Will AI hit an asymptote that prevents it from causing the upheaval predicted?

I’m open to discussion on this, and I’m not even sure I believe it. But, it’s a more optimistic view I’ve had recently pushing back against the fear that AI will render most humans obsolete.

Oftentimes we hear the narrative that whatever new technology “will only get better.” While this is may be true, I think it ignores the recent history that technology often hits a point where it gets better at a much slower rate. The derivative stays positive but the 2nd derivative goes negative.

Take TV. If you were born in the 90s like me, the TV quality completely changed in the first 10 to 15 years of your life, but since then, the changes have become much more negligible. The same could be said of video game graphics.

For example, compare the gap between 1997 GoldenEye and 2009 MWII vs the gap between ‘09 MWII and its 2020 remaster.





Yes, the 2020 remaster looks the best, but it’s still clearly a video game. Based on the ‘97 to ‘09 jump, many in 2010 were predicting video games would be indistinguishable from real life by 2020.

AI may similarly hit a plateau where it peaks as just a regurgitation of human knowledge, but never truly replicates human intelligence or even imitates humanity well enough to cause the upheaval many expect.

I believe this is the appropriate board for this post because it relates to the future of work.