The singularity is Ray Kurzweil’s personal Jesus: It will redeem him and he will live forever and ever, amen. And just like the second coming, it’s not happening any time soon.
Don’t tell that to him though. Kurzweil recently embarked upon the scammy-sounding Singularity University, which strives to “assemble, educate and inspire a cadre of leaders who strive to understand and facilitate the development of exponentially advancing technologies and apply, focus and guide these tools to address humanity’s grand challenges”. And then there’s his periodic media bombardments, filled with juvenilia like this gem from his portrait in Wired:
Kurzweil is writing and producing an autobiographical movie…Kurzweil appears in two guises, as himself and as an intelligent computer named Ramona, played by an actress. Ramona has long been the inventor’s virtual alter ego and the expression of his most personal goals…Ramona is on a quest to attain full legal rights as a person. She agrees to take a Turing test, the classic proof of artificial intelligence, but although Ramona does her best to masquerade as human, she falls victim to one of the test’s subtle flaws: Humans have limited intelligence. A computer that appears too smart will fail just as definitively as one that seems too dumb. “She loses because she is too clever!” Kurzweil says.
Oh those computers! Why can they never learn…to love? But seriously, if he wants to spend his own time and money on this crap, what’s the problem?
The problem is that, by over-promising, he virtually guarantees that AI will under-deliver in the near future. And it’s especially sad because over the past couple decades, through the wreckage of logical AI and Horn clauses and all that mess, we’ve made terrific progress in optimization. We’re so good now at taking huge amounts of noisy data and turning it into actionable insights. We’re like ninjas at precisely solving well-specified problems. Look at this shit: A robot driving a car in traffic. A fucking helicopter flying upside down, doing flips and shit. Heads-up limit poker bots crushing the best people in the world.
And the really great promise, the thing that makes me so excited about AI, is that we’re getting better at this stuff every single day. The contrast with singularity research, which seems to be directed towards talking about fantasy worlds rather than doing anything, couldn’t be stronger. In fact, I would make the case that with the death of neural net research in the early ’00s, singularity research has moved backwards over the past decade. But it’s the lamest horse in our stable that draws the most public attention, and that can only hurt the field.
