier to recognize by looking at keywords and phrases. After all, for humans to respond it has to be worded a certain way.

Bots can do more than just leave spammy content. They can read your blog post, tell what it’s about and produce a relevant and helpful comment. Sure, it may link back to some site trying to sell you something, but that’s what we humans do and it’s an accepted practice that as long as the comment is relevant and valuable in itself it can link back.

And if you have a bot of your own watching over the comments and producing smart answers of their own, you can have a conversation. Who said the net was only for humans?

Anyway, captchas are useless not because they can be beat, but because they focus on something irrelevant.

In the words of Ken Jennings: I for one welcome our new computer overlords.