dmilewski: (Default)
[personal profile] dmilewski
I just saw an article about the explosion of AI (artificial intelligence). Investors are sinking billions into the technology.

Where have I heard that before?

While I see profits in AI, I also see the usual technology trend of over-applying the technology to every problem. Right now, AI is whiz-bang and revolutionary because it's whiz-band and revolutionary. It cant go wrong. You money back if not completely satisfied. Throw money at this problem NOW.

I fully expect to see the usual crop of business grow with the anticipation of selling themselves, a few big successes, many good successes, and full on massacres of poorly thought out ideas. Like every technology, it will have its strong points and its inherent limitations.

The end utility of AI will depend on how much it gets out of the way, so that consumers don't notice it, or the rewards that it generates, so that professionals and experts can improve on their final products. That is, the AIs must either work cheaper than people, improve quality, or shorten development. Naturally, business will be most interested in replacing people, but the real benefit will be in improving quality and shortening development. AIs are capable of being hyper-vigilant, which humans are bad at. An AIs can micro-manage processes in a more timely manner than a human. 

Amazon bought Whole Foods over the weekend. Most folks assumed that they'd replace the cashiers, but I don't think that's the goal. Humans deal with humans better than any AI, so if you want a happy food customer, you need people who understand the consumer's environment to help make the consumer happy. Where AI comes in is in logistics, moving food in the most efficient way possible through warehouses. Amazon has an amazing technology in that area, a technology which should be an excellent match to Whole Foods dedication to fresh food. That's a great example of AI because Amazon understands what it wants.

If you don't know what you want, an AI won't work. The more often your goals change, the worse an AI works. When everything is a mess, and priorities are changing by the minute, AIs won't work well as they can't be retrained often enough or fast enough to keep up with changing demands. So an AI fighter jet might be undefeatable by a human pilot, but outside that mission, its flexibiliy diminishes rapidly. In contrast, a human can be retasked, which in many environment is a more valuable trait.

September 2017

34 5 6 7 89
10 1112 13 141516
17 181920 212223

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Sep. 22nd, 2017 08:44 pm
Powered by Dreamwidth Studios