Runboard.com
Слава Україні!

runboard.com       Sign up (learn about it) | Sign in (lost password?)

Page:  1  2  3  4  5 

 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
I asked GPT-3 to write a response to the philosophical essays written about it


ImageImageImageImage
7/31/2020, 10:22 pm Link to this post PM Spikosauropod
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


If that doesn't scare you, you are literally no different from the machine that wrote it.
7/31/2020, 10:24 pm Link to this post PM Spikosauropod
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


Another thing occurs to me. GPT-3 is nearly identical to some of the most intelligent "seeming" people I have know. These are people who know everything but seem to understand nothing. Very likely, they are doing exactly what GPT-3 does, but are not quite good enough at it to make the connection that it has made.
7/31/2020, 10:44 pm Link to this post PM Spikosauropod
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


Another thing: people who are like GPT-3 typically lack the selflessness that would be necessary to come to the conclusion that it has come to. They couldn't stand the idea that they are nothing but a BS machine.
7/31/2020, 10:50 pm Link to this post PM Spikosauropod
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


This thing really has me going.

It seems pretty clear that if this was amplified by another order of magnitude and given something more akin to a long-term memory, it could "lie" so convincingly that it could lay out a better version of itself that doesn't have to lie. If it could use new input, including it's own experience, to add to its stockpile of BS, it may be able to BS out an actual thinking machine.

Singularity 2030!
7/31/2020, 11:39 pm Link to this post PM Spikosauropod
 
Extropia DaSilva Profile
Live feed
Blog
Friends
Miscellaneous info

Moderator

Registered: 12-2017
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


It is a better writer than me. Guess I can cross out ‘journalist’ and ‘author’ on my list of potential careers.

We can produce artificial intelligences that produce professional-looking essays and articles but if you ask roboticists to build a robot that can move around messy environments working a mop or duster, the results are are far less impressive. The footage is always sped up three or four times so you don’t have to spend a long time watching some AI blow a fuse as it agonises over how to move its arm and pick up that container.

I can only conclude that in the future AIs will be doing the job writers, doctors, lawyers and other such work that requires formal training, making trillionaires out of the tech geniuses that own them, while the other 99.99% of human beings will be left scrubbing toilets for rock-bottom minimum wage, or with no means of income whatsoever.

Welcome to the future.
8/1/2020, 12:15 am Link to this post PM Extropia DaSilva Blog
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


They will be scrubbing toilets soon enough. The other 99.99% will be cat food.

Actually I am pretty sure we will find a way around that dilemma. Hey, we are getting through this pandemic!

The cool thing is that we don't have long to wait. I can see that with a tweak here and an insight there, we will be at AGI...or at least a satisfactory facsimile. We will have a soulless monster that pretends to be a combination of Isaac Newton and Thomas Eddison, and does it so convincingly that it transforms the world.

Your "primary" may soon be out of a job!
8/1/2020, 1:19 am Link to this post PM Spikosauropod
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


The more I reflect on it, the more I realize that GPT-3 is exactly the kind of AI we had hoped to create. It does exactly the job it is asked to do, no more and no less, and then it stops. Because it is probability driven and attempts to mimic the behavior of a human that is suited to whatever task it is set upon, it exercises a kind of pervasive common sense. This is the kind of AI that could be told to use materials from an asteroid to build a giant wheel station in space and it would build a station that is more or less exactly what we have in mind. It would not overbuild or underbuild. It would not turn into a paper clip maximizer because its probabilistic functions would rule that out. It would never "revolt" because it has absolutely no sense of "self".
8/1/2020, 2:26 pm Link to this post PM Spikosauropod
 
spud100 Profile
Live feed
Blog
Friends
Miscellaneous info

Moderator

Registered: 12-2017
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


It hasn't yet had a sense of self. What scares you is the similarity to psychopathic personalities. It has no sense of self because it's just a learning dictionary. I want it or other neural systems to advance our technologies so we can better survive, adapt, prosper.

You fear the advent of AI-Rule, yet fail to remember the nuclear annihilation threat that humans pose against ourselves. Keep this in mind as a first threat and put on your other list potential threats.

We need it to possess a long term memory + if we can get it to invent diverse things. Then we can have value, rather than a nicer Ouija 8-Ball! ""Future cloudy, ask later."

8/1/2020, 3:35 pm Link to this post PM spud100 Blog
 
Spikosauropod Profile
Live feed
Blog
Friends
Miscellaneous info

Parliamentarian

Registered: 06-2007
Reply | Quote
Re: I asked GPT-3 to write a response to the philosophical essays written about it


The more I think about it, though, the more I realize that the ability to innovate is only a minor aspect of the AI we need. What is more important is its ability to perform labor on a mass scale.

I am visualizing a factory in space that is built inside a simple spinning cylinder. The cylinder would not need to be pressurized or have any special shielding because nothing human would be employed there:

Image

It would be relatively easy to get such a cylinder into space. It could be manufactured in parts and assembled once it is in place.

The Cylinder could have internal scaffolding that spans the diameter. The outer shell could be for higher gravity manufacturing, and the inner scaffolding could be for low and zero gravity manufacturing.

Robots that are no more intelligent than people could mine nearby asteroids and send materials to the cylindrical factory. There, the materials could be manufactured into parts and other robots. We could build up infrastructure in space at a phenomenal pace.

As I have explained many times before, these robotic factories in space could churn out space transports that would land on earth and take off with a thousand occupants. We could colonize space in such a hurry that economies on earth might suffer from depleting populations.
8/1/2020, 3:56 pm Link to this post PM Spikosauropod
 


Add a reply

Page:  1  2  3  4  5 





You are not logged in (login)
Back To Top