AI Fatigue: Reflections on the Human Side of AI’s Rapid Advancement

A Communications of the ACM article (link) with that title by Victor Dibia, PhD of Microsoft Research summarizes how many people in the field (people whose job it is to keep up with AI research) feel, especially if they are older than 25 (I am, barely). I already reported feeling as overwhelmed as Andrej Karpathy, so it is nice to see the ailment recognized "officially".

AI fatigue, defined as "the collective exhaustion experienced by individuals and organizations in response to the unrelenting pace of artificial intelligence (AI) advancement [...], reflects the mental, emotional, and operational toll of trying to adapt to an unprecedented rate of change that has sustained for a relatively long period (several years) with little/no signs of slowing down."

The Three Firehoses of AI Progess: Papers, Models, and Announcements

Drinking daily from the firehoses used to be exhilarating. With thousands of new AI-related papers on ArXiv every day, new models popping up almost daily, benchmark-palooza and flashy announcement, the shock and awe of AI is taking a toll. The worst impact of all is the mental strain of having to keep up (move up or move on). When you spend that much time and effort keeping up, there is little time left for your own work, research and thoughts.

Unlike AI, humans need time to adapt and assimilate new concepts, ideas, technologies. Imagine how we will feel when AI starts really self-improving.