A welcome unpacking of a fuzzy concept, helping clarify the term "singularity" and other terms that get tossed around without much of an effort to explain some distinctions. As usual, Max More is at the forefront of human and machine evolution, making difficult ideas understandable. I look forward to reading the upcoming installments in this series.
"It is important to discuss whether we should expect an IE once we have AGI or SAI, and whether we can do anything to maintain AC but avoid IE and a PH."
Yes. And Transhumanism is essentially the answer to that question, which makes Transhumanism a moral mandate insofar as human morality is concerned.
Is it really true that technological change is accelerating? Consider two things: everyday life and productivity.
Over the past 50 years, we’ve seen digital technologies transform our world: the personal computer, the internet, the mobile phone, and ultimately the smartphone, a convergence of the three. But between 1925 and 1975 (another 50-year span) we saw the rise of radio, television, air conditioning, the flush toilet, and the full range of household appliances — refrigerators, vacuum cleaners, stoves, washing machines, and more. Which period brought a greater transformation to daily life?
The productivity data tell a similar story. In the United States, productivity grew by 110 percent between 1950 and 1985, but by only 79 percent between 1985 and 2020 (two equally long 35-year periods). From 1990 to 2005, growth was 39 percent, but just 21 percent from 2005 to 2020. These are not cherry-picked intervals: no matter which two consecutive periods of equal length we compare (since 1950), the earlier one consistently shows faster productivity growth than the later one—if we were truly approaching the singularity, we would see the opposite.
It would be difficult to quantify overall technological change, rather than in specific sectors. I agree that the picture is not clear and the trend of recent decades does little to support the singularity idea as a smooth exponential. However, many singularity enthusiasts expect a break in the trend -- singularity as phase change rather than as an exponential. You are correct about the lack of a trend toward faster economic growth. Possibly that will change with AI but so far we have seen no sign.
Unique perspectives. A truly enlightened piece.
Wait a minute. Gary Marcus doesn't think LLMs are the path to AGI??
;-)
I know. It's hard to tell with Gary. He is very shy and doesn't like to speak his mind. :-)
A welcome unpacking of a fuzzy concept, helping clarify the term "singularity" and other terms that get tossed around without much of an effort to explain some distinctions. As usual, Max More is at the forefront of human and machine evolution, making difficult ideas understandable. I look forward to reading the upcoming installments in this series.
"It is important to discuss whether we should expect an IE once we have AGI or SAI, and whether we can do anything to maintain AC but avoid IE and a PH."
Yes. And Transhumanism is essentially the answer to that question, which makes Transhumanism a moral mandate insofar as human morality is concerned.
Excellent overview and clears up some muddy thinking out there ... thank you! [Totally agree on LLMs]
Is it really true that technological change is accelerating? Consider two things: everyday life and productivity.
Over the past 50 years, we’ve seen digital technologies transform our world: the personal computer, the internet, the mobile phone, and ultimately the smartphone, a convergence of the three. But between 1925 and 1975 (another 50-year span) we saw the rise of radio, television, air conditioning, the flush toilet, and the full range of household appliances — refrigerators, vacuum cleaners, stoves, washing machines, and more. Which period brought a greater transformation to daily life?
The productivity data tell a similar story. In the United States, productivity grew by 110 percent between 1950 and 1985, but by only 79 percent between 1985 and 2020 (two equally long 35-year periods). From 1990 to 2005, growth was 39 percent, but just 21 percent from 2005 to 2020. These are not cherry-picked intervals: no matter which two consecutive periods of equal length we compare (since 1950), the earlier one consistently shows faster productivity growth than the later one—if we were truly approaching the singularity, we would see the opposite.
It would be difficult to quantify overall technological change, rather than in specific sectors. I agree that the picture is not clear and the trend of recent decades does little to support the singularity idea as a smooth exponential. However, many singularity enthusiasts expect a break in the trend -- singularity as phase change rather than as an exponential. You are correct about the lack of a trend toward faster economic growth. Possibly that will change with AI but so far we have seen no sign.