The A.I. Futures Project

[John Clark]

For the last few days I've been studying the following website, there's a lot to take in but I strongly recommend you read it, you'd be hard pressed to find a better use for your time:

AI 2027

It's a detailed month by month forecast of what we can expect in the field of AI over the next four years or so and was written by some of the best tech prognosticators in the field. They provide a brief summary of their findings:

"We predict that the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution. We predict that A.I.s will continue to improve to the point where they’re fully autonomous agents that are better than humans at everything by the end of 2027 or so. In early 2027, if current trends hold, A.I. will be a superhuman coder. Then, by mid-2027, it will be a superhuman A.I. researcher — an autonomous agent that can oversee teams of A.I. coders and make new discoveries. Then, in late 2027 or early 2028, it will become a superintelligent A.I. researcher — a machine intelligence that knows more than we do about building advanced A.I., and can automate its own research and development, essentially building smarter versions of itself. From there it’s a short hop to artificial superintelligence, or A.S.I., at which point all bets are off."

They divide AI abilities into 3 categories, currently existing, emerging tech, and science fiction. Interestingly they predicted all but one of the things in the science fiction category will move into the currently existing category by 2035, most of them before 2029, the one exception is "Dyson spheres capture a significant portion of the sun's energy", it will only be in the "emerging tech" category in 2035. The authors think that the world is going to be changing even faster than what Ray Kurzweil believes, and I think they're probably right, yes sometimes it seems ridiculously fast but that is the nature of exponential growth. Other things in the "science-fiction category" are listed below and the dates when we can expect to see them move to the "existing technology" category:

AIs are better than even the best human at everything that doesn't require a body (November 2027)

Humanoid robots that are actually useful (June 2028) 

AI is able to become a superhuman computer hacker (June 2028) 

AI discovers a cure for cancer (December 2028)

A fully self-sufficient robot economy (December 2028)

AI can make mosquito sized killer drones (December 2028)

AI finds a cure for aging (between October and December 2028)

AI successfully uploads a human brain (between October and December 2028)

Drexler style autonomous replicating nanobots (December 2035)

AI has the capability of producing biosphere destroying Mirror Life (December 2035)

 John K Clark    See what's on my new list at  Extropolis

wit

 

[Cosmin Visan]

Instead of masturbating with multiple AIs, how about you have sex with multiple girls ?

[John Clark]

About two weeks ago I told you what I thought about the AI 2027. It explains what we can expect to see in the improvement in AI and when we can expect to see it. This video is what somebody else thinks about it. andit  largely corresponds with my own thoughts. 

This Former OpenAI Employee Just Revealed Everything.

 John K Clark    See what's on my new list at  Extropolis

351

[Cosmin Visan]

The end is neah!!!

[Alan Grayson]

Let's get rid of Johnny one-note. AG