In the previous article in this series, we talked about the future of AI. To summarize, many people learn AI today. In fact, too many. And when the crowd does something, they are probably latecomers and the train has left.
In this article, we will try to justify this gut feeling. There is this recent (a few months old) article by Jerome Pesenti, the VP of AI at Facebook, “Facebook’s Head of AI Says the Field Will Soon ‘Hit the Wall’”.
But who is Jerome? He has a degree in Philosophy from Sorbonne, another degree in Cognitive Science, and a Ph.D. in Math. He then founded an AI company and after twelve years sold it to IBM, becoming a VP of Watson Platform in the process. And now Jerome is VP of AI at Facebook, replacing Yan LeCun.
So now let us read what he says. While giving the deserved due to the success and significance of AI at Facebook, Jerome says that AI will hit the wall soon or already has. “Deep learning and current AI, if you are really honest, has a lot of limitations.”
- “The compute power required for advanced AI is doubling every 3 and a half months.” This means that the progress we can make will reach the limit very soon.
- The cost of AI grows. “Right now, an experiment might be in seven figures, but it’s not going to go to nine or ten figures, it’s not possible, nobody can afford that.” Not even Facebook, much less the others in the world.
- Making AI practically successful is very hard and requires a new organizational approach. “When you start talking about technology transfer, it means you’re already lost the battle.”
Does this sound like a rumbling of a still distant but fast approaching and powerful thunder? If it does, then please watch for the next issue in the series where I give an outline of a plan and a solution. It is based on further analysis, on the success of a class that we are now running, and it will give you a way to adjust the direction and realize the ROI from what you already invested in AI and Computer Engineering. À la prochaine!