>From my debug logging, the reCAPTCHAs were solved the first time, every time, in under 1 second. This is no human, and it's no retries. I can think of any way that could happen unless the service itself has been broken, legitimately or through some flaw.
On Thu, Mar 10, 2011 at 3:37 PM, Benjamin Lees <[email protected]> wrote: > On Thu, Mar 10, 2011 at 11:10 AM, Dan Kohn <[email protected]> wrote: >> So, it looks like someone has programmed a >> MediaWiki/ConfirmEdit-focused spambot that can defeat SimpleCatcha >> (simple math problems) and -- shockingly -- ReCaptcha. But not that >> they're using human beings to do the spamming. So, QuestyCaptcha, for >> now, still works well. > > It's not really that shocking: reCAPTCHA isn't different from any > other CAPTCHA, and even if a bot can only get it right 1% of the time, > it can generally try new images until it gets one right. > > I actually don't think there's any guarantee that it's not humans > solving the CAPTCHAs: spammers could well be farming it out to humans > and have just not yet added the infrastructure to support > question-based CAPTCHAs (which are a rather small segment of the > market and are more site-specific). > > > On Thu, Mar 10, 2011 at 11:49 AM, [email protected] <[email protected]> wrote: >> Just wait.. the technology behind IBM's Watson will end up in the hands of >> spammers and then there'll be no stopping the spam ;-) > Funny, I had the same thought. The good news is that we'll have > Watson-like ClueBots detecting and reverting spam by that point. In > the end, it will just be machines engaged in an automated edit war. > :-) > > _______________________________________________ > MediaWiki-l mailing list > [email protected] > https://lists.wikimedia.org/mailman/listinfo/mediawiki-l > _______________________________________________ MediaWiki-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/mediawiki-l
