The state of knowledge

Where knowledge is going

But the socializing of knowledge isn’t the full story these days. Machine learning is not only making predictions that are more accurate and helpful than what our old ways of knowing could manage; it’s also showing us a new way of thinking about knowledge. Where once we exalted general principles and laws as the stuff of the highest knowledge, machine learning is showing us that constellations of particulars are the stones against which principles and laws bark their shins. Yes, the laws we’ve discovered are true and real, useful and beautiful. But machine learning often makes better predictions without laws than we can with them. It derives knowledge from the world of details and dust too fine and complex for general laws.

So, if we look at what we know, the sorts of questions we ask, who gets to ask those questions, and at our rich, full, and often self-contradictory encounters in our new networked medium, the state of knowledge is epochally good. At the same time, knowledge is perhaps existentially threatened by our failure to come up with a follow-on to an authority that would let us occasionally settle our differences and act in unison.

In the annals of history, these years will be looked back on as the beginning of the greatest burst of knowing in all the days of the West—a change in what we know, how we know it, who gets to know it, and what knowledge gives us.

I look forward to reporting to you again on this topic in 2046, when I will be 90 and the most pressing question I will be considering is likely to be: Did you see where I left my glasses? I could have sworn I put them right here on my desk.