#ai #ai-alignment # [[Epistemic status]] #shower-thought #to-digest 2022-07-04 # Related - [[Computing/Intelligence/Intelligence has always been about generalisation]] - [[Computing/Intelligence/Alignment/Intelligence explosion]] - [[Computing/Intelligence/Artificial intelligence knows us better than we know ourselves]] - [[Computing/Intelligence/Alignment/Availability bias is strong in AI alignment research]] # TODOs # Pessimistic vs optimistic [[Artificial intelligence|AI]] ## Narrow [[Philosophy/Rationality/Intelligence|intelligence]] is fast, general [[Philosophy/Epistemology/Knowledge|knowledge]] is slow Pessimistic and optimistic AI are two different ways of looking at computer intelligence. Pessimistic AI means that computers can only do certain things fast, like adding numbers, but they can't learn new things quickly. Optimistic AI means that computers can learn new things quickly. ![a beautiful ultradetailed vintage photo of a futuristic cybernetic cyborg male wearing a tall white chef hat and an apron, by tom bagshaw and anna dittman, portrait, 3 5 mm lens, golden ratio composition, detailed face, studio photography, very detailed, humanoids, industrial robots, artstation, 8 k, highly coherent ](https://lexica-serve-encoded-images.sharif.workers.dev/md/01fb5868-188a-4b53-ab5f-414d2033f677) >Although we can make intelligent machines that run a million times faster than a biological brain, they cannot acquire new knowledge a million times faster. >~ [[Jeff Hawkins]] ![front shot of a court jester!!!! character, polished, glossy, holographic crystals, bionic futurism, robotic cyborg, 8 5 mm f 1. 8, ominous elegant detailed concept artstation sharp focus ray tracing cinematic masterpiece fine portrait, by artgerm greg rutkowski tooth wu beeple ](https://lexica-serve-encoded-images.sharif.workers.dev/md/0fa7e933-bb64-43cf-b6bd-380e15cf5a18) Jeff Hawkins said that although computers can do certain tasks a million times faster than humans, they can't learn new knowledge a million times faster. That means it takes a lot of time to learn new things. Even though computers are one million times faster than humans for narrow tasks, it takes them a long time to learn new knowledge.