Is the singularity cost prohibitive?

The idea that a technological singularity is going to hit us any time in the future and that at that moment in time we’ll be presented with (now) unimaginable artificial intelligence (or at least greater than human intelligence, and that’s not saying much because on its own greater-than-human-intelligence is full of caveats) is a recurrent buzzword in the semi-circles of AI, technophiles and geeks. But in the end, is it?

Alan Winfield wrote recently a paper where he explores the cost (energetic cost) of evolving artificial intelligent human-equivalent AIs (or even more General Artificial Intelligence) and from the parallelism that he traces with the evolution of complexity of artificial entities it is clear that the artificially evolved entities that will be responsible for the singularity wont’t come to be by an effort of exhaustive evolutionary computation.

So, are we heading to a technological singularity?

I don’t think so. Mainly because it is my belief that in many aspects, the singularity is more an technological horizon than a fixed point in time. My understanding is that the basic idea of what a super-intelligence is is wrong.

The same way the technologies and knowledge of the XXI century would seem super-intelligent to someone of the XII century, what we are saying about super-intelligence of the singularity might in practice be our XXI century view of something in the future. So it might be that the singularity is nothing more than a technological horizon.

If in the XII century this horizon was maybe of 600 years, it might be that now we can’t really predict the future, with any kind of certainty, further then 50 years. This doesn’t make the horizon collapse to zero in the future. It might be the case that the singularity behaves like the speed of light. You can get very close to it without ever reaching it. You might be able to bring down the technology horizon to a few years, months or even days, but it will always be something unreachable like tomorrow is always 24h away.

The subject of singularity is of interest, mainly because of the debates it generates and because the idea forces us to us one of our most powerful tools to create new things: imagination.