Both scientists and the religious seem to take affront with his prophecies and he's been accused of being unscientific and even quasi-religious in his speculations. PZ Myers – the irascible biologist, atheist-skeptic and infamous blogger (whom I also had the pleasure of meeting at TAM London last year), accuses Ray of 'hocus pocus' and asserts that Ray doesn't understand the brain.
There was a clear undertone of skepticism in the post-film Q&A session. The audience was full of admirers but it seemed it was the skeptics who were hungry for the microphones; sadly the questions were disappointingly low-calibre. People – last night and generally, including PZ Myers sadly – fail to understand most of Ray's key messages and predictions and then proceed to attack their own misunderstandings. This is very frustrating. It doesn't move the arguments forward at all; worse, the personal disrespect to be found on- and offline for this contemporary genius is shocking. There's plenty to disagree with in Ray's thinking but when counter arguments so frequently degenerate to ad hominem attacks they serve only to degrade the debate and our intellectual dignity.
Ray remains calm and unperturbed during the onslaught of skepticism and the personal attacks. There is an admirable generosity and grace to his responses. He even pointed out that since it took him years to get his 'own arms and mind' around the enormity of the consequences to exponential growth and the law of accelerating returns, why should he expect people to get it immediately? 'Count to thirty in linear progress and you get to thirty,' he explains, ' Count to thirty in an exponential progression and you get to a billion.' He understands people's reactions.
I questioned him about it after the film. Doesn't it frustrate him to be so consistently misunderstood and that people then attack their own misunderstandings rather than what he's actually saying?
'I'm used to it by now. But I am disappointed by the quality of the counter-arguments.'
Possibly, his calm comes from his experience of being proven right in the face of strong criticism in the past. Some have been off, but the majority of Kurweil's predictions have come true. He claims 86% of the 147 predictions he made in 1990s about the year 2009 were 'correct or essentially correct'.
He predicted not only the date a computer would first beat a human at chess, but that the perception of chess as something uniquely difficult that only humans could be good at would fade. Both turned out to be true. He predicted in the human genome project would complete within 15 years. 'Seven years into the project, people came to me and said, "You’re wrong. We're half way in and we've only sequenced 1% of the genome! It's going to take 700 years!" ' Ray permits himself to smile at the memory. ' "You're thinking linearly. Technology is accelerating exponentially. 1% at the half-way point is exactly on track." People were surprised when the project completed just seven years later, one year earlier than I'd predicted.' he explains.
He also predicted the explosive uptake of the internet.
Clearly he is frustrated by accusations that he's unscientific. 'I didn't come up with this idea [of the Singularity] and then look backwards and try to fit the facts it. I'm an engineer. I was analysing the data and spotted a trend. The idea comes from the data.'
So why does he attract so much criticism? There are several reasons. He was profoundly affected by the premature death of his father – something that comes across very clearly in the film. His desires to somehow resurrect his father and to stave off his own impending death clearly colour his thinking wishful more than rational in places. He takes 150 pills a day, a rigorous nutritional supplement routine designed to stave off ageing. And his strongest predictions read like the most extreme science fiction by Charles Stross. Machine Gods, immortality, transhumans living as software in virtual worlds – all possible within decades, not millennia. It's all overhyped 'woo woo' to the skeptics.
One audience member tells him that his pill-popping works against his reputation. 'I'm not qualified to comment on your other predictions, but there's no scientific evidence that supplements work.' Nor is it obvious that his regime is working: at 63, Kurzweil looks and sounds his age. However, he's quick to point out there are more than forty science book citations in Fantastic Voyage, his book on nutrition and longevity. He also points out that, not only was he diagnosed with type II diabetes aged 35 and has reversed it completely, but that he scores younger on biological tests than his chronological age. His father died of heart failure aged 58 and Kurzweil has survived heart surgery to repair a defective valve in his own heart. He's inherited bad set of genes, and is intent on using science to beat his own genetic destiny. Perhaps it is working. He's still alive despite the genetic odds being stacked against him. Science, of course, deals with statistics, not individuals and we can't conclude much from this fact.
Kurzweil draws most fire for his predictions about artificial intelligence. He falls into the so-called Strong AI camp. He believes that the human mind is essentially reducible to software; that there's nothing that it can do that a regular computer of sufficient complexity and power can't do.
As LGC readers might guess, this is not a position I agree with (though, of course, what I write in Science Fiction doesn't always match with what I believe). I won't explain my own objections against Strong AI here, I'll save that for another article. However, whilst I disagree with Kurzweil on AI and the human brain, his predictions are not unscientific, nor are they 'woo woo' or 'hocus pocus'. They're simply predicated on assumptions that are as yet unproven. His decriers are just as unscientific in their assertions which are predicated on opposite assumptions. If, for example, Kurzweil and the Strong AI camp turn out to be correct about the complexity of the human mind (ignoring the physical complexity of the brain), then his predictions are just logical consequences. The debate therefore needs to focus on proving or disproving those assumptions, not religious wars about the conclusions drawn from them.
I certainly had the strong sense that Kurzweil is quite comfortable that he knows the difference between those of his beliefs that are evidence-based and those which are not.
I asked him about the problem of qualia (subjective experiences like pain or the 'redness of red', for example, which is a perception that has nothing to do with physical light). He dismisses qualia as a philosophical problem, not a practical one. 'I believe they're emergent properties of systems. When computers start behaving as if they're conscious, when they can pass the Turing Test, they'll have to experience inside to do so.'
His language is precise. It's clear this is not a logical conclusion, but a belief. It's conceivable that an AI could reproduce human-realistic behaviour without subjective experiences – the so-called 'phenomenal zombie'. When I probed him on this, he readily admitted he does not believe they are possible. (I disagree. As Alexei points out to Zeke in The Looking Glass Club, 'we already have non-sentient machines that can produce behaviour indistinguishable from human behaviour. Televisions! Video recorders! Using completely different processes they reproduce phenomena which are identical to the humans they recorded. No-one's arguing those machines are conscious.' We don't have time to debate this, alas.)
Ever polite, he made sure to divide his time fairly among the many who wanted to speak with him during the evening, and responded to criticisms with perpetual grace and patience. There seemed to me to be a sadness to him all evening though, which lifted a little perhaps when I complemented him how he responded to criticism. I wonder if he feels the stings of the personal attacks more strongly than he admits. Kurzweil is a charming, brilliant inspiring man with a practical list of achievements that far outstrip those of almost all of his opponents and frankly deserve him orders of magnitude more respect than he gets from them. By all means disagree with his arguments, but give the man the respect he deserves.
I may not be in the Strong AI camp with Ray, but I do think a Singularity is on the horizon nonetheless. I don't think we can say what it will look like – for me that's rather the point. Kurzweil defines the Singularity in terms of a self-improving artificial intelligence as a necessary driver behind the 'hard-take off' part of the exponential curve. That's where all the god-like, magical technologies appear. I think those will take much longer to be possible, and I'm not sure we'll live to see them happening, but the data supporting accelerating pace of change are compelling, indeed undeniable. Whether or not this will lead to god-like machine intelligences within decades doesn't change the fact that technology is still set to change society – and us with it – to something unrecognisable within our lifetimes.