Repeating Nature

Our goals and hopes for AI have a lot in common with nature.

If we look at nature, it has tried a trillion trillion things and a small percentage of those have succeeded and flourished. We can look back at this and think of it as intelligent or smart design although it is ultimately survival by success rather than intentional improvement. We as humans can learn from the successes of nature’s progressions. Yet we often think we know better and repeat the failed paths already tried by nature only to come back and learn from the master innovation laboratory. We won’t necessarily see and or understand all the failures – they don’t exist anymore. In the same way, what we perceive as AI can be shown all the things humans have learned, things that work. It may not yet know how to explain everything with context. It may not yet know to try, fail, and evolve – but it will. We are seeing evidences of this – where we are programming it to behave in this manner – allowing it to try, fail, learn, advance the spiral of its knowledge. 

I don’t think we will be able to gate and corral this success cycle indefinitely. If, by then, humans are not evolving in how we learn and collaborating in our advancements, we risk being surpassed by our own creations.