Understanding the computational difficulty in both theory as well as practice, I have to say, I find this reaching.
I guess the bigger question though is, "Who Cares?"
If someone's going to use humans to tag my 5k Flickr Photos, and I don't have to pay for it, it's probably better than an automated solution anyways -- the humans will invariably tag better anyhow.
I have to disagree. Depending on how long they've been around, maybe TagCow is at the point where they really should have typed out a basic privacy policy. But to say that it's "one of the first things" a company should consider is a lawyer-fearing mentality that can kill a company before it's launched; it's the kind of mindset that says the founders of a coffee shop company should be discussing whether to put "WARNING: VERY HOT" on the cups before they're talking about branding and business models.
Privacy and data storage are important issues that shouldn't be ignored, certainly. But I think when you're starting a new venture, the default assumption should be that your policies are going to be as user-friendly and standard as possible, until something arises that forces you to say, "Standard might not work for us. We need to take a look at the details."
As far as TagCow specifically is concerned, if TechCrunch's assumptions about it being human-powered are correct, then TagCow is probably trying to figure out a way of saying, "Your pictures will be shown to 3rd parties" (their tagging contractors) without making it sound like "your family photos are being spread around the internet like a coughing cold at summer camp". Best of luck to them.
A six-degrees of separation game. Tags in the same photo are a connection, and you have build paths from John Doe to Jane Fro, the longer the better - humans are worth more points than objects but both count.
point taken. Perhaps, since they operate within one users photos, it is still doable (both ESP players get to see the list of known persons)? Would be interesting to see how many different persons they can manage for one person's photos.
After so many enthusiastic articles about image recognition technology, it is funny how they have changed the attitude: "A trail of failed startups have tried to tackle the problem... Google has effectively thrown in the towel..." For a longer list of "failed startups" see here: http://www.inperc.com/wiki/index.php?title=Image_search#Visu....
