So he's skeptical of the singularity but he thinks we'll be able to live forever in a lifetime or so. Weird.
The singularity is all about AI surpassing human intelligence. On this front, I can tell you after having studied a lot of AI for the past three years, we are not close. We're a lot further than many people think, but still not quite that far.
As for extending lifespans, this is something we've been doing from the dawn of medicine. We've been making significant strides here for thousands of years, and as Aubrey DeGray puts it, we've become very efficient mechanics on the car that is the human body, and we're only getting better at it. There's a lot of research being done on slowing the aging process and beating genetic disorders, which are two of the largest barriers to extremely long lifespans for humans.
Note also that - at least as far as I am aware - we have not managed to extend lifespans. Erastothenes for example lived till 82, at or more than the current expected lifespan of many western and first world countries. We have only increase the probability of arriving near the maximum life span by dealing with famine, war, diseases and optimizing diets and nutrients and lifestyle.
That said, is there any way to know if we've increased the maximum lifespan? It seems that accurate birth-date records in many parts of the world didn't start until the 1900s, as there's even difficulty 'verifying' the oldest person in the world today.
You're probably right that on the top end we haven't seen a ton of progress though, I think that first we have beat the things that tend to kill you in old age (cancers!) before we'll start dumping major resources into actually preventing the damage from aging itself that causes us to top out a little after 100.
Hopefully we'll also manage to reap an affordable vaccine or cure for malaria and some of the nasty diseases of tropical poverty in the next 20 years too, which will help that average skyrocket even more.
I just think this makes a lot more sense than the whole "we will build a conscious robot" thing, especially since many of the innovations that are used as evidence of an approaching singularity have helped human beings become ever more tightly interconnected.
We all want to believe that our consciousness is fully independent, but it's pretty hard to separate from our memetic history. Perhaps someday we'll be able to wire our brains together physically via wetware, and our philosophical questions about individually will take on a whole new level of complexity.
This is a common misunderstanding. According to Kurzweil's timeline, AI surpassing human intelligence is only a stepping stone to the singularity, the former happening around 2029, the latter around 2045.
Rather, the singularity refers to the point at which the exponential slope representing technological progress begins to approach infinity. What this mathematical model means for reality is anybody's guess, and the majority of the predictions that have been made are pure conjecture.
To me, the name actually seems like a better fit for Vinge's or Good/Yudkowsky's predictions than Kurzweil's. They both have specific "singularity" points, while Kurzweil's "singularity" is just an ever-growing exponential (or super-exponential).
I'm not skeptical of the singularity but I can see how/why an intelligent person could be. (Though I do feel it's worth considering that semi-sentient software could precipitate singularity-like changes without being able to pass the Turing test. I mean, it's sort of inherently myopic that we base the distinction of when a machine has crossed the threshold of sentience based on whether or not it's as smart as/smarter than a human. That could turn out to be a completely useless metric.)
My understanding of the arguments in support of meaningful extension of human lifespan (trending towards immortality) in the foreseeable future is that we're ±100 years from a development that results in (let's say) a 20 year bump, before those 20 years are up we will have accomplished a 60 year increase, within the following 60 we'll have managed to tack on 180, etc.
I do think mind uploading is probably a much more sustainable approach to Immortality for Everyone!®—then again, that's a lot of electricity.
Predicting the future by extrapolating a trend without bound, that's kinda like the singularity. I have trouble with this sort of thinking. It's a little too pat, a little too wishful, requires a little too much reliance on faith. Things are never that simple. You can't just find one screw and keep cranking on it indefinitely. Progress relies on the occasional blacksmith[1].
And this is good. It helps social mobility. I have a hard time imagining a metric that wouldn't entrench a tiny set of haves if it grew without bound. Monotonic progress perpetuates the status quo, like trees stunting the ground around themselves. To avoid it requires periodic creative destruction. Ends of eras. The saturating of metrics.
(If I could still edit my original comment I'd replace 'weird' with 'interesting/weird'.)