[ExI] Fwd: The present and future of AI

[Jason Resch]

(Sorry re-sending, sen't to wrong address)



On Fri, Jun 12, 2020 at 3:14 PM Robin D Hanson <rhanson at gmu.edu> wrote:

> Yes of course I’m considering exp progress in stuff that has been exp, and
> linear progress in stuff that has been linear.
> The hard part about human AI is doing every thing WELL. You can’t really
> pick an ability X and say “no AI can do X”.
> Almost always, there is some system that does X, except just very BADLY.
>

I came up with the following list of human-specific traits of intelligence:


   - Communicate via natural language
   - Learn, adapt, and grow
   - Move through a dynamic environment
   - Recognize sights and sounds
   - Be creative in music, art, writing and invention
   - Reason with logic and rationality to solve problems

Is there anything important that I've forgotten?  Because there has been
major progress in each of these domains. It's hard to imagine that after a
1,000 or million-fold increase in compute they won't be vastly superior.

This is where we stand today, if we put everything together in one system:

"We would get an AI that could hold a conversation in any of 100 languages
over the phone. It would beat everyone at Jeopardy, Chess, Go, Poker, as
well any old video game. The AI would be able to recognize any object or
face — even detect cancer better than most doctors. It would also be
accomplished and creative, having invented things, discovered laws of
physics, and identified new drugs. The AI could compose as well as Bach and
paint as well as van Gogh. It would also be seen as highly original in its
own artistic styles."

‘Let’s put all these together,’ and then it will be smart. -- Marvin Minsky


Jason



>
> On Jun 12, 2020, at 3:12 PM, Jason Resch <jasonresch at gmail.com> wrote:
>
> On Fri, Jun 12, 2020 at 1:57 PM Robin D Hanson <rhanson at gmu.edu> wrote:
>
>> There have been substantial advances in the last decade. But then there
>> have been substantial advances every decade since 1950. We are still a LONG
>> way from broad human level abilities; it will take a LOT longer than 10-30
>> years to get there.
>>
>> What do you see as the most difficult aspects of human ability we have
> yet to achieve?
>
> Also, are you considering exponential progress in your 10-30 year
> projection over the future (which could see a billion-fold increase in
> computing power?
>
>
> Robin Hanson rhanson at gmu.edu
> Future of Humanity Inst., Oxford University
> Assoc. Prof. Economics, George Mason University
> See my books: http://ageofem.com http://elephantinthebrain.com
>
>
>
>
>
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20200612/b71e146d/attachment.htm>