ENGINEERING INTELLIGENCE 2025 The next quarter century will most likely determine whether the purpose of human species is : multiplying 10 times more health and wealthy for all or blowing up earth. Thos who read The Economist from 1951 have had a previewbecause war time editior Geoffrey Crowther decided weekly journalism of what economists ruled made no point wiythou understanding what (and where) engineers were accelerating . Let's review (approximate- edits welcome rsvp chris.macrae@yahoo.co.uk) the 150 year journey that took us (today 8 billion beings - 4 biollion millenial generator) to this choice through bad stuff like 2 world wars and well as happier advances in intelligence engineering
| Some people feel THE NET (Neumann Einstein Turing) only gave us nuclear science.Actually within a decade of their death in mid 1950s, we had started to play with their 3 exponentially acceleraqting legacies which change humanity : chips, computers, satellites. Moreover as early as 1905 (Einstein's e=mcsqyared) it was declared that humans did no understand how the earth was energised- natures games with physical, chemical and biological science required far deeper maps than Newtonian science had systemised. | Today Nvidia has become a hundred of the world's most inspiring intelligence partnerships as well as at least 50 engineers doing their lifetimes work over the last 20 years. Jensen Huang 33 years as ceo is longer than any other valley company. While his company sells the machines deep AI is made of, he also has his own preferred foci - eg dont use chats on their own. (Intriiginly AI as cultural mediator loves translating Asian pattern languages into Western A-Z as vice versa)  In 2025 humans have not pretrained on energy dynamics the way von neumann advised. This is going to eg make costs of electricity very uneven. But first thing to do is to be transpoarent - eg if energy cost of a container ship is as much as a nvidia AI supercoputer epiecenter, lets retire a few ships and smartly map what ai contexts to advance where e  | .In 1963 the chip was developed by Intel's Noyce; within 2 years Gordon Moore offered engineers MOORES LAW - a partnership - doubling capacity of chips every 18 months; that 100 fold innovation capacity over a decade ; a million foldf by 1995. You can study where trades and what microelectronics engineers did with chips. Although Intel was based in santa clara the top of what was soon branded silicon valley and bay area, it was Japanese engineers who were intels first clients with microelectronics of calculators, electronic watches, entetainment games as well walkmans, as well as civil engineering (Tokyo is benchmark infrastructure as world's largest city and sme just-in-time supply chains- eg some machines have 5000 moving parts). The Jpaanese (today 130 million brains) followed by South Koreans (today 50 million) and Taiwanese (about 20 million) had been mentored by Deming in quality engineering; this seeks to iteratively improve societal value (now every 18 months). Deming had clashed with american (it system aint boroke dont fix it) view which is priortising continetal wide sales. So it was that Pacific trade between us west cpoast and far eastern coastal places compounded mots of moore's laws most radical value multipliocation leaps. Computers sold by IBM from 1964 werent designed round chips. So for 20 years computing stayed impersonal: mainly top down processis in big gov or big corporations. What we might cal Jensen's law starting in 1993 asked what happens if we design computers to be oppoiste of ibm central processing - faciltating many parallel interactive advabces. Japan's property bubble of late 1980s prevented if from investing in linking Jensen's research. Fortunatley his nation of bitrth as a Taiwanese American was ready. A 3rd phenomenon whih Economiost called satellite's death of disdtance in sharing data clouds also began accelerating the new century. In 2025 we have million cubed more tech to play with than 1965 and matematically alphafold3 has show some ai have billion times more brain power than individual human brains. |
| Saudi UAE India Japan Taiwan HK UK France Germany/EU Africa-South Doudna- Berkeley Colossus. | . Possible questions: whats unique data sovereignty ise india as world's largest population wanst to ai first; or souther anfrica as continetal lab for half of 2040s youth; what's smart for criticla mineral constrained isles like japan and taiwan to collab on: what will the arabian corridor as richest per capital and cheapest energy owners? can uk france and gemerany win-win with ai and renew all europe's postive dreams; which are america's unique ai computers (eg doudan biotech and science at top of valley in Berkeley; Memphis Collosus ...) |
No comments:
Post a Comment