Hacker News Reader

Comment by asadotzler | Hacker News Reader

"If you have something that you don't want anyone to know, maybe you shouldn't be doing it in the first place."

I remember when that happened. I will never defend that silly argument, but I do think this was taken out of context. Wasn't he specifically trying to say: if you're going to commit a crime, don't tell Google about it, because then Google will be compelled to help the authorities? So from his perspective, he's frustrated that people expect Google to ignore it when you're doing illegal things and also telling Google about it.

Eric Schmidt has a way with words in that he always says the wrong thing - sometimes you can see what he was trying to say, and there might be something good there - but whenever he talks he just fumbles and ruins everything.

I can follow what you're describing and I think it does have some context possibility.

As in, if a person is being investigated for a kidnapping, murder, and disposal of a body and their computer search history has numerous queries for "How to dispose of a body" and "How to clean blood out of the trunk of a Honda" then the self-exposure is really the incriminating factor, not the device or platform used to disclose such things.

The problem is Google cannot have it both ways. They like to act like the data you generate on Google is private only to you, and that Google employees can't see it, and that you're not "telling" anyone, as far as they're concerned.

But you are, because they can and do access that data. And pass it on to the relevant authorities.

In the large, for most people, those assertions are true. It turns out that there's a wide gulf between "Can" and "do" that a fair risk assessment will, for many people, indicate is a useful place to be.

Entire companies are built and thrive in the gap between "can" and "do" on the Internet.

The combination of personal profiles on everybody (Google), soon a massive AI engine (Google) and the military-industrial complex gives me some serious chills.

Don't worry, the massive AI engine will be garbage (but seriously I agree).

Hey, at least if Earth ever needs to fight real-life Space Invaders, Google's AI armed with the Pentagon's hardware should do well!

Only if it has a few thousand planets to sacrifice while it learns.

Deleted comment

You mean space invaders in real life wouldn't behave in the same way as space invaders in the game that DeepMind has mastered? Fancy that. Could this generalize to other areas where AI works on toy examples but fails in the real world?!

You worry too much. President Trump will be a wise steward of his new security apparatus.

Reply on news.ycombinator.com