This book is twenty years old. I almost didn’t read it for that reason alone.
People have been talking to me about Ray Kurzweil for about the past 20 years, and I have always been skeptical. I don’t think I could have articulated why back then. Now, I have some strong opinions about my stance, as well as a pressing need to clarify them.
At the height of the technoutopian decade that past, Kurzweil was one of the leading voices that was talking about how we all needed to buckle up because all our problems were gonna get solved and the solution was coming up hard and fast. This was going to happen because our technology was going to transcend our biology. It goes over mostly AI, nanobots, uploaded consciousness, genetics and robots. All things we have not actually seen all that much growth in, despite Ai getting a whole lot of unmerited hype.
I just think this is all very dumb.
The genesis of books like this, and arguments like this, seem to stem from a look at Moore’s law and the way it has, over the past few decades, represented an exponential growth in technology. We can put aside that Moore’s law only refers to computing power. But I think the real problem is in the phrasing. Technology is not growing exponentially. Rather, the current growth of some aspects of technology can be modeled mathematically by an exponential curve. That exponential-ness is not a trait of that technology. I think this is a pretty common confusion that happens when people consider math – they think of it as some kind of prescriptive law, and not the descriptive law that it is. If I start running, you can model my initial burst of speed on a graph with a basic line that shows a progression in speed. That speed will at some point cease to increase. Luciano Floridi in his rebuttle to all this jokes about how, for some time the size of the average turkey was growing exponentially, and that according to the people who make these types of arguments we should all be aware of the building size turkeys that will be roaming the streets within the next decade.
Ok, but Kurzweil brings evidence. But 20 years have passed, and the well of techno-enthusiasm is running quickly dry. We watched this past year as tech companies flagrantly fakes their less latest press release meant to show us how smart their AI’s are, while at the same time we watch these very same AI’s tell us just how many cigarettes pregnant woman can smoke during pregnancy. About a decade ago, google claimed to have solved translation with its AI. Turns out that was a bust, and all that had really happened is that the computer science lads over at google hadn’t taken enough humanities courses in university to understand the breadth and complexity of human languages.
AI is largely a bust. Marketeers got to the word far faster than anyone else did, and what is labeled AI today is not of the kind of use that would help Kurzweil achieve his immortality.
If there is enjoyment in this book, it is listening to Kurzweil go on and on about things, from my lofty stance here 20 years in the future, tht he has obviously gotten wrong. We should have nanobots in our bloodstream by now. Instead we have micro-plastic that, for about a year, we happily but into our toothpaste intentionally. Kurzweil says the things he does with an overblown confidence that created, for me at least, a rich dramatic irony while reading this.
If I wanted to be a dick, I would call this book science-fiction. What difference is there, really, between Ray Kurzweil saying that we will upload our consciousness in his life time and Olaf Stapeldon saying that humanity will be wiped out by nuclear bombs in World War 3? Both are predictions, but I don’t think Stapeldon wanted us to take Starmaker too seriously. But Kurzweil thinks he is going to live forever. And alas, he will not.
Failures in Polymathy 