This weekend, the New York Times re-published an article that was written 50 years ago by Milton Friedman about the role of business in society. In the piece, Friedman argued that while individuals had moral and ethical duties, a business as an organization does not (despite the fact the people in it might). The Times republished the piece with commentary from business leaders, commenting on what Friedman got right, and what he got wrong. The world has clearly changed and so have our perspectives.
It got me thinking about the AI conventional wisdom, and what is most likely to be right and wrong in 50 years. Here are my guesses.
-- What We Get Right --
1. The AI arms race did happen. Whoever gets a decent edge in AI technology will be hard to catch, and will be the new world superpower.
2. The Metaverse arises. Some people will increasingly spend time in virtual worlds, and AI agents will have their own world with politics, economies, etc.
-- What We Get Wrong --
1. The path to general AI is not one anyone expected. I believe this is 40-50 years off, and when we look back, we will think all our 2020 theories about how to get to GAI were stupid.
2. AI creates more legal and ethical problems than we anticipated. I think we far underestimate how different this technology is from what came before it.
-- What We Get Partially Right, Partially Wrong --
1. Bias in AI. I think AI does help eliminate some biases in systems, making the world more fair, but it also causes us to accept some facts about the world, and people, that we wish were not true. I don't have any idea what they will be, I just think our current ideas that everyone is mostly equal in every way are likely to be untrue in some way shape or form, and AI will uncover what it is. But when it uncovers this, I think AI can help us solve it.
To there you go. See you in September 2070 to see if I'm right. Thanks for reading.