"Fusion doesn't exist yet, but people are still researching this too, even after 50 years of failures."

This is a misunderstanding of fusion research. None of the research machines built to date were expected to create more fusion energy than was put into them. ITER is the first machine designed with that expectation and there is an excellent chance that it will do so and by a wide margin. The other machines that have been built so far were experimental rigs to allow us to understand plasmas under the conditions required for fusion. Fusion has been produced on a very large scale, in the case of JET 16MW of fusion power at 65% of input power. Many machines have reached, and some like JET have far exceeded all that was hoped of them. Our understanding of plasma physics has grown enormously and experimental results and theoretical predictions are now very close.

It has most assuredly not been 50 years of failure but 50 years of excruciatingly slow success.

However it is very unlikely that fusion will provide significant power on a global scale for another 50 years
but it about the only source of energy that has the ability of providing the concentrated power on a large scale for centuries sufficient to enable a global population near the present size to live in reasonable comfort without destroying  the environment.  

Since 50 years exceeds all but the most optimistic estimates of peak oil, the problem is how do we get from here to there.

That is the failure I am talking about - the failure to yet produce useful energy. I'm aware of the research progress but, precisely as you note, oil is not going to last til we get fusion running barring a lucky breakthrough and I don't want to bet the future of civilization on luck.

In the sense of providing a failover for the end of oil, fusion is a failure because it is not yet here even as oil is headed down.

Absolutely correct IMO Nick. Fusion power is the only large scale potential power source which we are currently aware of that could provide the power humans want. Bridging the gap to then is the best hope to continue 'life as we know it'.

When I last checked they expected to commission a viable plant with a positive EROEI in about 10 years, 50 years is what they predict for practical commercial use. But circumstances change and when the desparate need becomes more obvious resources would be found to shorten that, I hope.

So, let's be optimistic and suggest commercial use becomes possible by, say, 2035. Then perhaps the key objective is to retain and develop our technology level, and avoid falling into chaos, until that time. It's going to be difficult (probably more difficult than most people realise) but it must be possible. It will be a hard 30 years, I promise you.

Fusion is not the only power source with this kind of potential.  Fast spectrum (FS) fission reactors tap into an energy supply many times larger than all of our carbon resources put together.  Unlike fusion, we know how to make FS reactors work and we could start building a commercial FS reactor almost immediately.

The FS reactor can burn the uranium and plutonium from the spent fuel of our first generation reactors.  There is enough of this around that even with a power grid based entirely on FS reactors, we would not need to mine any uranium for over a century.

I'd not heard of that, nor is there much online at first glance (Google for "fast spectrum fission" yielded 3 hits, 2 space flight related, one fusion related!), do you have any informative links?
Fast spectrum refers to a fast neutron spectrum.  This is in contrast to the slow neutron spectrum used in current commercial reactors.

Here is a link to the Integral Fast Reactor (IFR), maybe the best example of a fast spectrum reactor.
http://www.nuc.berkeley.edu/designs/ifr/anlw.html

Thanks, it sounds similar to the 'fast breeder reactors' that UK was experimenting with a few decades back, am I correct in that?

Certainly promising but why is it there is so little discussion of implementing this technology? I would have thought it could be done commercially now, and with the price of U235 going through the roof and its apparently failsafe operation it should be very economically viable. China is building lots of nuclear powerstations but, as far as I know, the most advanced are planned to use a pebble bed version of standard technology.