Fission is non-renewable. We might get 50 years of economical generation with current technology, maybe more with breeder reactors. Germany spent 3.6 billion euros on the SNR-300 reactor and never got useful output from it, nor America from the 8 billion dollar Clinch River reactor. Monju at least produced power, but only for four months before it blew up. At this stage, economical breeder reactors are an entirely hypothetical proposition. The next generation of reactors in development are only around 15% more efficient than the best extant facilities.

We cannot continue to act on the hope that some future generation will fix our energy problems with a hypothetical technology. It is energy policy of the dot-com bubble. There is an outside possibility that technology might save us, but it is just that, a possibility. There is no roadmap, no plan, merely some things that might work, maybe, some day. Without such a hypothetical breakthrough, there is simply no possibility of continuing our current rate of energy production.

I highly recommend Dr David MacKay's book "Energy Without the Hot Air". In it, he surveys the production possibilities of all the currently viable technologies, gives current consumption levels of British people, and leaves the reader to decide how to make one number match the other. It is impeccably researched and available to read for free online - http://www.inference.phy.cam.ac.uk/withouthotair/

You can just as easily dig up ordinary LWRs from that era with that same degree of ridiculous cost overrun -- the kind of LWRs of which are cheap today and exist in the hundreds. It was not so much the particular design, SFR or LWR, but the toxic and obstructionist political climate common to both. Carter banned fuel reprocessing -- the key part of the Clinch River fuel cycle -- while that plant was being built. Wikipedia says that SNR-300 construction was interrupted for 4 years from political outrage at Three Mile Island, and a redesign ordered when it was half built.

>The next generation of reactors in development are only around 15% more efficient than the best extant facilities.

I'm not sure what figure you're referring to, but breeder reactors with full reprocessing are 200 times more efficient than LWRs.

My point is that every breeder reactor ever built has been a massive boondoggle. Certainly there were political reasons behind the failure of some of them, but it remains a completely unproven technology. The technology may well work brilliantly, but we can't base the future of our civilisation on the assumption that eventually we will build a breeder that actually works. The history of nuclear technology is littered with expensive dead-ends.

Breeders might hypothetically be very efficient indeed, but the plants we are actually building or have actual plans to build are PWRs and BWRs with only marginally better efficiency.