I am having quite a tussle with Joerg these days, he keeps sending me these long interesting mails and I simply can't keep up with him. This time the topic is Ray Kurzweil and my Deflation Page . Basically I am arguing that it is the whole question of the unit price of information that we need to examine, and how a Moore's Law type process can extend throughout the 'information' economy. My conclusion is basically that there may well be deflationary tendencies at work here, especially if we cannot set the global economy on a significant and sustained high growth trajectory. Joerg begs to differ, but that's what blogging is all about, isn't it. Incidentally, he doesn't know yet (because I haven't explained it to anyone) that I have another argument up my sleeve, and that is the 'gift economy' type characteristics of the internet-driven information revolution (just a hint: think SETI@Home, think file sharing (not just music, but working papers etc) think Linux, think internet archive, think blogging.....).
I am absolutely convinced that you are wrong about the productivity issue. Kurzweil´s graphs and diagrams are one part of the picture. However, they do not translate directly into output data. They demonstrate how much more efficient new building-block-type products become over time - and how this development is non-linear. Assembling and configuring these products - ICs, disks, etc. - to automate production processes necessarily turns out to be a similarly non-linear process. Chinese workers and engineers have 1:20 advantage in terms of attracting investment to China as a low-wage country and securing employment in due course.
Once a company decides to be ready to automate production this comparative advantage evaporates instantly in a completely non-linear fashion. You cannot possibly automate employing last year´s technology. NASA still buys Intel´s earliest chips on Ebay which happen to be part of critical systems - they are missing out on productivity advances because they have no funds and no mandate to redesign the systems in question. The upshot is that - at any given point in time - there are large unrealized gains from technological progress which will be realized as soon as some critical threshold has been surmounted. The point is that these clustered innovations are not implemented as their isolated components are invented - "as time goes by", one might say - but as soon as the profitability threshold of the whole package has been overcome. Automation occurs not when the benefit to be had is larger than your workers´ latest wage increase, but when the benefit outweighs the twentifold advantage the Chinese currently (seem to) command. These technological gains are not prone to evaporate like unrealized investment gains on paper assets
- as long as we can be sure that we do not reverse the process of civilization and technological advancement. (That is a very real "if". Natural processes are not like that. What are the lessons to be drawn from history? What progress was there during the centuries following the fall of Rome? There is such a thing as an inbuilt cultural bias for picking one timeframe over another. Asians are less liable to sort of "daytrade away" the resources at hand. Differences in national savings rates kind of manifest our varying collective attention spans and agenda preferences.)
No comments:
Post a Comment