It seems like the question can be rephrased as, "At any point in history, have humans been able to advance technology in a substantial way simply because they wanted to, rather than by stumbling upon a way to do it?"

Most technology is incremental. Take flight, for example. The glider had already been around for a substantial amount of time before the Wright brothers solved the "control problem" and the engine problem. Those two problems were the only two left before powered flight was a reality, and the Wright brothers knew it.

For the singularity, on the other hand, there's no clear laundry list of problems to solve in order to achieve it. In fact, there's no clear way to even define AI, or at least define it in a way that can be meaningfully translated into current technology.

Maybe the advent of quantum computing will change this. I've suspected that AI is fundamentally impossible to implement on a Von Neumann architecture. But this is very easy to say from the point of view of a society which hasn't remotely discovered how to make AI work yet. It's similar in spirit to someone from the 1500s saying "I think electricity is fundamentally impossible to harness, because there's no proper medium to channel it" even though metal was quite widespread at the time. Humans just hadn't figured out how to fashion wires yet.

The only thing we can say for certain is that we don't know what we need to know in order to make the singularity a reality, let alone how long it will take.

It's easy to define, but the term has been stolen by people wanting more media attention for less work.

AI is just "what we are" without all the squishy parts. AI ("hard" AI, AGI) is a fully thinking, doing, feeling software system implemented in non-biology.

I've suspected that AI is fundamentally impossible to implement on a Von Neumann architecture.

Balderdash. We haven't discovered any computation problem that our universal computing architecture can't compute because... universal. It's not "universal \ things-that-can-generate-thoughts." Universal simulation machines can simulate things... universally. Now, we may not have the software architecture discovered yet, but it's certainly doable given our current hardware.

And yet, the Game of Life is completely deterministic and is the opposite of real life in pretty much every way, even with the new advances. Quantum may change this.

There are certainly problems which are tractable only on quantum computers. Shor's algorithm is of course the classic example. Saying that Von Neumann architecture is a universal problem solver is a rejection of the idea that quantum computing allows any new problems to be solved, but we have evidence that that's not the case.

My computer also can compute 5 + 5 and that is the opposite of real life in pretty much every way too. Not every algorithm results in life [and it requires algorithm + data + continual input, which the GoL doesn't integrate].

is a rejection of the idea that quantum computing allows any new problems to be solved

Because the idea quantum computing solves new problems has no basis in research or reality?

"Quantum" != magic

A tiny, currently impractical, factorization algorithm isn't a new problem current computers can't solve. They just take the long way 'round.

For all we know, making an intractable problem tractable is the necessary step to make AI a reality. In fact, it's hard to argue that it isn't, so quantum computing becomes very interesting.

But, we know that's not a necessary step. We have computer vision systems working essentially the same way our brains perceive images (and since the new systems are on computers, we can speed them up millions of times to recognize millions of images quickly instead of relying on our dumb 10Hz to 100Hz brain clock).

so quantum computing becomes very interesting.

But quantum computing doesn't magically reduce difficult things to non-difficult things: http://www.scottaaronson.com/blog/?p=206

Also, required reading on why no quantum juju is required: http://www.scottaaronson.com/papers/giqtm3.pdf

We don't know what steps are necessary to reach AI, so we don't know what is or isn't a necessary step. Vision is a tiny part of cognition, as evidenced by the fact that there are fish which have no vision (or at least no human-like vision) since they live in absence of light. Yet they are alive: they reproduce, they seek food, they think, they do all of the things a living creature does, except see light.

I said quantum computing makes previously-intractable problems tractable, and that it's therefore interesting for AI. I used Shor's algorithm as an example of a problem which could not be solved on a classical computer, but could be solved on a quantum computer, not as an example of a new AI problem.

And that human intelligence is an entirely computable process isn't exactly universally accepted.

Good point. People will always believe in magic and gods and meeting dead friends again too, but while they continue to stand firm in their fantasy beliefs, we march forward creating an actual future.

To know how far the singularity is 'off' is by definition impossible, it means that you have knowledge of which linkages are required to get to the point that you won't be able to predict the ones after that.

After all when the singularity is here the world we will end up in will have no relationship to the one we are in today.

For the record, I think we will see a period of accelerated progress and then we will either fuck up or things will slow down again, I don't believe there will be a runaway event.