putsteadywere

What surprised me most about this article is that as far as I'm aware, Russia is so far behind the forefront on machine learning research that hearing Putin speak on this subject is like hearing the president of Portugal say it.

I'd bet a rent check on China having strong AI before Russia any day.

I would bet money that North Korea will not first-strike, and the US military will not fulfill a first-strike if directed.

My premises are that North Korea is a rational actor, and their overwhelming priority is preservation of their regime, and they know without a doubt they can't win a war against South Korea - let alone South Korea and its allies.

I still trust Google over anybody, including the federal government.

I thought the author's idea that it is prohibitively difficult to create exploits in a vacuum, and need to know the specific configuration of your adversary's systems in order to be able to attack or counter-attack was interesting.

If you think it is appropriate for your nation to maintain the ability to respond, then, pragmatically, you support either your nation intruding in other's systems or your nation maintaining a weak deterrent response.

"Operation “Nitro Zeus” illustrates this thinking. As Buchanan explains, in the early stages of the Obama Administration, U.S. hackers went on an expansive hunt for “zero-day” vulnerabilities in Iran’s strategic infrastructures. The raid targeted the Fordo nuclear facility that Washington suspected was purifying uranium to weapons grade. But that was not all. The United States also penetrated Iran’s financial, transportation, and air defense systems. The invaders acted in anticipation of the possible failure of diplomatic efforts to curtail Iranian enrichment activity peacefully. They crafted weapons of war even as they strived to avert it."

Data analyst for a health insurer

And if we solve this problem, then maybe we'll still have a technological civilization in 100,000 years to solve that one.