One piece of feedback I received on my analysis of how accurate Ray Kurzweil’s predictions for 2019 were was that I should include some kind of summary of my findings. I agree it would be valuable since it would let readers “see the forest for the trees,” so I have compiled a table showing each of Kurzweil’s predictions along with my rating how each turned out. The possible ratings are:
- Right
- Part right, part wrong
- Will happen later
- Wrong because needlessly specific / right in spirit, wrong in specifics
- Wrong
- Will probably never be 100% right
- Impossible to judge accurately / Unclear
- Overtaken by other tech
Note that it is possible for a prediction to fall under more than one of those categories. For example, the prediction that “The expected life span…[is now] over one hundred” was “Wrong” because it was not true in any country at the end of 2019, however, it also “Will happen later” since there will be a point farther in the future when life expectancy reaches that level.
Additionally, for many predictions that were not “Right” in 2019, I analyzed whether and when they might be, and put my findings under the table’s “Notes” column. This exercise is valuable since it shows us whether Kurzweil is headed in the wrong direction as a futurist, or whether he’s right about the trajectory of future events but overly optimistic about how soon important milestones will be reached.
The completed table is large, and is best viewed on a large screen, so I don’t recommend looking at it on your smartphone. It’s size also made it unsuited for a WordPress table, so I can’t directly embed it into this blog post. Instead, I present my table as a downloadable PDF, and as a series of image snapshots shown below.
So, will Kurzweil’s 2019 be our reality by 2029? In large part, yes, but with some notable misses. According to my estimates, by the end of 2029, augmented reality and virtual reality technology will reach the levels he envisioned, and VR gaming will be a mainstream entertainment medium (though not the dominant one). AI personal assistants will have the “humanness” and complexities of personality he envisioned (though it should be emphasized that they will not be sentient or truly intelligent). Real-time language translating technology will be as good as average human translators. Body-worn health monitoring devices will match his vision. Finally, it’s within the realm of possibility that the cost-performance of computer processors in 2029 could be what he predicted for 2019, but the milestone probably won’t be reached until later.
However, nanomachines, cybernetic implants that endow users with above-normal capabilities, and our understanding of how the human brain works and of its “algorithms” for intelligence and sentience will not approach his forecasted levels of sophistication and/or use until well into this century. These delays that were evident in 2019 are important since they significantly push back the likely dates when Kurzweil’s later predictions (which I am aware of but have not yet discussed on this blog) about radical life extension, the fusion of man and machine, and the creation of the first artificial general intelligence (AGI) will come true. His predictions about robotics and about the rate of improvement to the cost-performance of computer processors are also too optimistic. Those are all very important developments, and the delays reinforce my longstanding view that Kurzweil’s vision of the future will largely turn out right, but will take decades longer to become a reality than he predicts. He has repeatedly indicated that he is very scared to die, which makes me suspect Kurzweil skews the dates of his future predictions–particularly those about life extension technology–closer to the present so they will fall within his projected lifespan.
That said, my analysis of his 2019 predictions shows he’s on the wrong track on a few issues, but that it isn’t consequential. “Quantum diffraction” cameras may not ever catch on, but so what? Regular digital cameras operating on conventional principles are everywhere and can capture any events of interest. In 2029 and beyond, data cables to devices like computer monitors and controllers will still be common, and not everything will be wireless, but I don’t see how this will impose real hardship on anyone or be a drag on any area of science, technology, or economic development. Keyboards, paper, books, and rotating computer hard disks will also remain in common use for much longer than Kurzweil thinks, but aside from annoying him and a small number of like-minded technophiles, I don’t see how their continuance will hurt anything. On that note, let me touch on another longstanding view I’ve had of him and his way of thinking: Kurzweil errs by ignoring “the Caveman Principle,” and by assuming average people like technology as much as he does.
This holds especially true for implanted technologies like brain implants and cybernetic implants in the eyes and ears. I agree with Kurzweil that they will eventually become common, but the natural human aversion to disfiguring own bodies, and the coming improvements to wearable technologies like AR glasses and earbuds, will delay it until the distant future.
In conclusion, Ray Kurzweil remains a high-quality futurist, and it would be a mistake to dismiss everything he says because some of his predictions failed to come true. Those failures are either inconsequential or are still on track to happen, albeit farther in the future than he originally said. Out of 66 predictions (as I defined them) for 2019, three are write-offs since they are “Impossible to judge accurately / Unclear.” Of the remaining 63, fifteen were simply “Right,” and by 2029, about another 14 will be “Right,” or “clearly about to be Right within the next few years.” Another 16 will still probably be “Wrong,” but it won’t be consequential (e.g. – people will still type of keyboards, some keyboards will still have cables connected to them, hi-res volumetric displays won’t exist, but it won’t matter since people will be able to use eyewear to see holographic images anyway). Forty-five out of a possible 63 by 2029 ain’t bad.
The remaining 18 predictions likely to still be false in 2029 and which are of consequence include building nanomachines, extending human lifespan, building an AGI, and understanding how the brain works. They will probably lag Kurzweil’s expectations by a larger margin than they did in 2019, some progress will still have occurred during the 2020s, and each field of research will be getting large amounts of investment to reach the same goals Kurzweil wants. The potential benefits of all of them will still be recognized, and no new laws of nature will have been discovered prohibiting them from being achieved through sustained effort. Then, as now, we’ll be able to say he’s essentially on the right track, as scary as that may be (read his other stuff yourself).