The current absolute land speed record is over 1,220 km/h. The average consumer vehicle can't go anywhere near that fast. Just because simulspace acceleration up to 60x is possible doesn't mean it is cheap, easy (may require extensive implants to use), safe (may cause insanity like the Lost project, only faster), effective for R&D purposes (it might only be able to simulate a white featureless room), able to be run for more than a fraction of a second, or available to consumers (the only place in the system to get 60x acceleration might be a single research lab at UoM where they research cutting edge VR and they are only doing it to show up TAU who have only been able to achieve 58x, mostly because TAU doesn't have indenture 'volunteers' to test it on).

I think it is sensible to assume fast simulspaces are more expensive, but that doesn't mean much on a corporate level: 5000 credits is nothing on a corporate budget, even if it deters thrill-seeking private people. If you can speed up things S times at a cost of C(S) times higher (simulspace cost and, perhaps, higher salaries) you end up with finished products earlier (market positional advantage!) and can do more (productivity per person goes up by S). If C(S) is linear, then either it is almost never worthwhile to use (the exception might be if the positional advantage is very valuable), or it should *always* be used and run as fast as possible. In the more plausible case that C(S) grows as a concave function of S (maybe an exponential, say), then there will be an equilibrium point where the productivity gain of speeding up is counteracted by the higher cost of faster processing. If most of a concave C(S) comes from technology costs, then as technology develops the speed uses will increase. If C(S) is more due to overheads - salary and other HR costs, having to handle teams working on different timescales, linking internal simulations to external equipment - then things will remain more stable. But in that case different businesses will likely have very different equilibrium points: programming and indenture-using companies will have more use of fast teams than physical engineering projects using highly paid experts with family. The real economic bomb is, as I often repeat, forking. Speeding people up just makes their results show up earlier. Forking them mean that rare human capital can be made common at nearly no cost. Currently, as far as I understand the literature, that is worth much more. There is of course an optimal trade-off (if you have limited computing resources) between number of copies and speed, but multiplicity also allows neat tricks for problem solving like running several permutations of design team membership simultaneously, and then comparing their different solutions and choosing the best. What I think happens in EP is that most big companies have R&D departments that are fairly posthuman: they have a reason to use the most effective methods available to come up with great solutions before anybody else does. But this is balanced by the sheer complexity of the technology they are dealing with. They are unraveling new physics, the functions of the brain, how to manage legacy code systems with more than a century of development, and that takes enormous ability. Technology progresses, but probably not quite as neatly as anybody would want it.

Erenthia wrote:

I think Ray Kurzweil is definitely on to something with this ideas about accelerating change, but after reading his responses to critiques of his 2010 predictions, I fear he is going down a path very similar to Sigmund Freud in his complete unwillingness to see his theories and mathematical models can now be upgraded to include further data.

Well, unlike Freud, Kurzweil has made quantified predictions. Accelerating change isn't some weak unscientific hypothesis like psychotherapy. The actual pace of development will fit his predictions or not, and if they don't he'll have to adjust or look like a fool (whereas psychotherapists had the luxury of not being right for a century without being obviously wrong)

Erenthia wrote:

Kurzweil has a phenomenal opportunity to look at the data and see where his predictions went wrong, modify his models and make better predictions, but he's too attached to his previous work. It's a very unscientific attitude.

Well, this is how most people making predictions about the future handle them. A surprisingly small number of predictions are sharp enough to be tested as strictly true or false, even when they sound sharp at first. There is a whole literature on the reliability of predictions, and the short answer is that *nobody* is good at making predictions, and that both people making predictions and the people who listen to them don't seem to care much. Kurzweil at least tries to make money from figuring out when certain technologies become viable (although he also makes money from having lots of people paying for him as a great futurist - the more money he makes from this compared to tech, the less I trust his predictions). A colleague of mine has been analysing Kurzweil predictions in a bit more detail: http://lesswrong.com/lw/diz/kurzweils_predictions_good_accuracy_poor/ His view is that Kurzweil is better than many people think, but not as good as he himself thinks. He has also done a nice analysis of people's past predictions about AI: http://lesswrong.com/lw/e36/ai_timeline_predictions_are_we_getting_better/

One important thing to remember about the expectation of accelerating change is that it means you should try to invest as much as possible of what you have now - it is going to grow tremendously, and you will be able to buy very cool things with your profit in the not too far future. If you think change is slower saving becomes less important: use your capital now to get a good life. Which of course means less investment and so on. Maybe this is one of the fundamental differences between the PC (and other technophile fractions like the Extropians), and the more laid back Autonomists.