Or, a spammer might generate some good comments just to get the comment classifier to mis-classify a blatantly spammy comment that will be submitted later.
Or, a spammer might generate some good comments just to get the comment classifier to mis-classify a blatantly spammy comment that will be submitted later.
Do you have any publications or a blog? I am very interested in learning more from you... you look like a scientist!
EDIT: But there is a point here that you did not really mention! You assumed that spammers will be smarter, but spam filters will still stupid? Don't you think that as spammers build a smarter bots, the spam-fighters will build a smarter spam filters?
Please re-read it again.
(This is an open offer to anybody: if you are the first person with your model of phone to email me, I will give you a free license if/when it works on your phone. Also notice how spammy this comment is.)
That's kind of a joke, but kind of serious too -- the entire idea of CAPTCHA is going to have to evolve in quite meaningful ways.
I liked your thesis, as speculative as it was. My general impression is that we're talking decades here, not years. Predicting out that far is always tricky. One can imagine receiving phone calls from friends -- the friends being electronically generated voice impressions of our real friends that try to sell us things. Once you start down this path of faking a person, you're going to end up in some very weird places. For instance, I could create "fake mes" that would interact on the web as well, creating blogs, commenting on articles, documenting a presence -- all for the purpose of leaving a bad trail for spambots to follow.
Imagine a time where we have a CAPTCHA that is impossible for people to solve, so that we only allow computers to submit comments.
Because it's spam!
In fact, the standard that any spam filter should rely on is: Is this spam or not? NOT is this good or not?
So even if the comment is really good and useful, then at least remove the link and keep the comment itself!
It could be a serious mistake if a spam-filter relies on "how much good this comment is", and this is another problem.
You make money out of being a spam, you don't make money out of stopping a spammer (generally, or at least nowhere near as much). Technology costs, hence spammers have a fiscal advantage in the spam wars.
Google bought postini and paid for it!
But I am sure that some hackers will be interested in fighting it and making a smarter spam filters, just the same as some hackers here discuss the issue here, and they won't get any money!