AI and the future

[surprising]

In the last couple of years, what was once the realm of science fiction is now becoming a reality. The rate of progress is staggering, and I am starting to wonder what the world looks like in 5, 10, 20 years. Let's discuss possible futures.

There will come a time when we perfect robots. These robots will be able to replace many human jobs. Many many jobs will be no longer necessary, and many many people will be out of work. The service industry is like 80% of the jobs in the US (according to ChatGPT). How do we survive this? Do you provide a basic income for everyone at that point? What will it mean to even be human? Without need of a job, will humans pursue base pleasures, or will they pursue interests that better themselves and/or the world without worry of money?

AI infrastructure will continue to be dominated by the top tech companies and potentially one will win out and control all of AI (like Buy n Large in Wall-E). My bet would be Amazon as they already have the infrastructure to provide all of us humans anything we want overnight. They will provide anyone with a Prime membership a fully customized home robot, built based on your decades of Amazon purchases. It'll know you better than you.

 

[GrayHare]

Humans will be needed to improve and repair the robots. Automobiles put wheelwrights out of business but also created new jobs.

 

[braumeister]

They could build a flying car in the 1950s, but I'm still waiting for mine.
So I've learned not to expect much.

 

[Sojourner]

There will come a time when we perfect robots. These robots will be able to replace many human jobs. Many many jobs will be no longer necessary, and many many people will be out of work.

While these statements are absolutely true, they don't quite convey the breadth and depth of the coming AI revolution. Not only will robots be "perfected" to the level where they can replace human jobs, AI itself is on track to become unimaginably powerful in a more general sense. In fact, robots will only be perfected to the level of near-perfect human replicas after the dawn of artificial general intelligence (AGI). Until that time, AI will continue to be specialized and limited in its capabilities. Specialized AI, along with specialized robots, will become more and more powerful and capable, replacing more and more human jobs and endeavors over the coming years. The field of radiology is a good example. Specialized AI that can read X-rays and other medical images with much greater precision and accuracy than any human will become ubiquitous over the coming decades, completely replacing all human radiologists.

AGI, however, will by definition do every job, task, and human endeavor better than any human. This is because AGI will be capable of self-analysis and self-improvement at speeds vastly beyond anything humanly possible. Once AGI comes into existence, it will rapidly evolve itself to superhuman AI (and continue to do so), outpacing humans in all conceivable domains of expertise.

The service industry is like 80% of the jobs in the US (according to ChatGPT). How do we survive this? Do you provide a basic income for everyone at that point? What will it mean to even be human? Without need of a job, will humans pursue base pleasures, or will they pursue interests that better themselves and/or the world without worry of money?

Many AI researchers and futurists are concerned about this very topic. The concept of universal basic income is being widely discussed and debated among this group. I've read various books and articles postulating that UBI is inevitable for our global society once AGI comes into existence, because there literally will not be one person who can do even one job better than superhuman AGI and "perfect" robots that will be ubiquitous.

 

[rembrandt]

It's worse than that. Since the machines will do all of the thinking for us, our brains will atrophy from not being challenged. This will result in de-evolution. It's already starting with cell phone apps doing many tasks for us. Not a pleasant thought.

 

[plank2505]

Energy generation and transmission is a major bottleneck to scaling AI to widespread practical uses.

 

[GrayHare]

"Combined with a form of fusion, the machines had found all the energy they would ever need." AI has been well covered in fiction. I expect neo-Luddites (pun intended) will block AI from taking over everything.

 

[rembrandt]

"Combined with a form of fusion, the machines had found all the energy they would ever need." AI has been well covered in fiction. I expect neo-Luddites (pun intended) will block AI from taking over everything.

"I'm sorry, Dave, I'm afraid I can't do that"

 

[GenXguy]

I'm not worried about any of this stuff. I'm still waiting for the time when ChatGPT and other paid IA models can actually answer a majority of my basic questions correctly.

 

[plank2505]

I'm not worried about any of this stuff. I'm still waiting for the time when ChatGPT and other paid IA models can actually answer a majority of my basic questions correctly.

Take a look at this (sober) summary of the current state of these models. The o1 model is much advanced and internal claims (that cannot yet be publicly verified) of results from the o3 model shows significant advancement over o1.

Prophecies of the Flood

What to make of the statements of the AI labs?

www.oneusefulthing.org

The tech looks like it will get there and faster than most would have thought. The problem is we don’t have the power to use the tech at scale.

 

[OldShooter]

In my experience, mostly in high tech, the latest greatest technology has never managed to achieve what is near-in (a few years) has been forecast for it. (Anybody remember magnetic bubbles?). My current amusement is to wonder when we will start referring to "the AI bubble." My guess is a couple of years.

 

[plank2505]

In my experience, mostly in high tech, the latest greatest technology has never managed to achieve what is near-in (a few years) has been forecast for it. (Anybody remember magnetic bubbles?). My current amusement is to wonder when we will start referring to "the AI bubble." My guess is a couple of years.

Your smartphone has more computing power than most expensive and powerful computers 30-40 years ago.

Technology has completely restructured our lives in ways we could not have imagined decades ago.

You may not feel this way because (a) it’s hard for people to remember the details of how life was back then, and relatedly (b) the tech is so common and ubiquitous that we take it for granted.

It’s often only when we compare specific tasks then and now that we can appreciate the transformation.

The same is true for standard of living. Our standard of living has improved dramatically and consistently, but you will hear people talk as if things are tougher/unaffordable now than in the past. This is because people misremember the past standard of living and take the positives of today for granted.

 

[OldShooter]

None of that has anything to do with what I said and, actually, when working I lived in the world of computers, microelectronics, and spacecraft. My customers (like JPL) really were rocket scientists. Your issues are yours, not mine.

 

[JoeWras]

I guess it depends...

Airplanes made absolutely incredible advancement in 20 years, or heck, 5 years. If you look at the Wright Flyer's first flight in December 1903 in North Carolina (First in Flight*), and then move forward to 1910, the advancement was unbelievable.

The thing was, at that time, there were also huge parallel advances in materials (wings) and engines. A lot of people put their heads together on multiple disciplines and made big advancements. Sadly, most of this was the for The Great War (WWI) effort.

I'm not quite seeing that with AI, yet, but I think there is a chance we'll see it go asymptotic. I hope I'm dead by then.

BTW, I don't want to poke a bear, but we had multiple threads about 8 years ago about self driving cars. Most of those threads became very contentious and were closed. Today, we can't open a thread about self driving cars because of the controversy. What am I saying? I'm saying that it has gone slower than many expected. I was a doubter, but that means nothing if AI really works.

If AI works its real magic beyond the marketing words, self driving can happen. I'm still dubious about a lot of AI because I see the term employed too much for marketing and not real work. Real AI exists, but it is not widespread. But it is expanding and when it really kicks in (my opinion, 10 years) watch out. Right now, it is is an overused term, just as self driving was a few years back, and maybe even a bit today for Tesla's software roll outs.

* - "First in Flight". Actually, Ohio, we love you. Let's be friends. You are the Birthplace of Aviation, but NC was First in Flight, like it or not. Ha ha ha ha.

 

[GrayHare]

I'm not quite seeing that with AI, yet, but I think there is a chance we'll see it go asymptotic. I hope I'm dead by then.

Agree. Between advances in fusion (lots of energy) and quantum (lots of computing without huge energy consumption), AI won't be held back by a lack of power for long. Heck, it might even foster itself by solving challenges in those fields.

 

[Gumby]

.....

* - "First in Flight". Actually, Ohio, we love you. Let's be friends. You are the Birthplace of Aviation, but NC was First in Flight, like it or not. Ha ha ha ha.

Sorry, Connecticut is the birthplace of powered aviation. Gustave Whitehead flew the No. 21 in Bridgeport, Connecticut on August 14, 1901.

 

[JoeWras]

I hear rumors that at the summer Olympics in Brazil, they made a very, very strong point that Brazil invented powered flight. Go home Connecticut.

But back to the topic, because it ties in. I think powered flight is a good analogy what for what is going on with AI. There are many players: Ohio, NC, Brazil, France (never forget France, seriously), and so on. And then one day, when nobody is thinking about it, there are airplanes buzzing about and rich guys putting out rewards to fly across the USA, or to Hawaii, or to Europe.

We are seeing this with AI right now. How far away are we from the AI analogy to cross-Atlantic passenger jet service? Hard to tell, but not too far away.

 

[sengsational]

There's a lot of people who have ideas about AI, and are willing to share their perspectives, but not many people who have thought deeply about possible long-term effects. Here's a post I made that's applicable here:

I just got done reading "Nexus" (Harari) and it's a fascinating view of the progress of humanity, stone age to AI. It's a depressing look at what we're up against as AI becomes more integrated into the world. He argues pretty convincingly, from cultural movements of the past, that as AI gets integrated into more stuff, things will probably go south. He argues against what he calls "the naive view of information", which incorrectly posits more information will yield more truth. Plenty of justification that truth is not the obvious or likely result of more information.

He talks about how earlier technologies for managing information (the printing press, telegraph, radio, etc) were used and misused by the various cultural organization schemes on both sides of the iron curtain. Harari talks about a new "silicon curtain" or maybe not just one, where each grouping will have their own "mythology," or cultural narrative. Again, some pretty reasonable arguments about how AI will be different. The network, up until now, always required a flesh and blood person as a participant. If AI's become "members of the community", acting on their own, THAT is why "it's different this time," when compared to earlier human created technologies.

 

[OldShooter]

I guess it depends... ... Today, we can't open a thread about self driving cars because of the controversy. What am I saying? I'm saying that it has gone slower than many expected.

Agreed. Actually my point, well known in tech, is that the short term impact of a technical advance is usually much less than the hype projects, but it is also true that the long-term impact is often much greater than was expected. In addition to AI I think we will see that with quantum computing and (hopefully) with fusion power.

 

[disneysteve]

I am not a tech guy so I come at this as a regular person who does encounter AI in various places. So far, I've seen a lot of bad from AI. Bad "writing" from otherwise respectable news organizations. Awful item descriptions and photos on ebay. Inaccurate AI-generated summaries on Google and Facebook. Stuff like that. I'm sure AI is probably doing some good stuff, and maybe when it is, it goes unnoticed for just that reason, but the bad AI stands out like a sore thumb. Maybe some day it will be good, but that day isn't today.

 

[Walt34]

Sorry, Connecticut is the birthplace of powered aviation. Gustave Whitehead flew the No. 21 in Bridgeport, Connecticut on August 14, 1901.

Well, it depends on who one asks. Many hold that Alberto Santos-Dumont was first. Record-keeping then wasn't so good so there is still debate about who was first in controlled flight. Or even the definition of "controlled flight".

 

[rk911]

They could build a flying car in the 1950s, but I'm still waiting for mine.
So I've learned not to expect much.

View attachment 53808

Several have completed successful test flights but I can't see them taking off (pun absolutely intended ) until there is somewhere for them to land and then transition to ground transport to complete the journey. Flying cars might be most useful for short vacafion trips between airports but certainly not for commuting to most jobs.

 

[Koolau]

"Combined with a form of fusion, the machines had found all the energy they would ever need." AI has been well covered in fiction. I expect neo-Luddites (pun intended) will block AI from taking over everything.

In my best Travis Bickle/Robert De Niro voice: "You talkin' to me." (Resident neo-Luddite here.)

 

[GrayHare]

In my best Travis Bickle/Robert De Niro voice: "You talkin' to me." (Resident neo-Luddite here.)

I know several who worked in the AI field and have ceased doing so because of concern where it is headed. AI is a threat not only to the human way of life but also to species survival. I don't see that threat as imminent, but it seems inevitable with time unless humans actively stop it.

 

[Koolau]

Several have completed successful test flights but I can't see them taking off (pun absolutely intended ) until there is somewhere for them to land and then transition to ground transport to complete the journey. Flying cars might be most useful for short vacafion trips between airports but certainly not for commuting to most jobs.

Aviation is a whole order of magnitude more complicated than ground based transportation. Those who thought we'd have flying cars had not really thought it through - even though it was reasonably simple to design/build a vehicle capable of both air and ground travel.

Whether AI will run into similar "road blocks" (pun intended) remains to be seen and YMMV.

The elephant in the room (beside sengsational's contribution) is that most human inventions have been designed to be and have more or less remained man's servant. Could AI eventually be a de facto master? I'd swerve quickly into the political if I suggested why I think that way, so I'll just leave it at that because YMMV.