Be informed about free updates
Simply log in to Artificial intelligence Myft Digest – delivered directly to your arrived mail.
Two topics dominated the annual meeting of the World Economic Forum in Davos last week: Donald Trump and artificial intelligence. Of the two, the latter was more interesting and almost certainly significant. Much attention in the discussion was dedicated Deepseeksurprise of the Chinese sigh. However, we just learned that knowledge spreads: no country will monopolize these new technology. That surprised the market. With new technologies, such “surprises” are not surprising. But that does not change the big question, which is what for all of us advanced machine intelligence. (See cards.)
Human beings are both social and intelligent. This combination is their “killer”. This allowed them to dominate the planet. Human intelligence invented general purposes that shaped the world, From a fire of fire to a computer formation. But with computers they think, that could change. Blaise PascalThe French mathematician and the 17th century philosopher, said “Man is just a reed, the most famous thing in nature, but he is a thinking reed.“Does that uniqueness approach now?
In Davos, I attended two fascinating discussions about the rewards and risks of progress in AI. One was Interview of Sir Demis HassabisGoogle Deepmind co-founder and joint recipient of the Nobel Prize for Chemistry, Roula Khalaf, Editor of FT. The other was interview of Dario Amodifounder and executive director Anthropic -Ai author Machines of love of graceaccording to Zanny Minton Beddoeseditor of economists.
The interview with Hassabis emphasized the recent progress in our ability to do scientific analysis, especially in biology. More than 2 mn researchers use AlphafoldHe said, the Deepmind program developed. “We have made all the proteins of the famous science, all 200 mn. . . [T]The rule is that the doctorate is the whole doctorate to find the structure of one protein. So, 200 mn would take a billion year of doctorate. And that’s exactly what we gave to the world, for free. “This, elaborated,” science with a digital speed. “The possibility that opened at the time is from the great acceleration in medical progress. Indeed, we may have the next 50 to 100 years of usual progress in five to ten years.
Generally, Amodei claimed, we can imagine Ai as “Genius Land in the Data Center“, What the Chinese may have done is even cheaper than before. Still, are these truly geniuses? My test would be, given all physics by 1906, but nothing after that could produce Einstein’s general theory of relativity .
It seems likely that the impact of such a problem solving, whether the “genius level” or not, should be extraordinary. This could, among other things, accelerate the improvements of knowledge and thus the growth of productivity and the spread of prosperity. Both are desirable. In the last decades, an increase in the “total factors productivity” – the best measure of technical progress – was modest. Moreover, a huge number still lives in extreme poverty and, depressed, Progress slowed down.
Still, it is also obvious that accelerated progress could create difficulties. The structure of the labor market could change massively, for example, in this case, a sharp decline in demand for workers whose assets are trained but mostly routine intelligence. The forecasts of such effects differ. A 2023 paper Erik Brynjolfsson and Gabriel Unger notes that, as the truth during the computer revolution, the effects on productivity could be modest. Still, this time it could be different, with high productivity, but appropriately to great and devastating economic and social changes. Again, depending on how society reacts, a successful AI can lead to “techno-feudalism”, with even greater wealth concentrations. The invention of a huge number of new treatments could greatly increase the costs of health care, and also the cost of dealing with a much extended life, even if they are in balance, healthier. Are people ready to live together with their big grandparents? Therefore, obviously good things could create real challenges.
In addition, the development of the intended AI creates high risks. How is his use of thunderous actors, including enemy states, terrorists and mass murderers? What moral judgments does he allow Ai to bring in the war? How is the use of AI in control? Will “Big Brother” watch us forever? Again, what will we do with the production of false and false news? How does freedom survive all these threats?
Hassabis is clear that we need effective global boundaries of AI. In the time of broken international collaboration and contempt to the very idea of a “international order based on the rules”, will China and now work together to make AI safe? It seems unlikely, not the least because they have different views on how such technologies should be used.
Back in 2015 I wrote generally skeptical article About (modest) likely to influence the productivity of new technologies. For the next few years, it would finally prove to me that I am wrong. Still I noted that we were approaching instead of “singularity” – Artificial intelligence that outweighs all human intelligence – everything must change.
One of the great ideas in Frank Herbert Dune The series is that in the distant past (our future), humanity led a successful jihad against the machines they think. After that, people had to become superhuman. A leading character explains That “people have set those machines to usurp our sense of beauty, our necessary self -coins, from which we bring living judgments. Of course, the machines have been destroyed.”
That concern could prove to be wise. But I’m realistic: Ai is out of Pandora’s box.
