"me too. I really want to go further back in time. And, although I have some power consumption numbers from further back, I don't have access to the BEST power consumption numbers. I am not sure that people really kept those sorts of stats back then. Power wasn't really the issue of consideration back then. 

If anyone has access to better data, then by all means, let me know. "

"James - another very interesting chart,  with a long, historical perspective. It will be very interesting to see what comes from Henry Markram's Human Brain Project in the next few years."

"James - an interesting chart, I wonder what the longer term trend is? I believe you are correct about the cost of power, that is also why Google server farms are located near cheap sources of electricity."

"More data like this is available here: http://www.kurzweilai.net/the-computing-trend-that-will-change-everything?utm_source=KurzweilAI+Daily+Newsletter&utm_campaign=b0c3dd0f7e-UA-946742-1&utm_medium=email"

"very very interesting thanks! I think that there is little doubt that we will see a systematic flattening of the improvement rate for magnetic storage hdd's by 2020 at the latest.

But I am far less certain that SDD costs will flatten too by the same time. I think that their improvement rates will continue for some considerably longer time (just a guess though).

If I am right, all that means is that SDD and HDD price points cross sooner than my graph projects. Since I was already projecting that SDDs would eventually dominate, this shouldn't much affect the long term trend in price/tb. But it will be very interesting to see what really happens."

"I will get back to this after Christmas break. If you find anything substantial before then, let me know."

"I would love to find more detailed data. If you find some let me know. I will look too. That would be nice."

"Well, I love this graph a lot, but the x-axis is too wide. Does that link give you enough to make one from, say, 1500 to present?"

"Looks to me like they are still going up everywhere according to the article you linked to, just slower in the US. But as a futurist, I am not US centric. If the US falls behind, progress will continue somewhere else. I would love to see it happen in the US, but I am most interested in insuring that it happen.

If the problem is obesity as the article claims, we may have a solution for that coming down the pipeline (a cure for obesity isn't impossible). And Americans who care can do something about it now (exercise and diet). If its something else, we will have to see what can be done. But what we know now is that activity is the single most important predictor of life expectancy and health. For now, until the cure comes, it's time to get off the couch and get busy.

In any event, the future will happen, somewhere."

"If you can figure out a typical utility/value for space and speed for each application, I can do that. :-) The question is, how do you define "typical"?"

"And even if the end of Moore's Law were such a limitation, we will have already passed the threshold that would lead to full brain simulation before then, assuming a synapse level simulation is sufficient, (which seems likely). In other words, the end of Moore's Law will come too late to pose any real threat to the creation of machines with truly human levels of intelligence."

"Moore's law may well hit the quantum limits sometime around 2020-2025, but it should be pointed out that this limit is not a fundamental limit to super computer performance, (the brain serves as an existence proof that it is possible to do about 10^19cps on just 20 watts of power in a very small space)."