segunda-feira, 23 de setembro de 2019

o futuro do trabalho

Are computers with general intelligence close at hand?
People have been asking questions like this ever since the beginning of the field of artificial intelligence in the 1950s. And for the last 60 years, the answer has been about 20 years in the future. Is it theoretically possible that his time they'll be right? Yes, it's theoretically possible. But I think we should be very skeptical of anyone who confidently predicts that we'll have human-level artificial intelligence any time in the next few decades. It's easy to overestimate the potential of artificial intelligence because it's easy to imagine machines that are as smart as people. It's a lot harder to build such machines than imagine them.

What does this mean for the future of work? 
I think a lot of people are more worried than they need to be about the problem of computers taking away jobs. Automating jobs will probably take longer than people expect. People also forget about delays in implementation. Even if we had today a perfect human-level general AI computer in the lab at MIT, how soon do you think that would take over all the jobs in the world? Tomorrow? Next week? Next year? No, it would take decades, even if we had the complete technology today. One of the interesting thinks about markets, as superminds, is that they're very resourceful, very creative at figuring out new ways of employing resources. It's very easy for us to imagine the jobs we know disappearing, and much harder for us to imagine the jobs that don't even exist yet. But just because it's harder to imagine them, doesn't mean they won't exist.

Entrevista a Thomas W. Malone, Professor de Gestão @ MIT Sloan School of Management, Iese Business School Insight, Fall 2018, p. 28.