Artificial Intelligence - General Intelligence Breakthroughs

For discussion of liberty, freedom, government and politics.
Post Reply
endlessQuestions
Level 34 Illuminated
Posts: 6627

Artificial Intelligence - General Intelligence Breakthroughs

Post by endlessQuestions »

Y’all really need to understand this. AI has broken several boundaries people were saying it was going to take years to get to. The world is changing faster than we can possibly comprehend.
For those who can’t read things more than 5 pages long or watch things more than 5 minutes in duration, I apologize in advance. The good news is, the video does a great job of explaining how you have just been surpassed by technology you’ll likely never understand, except to know that one day you had a job, and the next you were replaced by something that has no limitation regarding how much it can “concentrate” and learn.

Link to the paper:

https://arxiv.org/pdf/2303.12712.pdf

User avatar
Wolfwoman
captain of 1,000
Posts: 2355

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Wolfwoman »

I watched the video. I didn’t even know that there was a GPT4 vs. a ChatGPT4.

It can create propaganda and conspiracy theories.

In the video he mentions an “embodied” GPT-4. That’s truly frightening!

User avatar
MikeMaillet
captain of 1,000
Posts: 1727
Location: Ingleside, Ontario

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by MikeMaillet »

Just a few days ago some of the creators of this technology displayed photographs of Donald Trump being arrested but the photos were 100% made up. Also, technology is now available that can change anyone's voice to make it sound like someone else. The time has arrived where we will not be able to judge any news story for fear that it was all made up by computer.; not just the words but the visuals as well.

Unless angels start appearing to me, the only voice I can trust is the still, small voice.

Mike

endlessQuestions
Level 34 Illuminated
Posts: 6627

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by endlessQuestions »

MikeMaillet wrote: March 27th, 2023, 3:58 am Just a few days ago some of the creators of this technology displayed photographs of Donald Trump being arrested but the photos were 100% made up. Also, technology is now available that can change anyone's voice to make it sound like someone else. The time has arrived where we will not be able to judge any news story for fear that it was all made up by computer.; not just the words but the visuals as well.

Unless angels start appearing to me, the only voice I can trust is the still, small voice.

Mike
I think that’s the key takeaway here. Thanks for pointing it out.

Interestingly, this would be a fulfillment of President Nelson’s words from several years ago, yes?

The question would be, were his words uttered through the spirit of prophecy, or uttered because he learned what was coming by meeting with Trump’s NSA advisor and fellow Latter-day Saint Robert C O’Brien, Jr. in 2019 (if my memory serves me right), who may have tipped him off as to what was coming…

Fascinating times we live in…

User avatar
Niemand
Level 34 Illuminated
Posts: 14231

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Niemand »

The obvious solution to a lot of these issues is to MEET IN PERSON and have PHYSICAL FACILITIES, but off go the greedy and do the exact opposite.

User avatar
Niemand
Level 34 Illuminated
Posts: 14231

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Niemand »

Short video (8 mins) on an AI art generator

User avatar
TheDuke
Level 34 Illuminated
Posts: 5923
Location: Eastern Sodom Suburbs

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by TheDuke »

so, lets be clear. The only true leap forward here is having enough power and memory/speed to process same old inferences faster. Then to have a great language processor to make the often meaningless results seem eloquent. Lastly you get many, many people playing so it gets filled with detailed facts (right or wrong).

What made the chess computers smarter? processing speed. same algorithms. Actually as speed increases the algorithms for chess got simplified as the older ones needed weighted heuristics.

What made Walmart successful? Volume and dumping money. All volume. What made google successful? Processing speed, access to broad areas of the Internet and many users.

Same with AI that speaks. What jobs will Chat like AI take over? Well, telephone jobs where you talk to a number of useless people. Usually useless because they have no authority to correct things, even the supervisor. Robots can build things but not AI. Opposite, you need deterministic computation to operate large robotics near people or people get dead (crushed, hit, drilled, hammered, etc...).

Got to quit convoluting AI, robotics, natural language processing and image processing.

I say for most lost jobs, the problem will be for the Chinese or Indian folks to move to another company as most of those jobs are already overseas.

User avatar
Niemand
Level 34 Illuminated
Posts: 14231

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Niemand »

TheDuke wrote: March 28th, 2023, 9:19 pm so, lets be clear. The only true leap forward here is having enough power and memory/speed to process same old inferences faster. Then to have a great language processor to make the often meaningless results seem eloquent. Lastly you get many, many people playing so it gets filled with detailed facts (right or wrong).

What made the chess computers smarter? processing speed. same algorithms. Actually as speed increases the algorithms for chess got simplified as the older ones needed weighted heuristics.

What made Walmart successful? Volume and dumping money. All volume. What made google successful? Processing speed, access to broad areas of the Internet and many users.

Same with AI that speaks. What jobs will Chat like AI take over? Well, telephone jobs where you talk to a number of useless people. Usually useless because they have no authority to correct things, even the supervisor. Robots can build things but not AI. Opposite, you need deterministic computation to operate large robotics near people or people get dead (crushed, hit, drilled, hammered, etc...).

Got to quit convoluting AI, robotics, natural language processing and image processing.

I say for most lost jobs, the problem will be for the Chinese or Indian folks to move to another company as most of those jobs are already overseas.
I've thought about this many times. There is barely a job between Accounting and Zoo keeping which some robot/computer won't be able to do within the next few years. 2050 according to one estimate, perhaps much less at the current rate. The Chinese and Indians will have the same happen to them.

It is sad to see this happen in the arts, but it proves that most people's cultural/artistic consumption is repetitive, unoriginal and predictable. I don't believe AI "gets" creativity, but it produces a simulacrum of it.

The Luddites worked out this process early on when their physical jobs were being replaced. Since then technology has provided some new jobs (such as IT), but currently the replacement rate isn't high enough and the pool of jobs is decreasing.

In an ideal world, sensible rulers would work out this situation could benefit everyone, and we would improve our minds and lives through experiences we would not have if tied down by work. But we don't live in an ideal world.

The end result will be billions of angry unemployed and a few hundred (maybe thousand), living in their castles and compounds, and luxury islands, off accumulated wealth. That isn't a good ratio and the powerful people know that. One solution is to reduce our numbers in their favour, another is to promote infighting among the masses, another is to corral us into small ringfenced areas, another is to put us into virtual reality/sedate us with "drugs and video games" (as Noah Yuval Harari calls it). I see all of those solutions currently happening at some level. The transgender thing has a dual/triple purpose perhaps — voluntary sterilisation, creating infighting and also a soft-peddle into transhumanism.

User avatar
TheDuke
Level 34 Illuminated
Posts: 5923
Location: Eastern Sodom Suburbs

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by TheDuke »

Niemand wrote: March 29th, 2023, 4:24 am
TheDuke wrote: March 28th, 2023, 9:19 pm so, lets be clear. The only true leap forward here is having enough power and memory/speed to process same old inferences faster. Then to have a great language processor to make the often meaningless results seem eloquent. Lastly you get many, many people playing so it gets filled with detailed facts (right or wrong).

What made the chess computers smarter? processing speed. same algorithms. Actually as speed increases the algorithms for chess got simplified as the older ones needed weighted heuristics.

What made Walmart successful? Volume and dumping money. All volume. What made google successful? Processing speed, access to broad areas of the Internet and many users.

Same with AI that speaks. What jobs will Chat like AI take over? Well, telephone jobs where you talk to a number of useless people. Usually useless because they have no authority to correct things, even the supervisor. Robots can build things but not AI. Opposite, you need deterministic computation to operate large robotics near people or people get dead (crushed, hit, drilled, hammered, etc...).

Got to quit convoluting AI, robotics, natural language processing and image processing.

I say for most lost jobs, the problem will be for the Chinese or Indian folks to move to another company as most of those jobs are already overseas.
I've thought about this many times. There is barely a job between Accounting and Zoo keeping which some robot/computer won't be able to do within the next few years. 2050 according to one estimate, perhaps much less at the current rate. The Chinese and Indians will have the same happen to them.

It is sad to see this happen in the arts, but it proves that most people's cultural/artistic consumption is repetitive, unoriginal and predictable. I don't believe AI "gets" creativity, but it produces a simulacrum of it.

The Luddites worked out this process early on when their physical jobs were being replaced. Since then technology has provided some new jobs (such as IT), but currently the replacement rate isn't high enough and the pool of jobs is decreasing.

In an ideal world, sensible rulers would work out this situation could benefit everyone, and we would improve our minds and lives through experiences we would not have if tied down by work. But we don't live in an ideal world.

The end result will be billions of angry unemployed and a few hundred (maybe thousand), living in their castles and compounds, and luxury islands, off accumulated wealth. That isn't a good ratio and the powerful people know that. One solution is to reduce our numbers in their favour, another is to promote infighting among the masses, another is to corral us into small ringfenced areas, another is to put us into virtual reality/sedate us with "drugs and video games" (as Noah Yuval Harari calls it). I see all of those solutions currently happening at some level. The transgender thing has a dual/triple purpose perhaps — voluntary sterilisation, creating infighting and also a soft-peddle into transhumanism.
Probably true, but only to AI, at least in mass if we all somehow agree AI is same as people and interactions with humans are of no value. Perhaps not true for robotics and they can become more profound at doing repeatable jobs.

But theoretically, in a perfect world, it just frees up people to do more useful and creative things. However, what we see is the tech in a few hands in $with multiple B in it, and many hundreds of thousands out of work, as they wait for the government to pay off their college bills for 8 or 9 years of undergraduate work to get a degree in Transgender Studies.

User avatar
Fred
Level 34 Illuminated
Posts: 7789
Location: Zion

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Fred »

Where it's going...

User avatar
Niemand
Level 34 Illuminated
Posts: 14231

Re: Artificial Intelligence - General Intelligence Breakthroughs

Post by Niemand »

endlessQuestions wrote: March 26th, 2023, 9:00 pm For those who can’t read things more than 5 pages long or watch things more than 5 minutes in duration, I apologize in advance.
People post videos on here which are everything from 30 seconds to five hours long, and more. They often give you no idea of how long they are. Fred's video in the last post there is half an hour long, for example, which is not horrific, but I would have had no idea unless I'd scrolled to the end of the video link to find out.

For obvious reasons it isn't always a good use of time to watch very long videos. Some of them are very long, but very thin on content or anything of use. I know this because I made the mistake of watching a few years ago. Some people have to learn the art of summarising.

In fact it is better if a long message is broken into segments so people can do all the things they need to do in their lives.

Post Reply