From what I’m seeing it seems possible that by 2024-25 there will be quite a lot of unemployment (at least in some countries) due to AIs being able to do more and more jobs or at least massively increasing productivity to the point companies using AI would have a massive advantage, getting rid of the companies still hiring large ammounts of people.
Also, as robots would take longer to appear this unemployment would start with office and WFH workers cascading to less demand for jobs that would support these people and thus to a lot of competition for jobs and ever smaller salaries compounding the situation and leading to massive social changes (with hopefully the people realizing how good an idea taking the means of production is).
Maybe not 2024-2025. Large language models and bots that copy art from the internet are cool and all but they’re not yet capable of matching humans and probably won’t for quite a while.
In some cases they can match and even pass humans, as there have been some research showing it, and even in the cases where it can’t yet do so, having workers using AI could still increase productivity to the point that humans aren’t as necessary for completing the jobs. And if the big corporations don’t start using it startups are likely to arrive to do such jobs, like callcenters which already have AI systems that can do it to some extent and might be fully automated by the time the infrastructure for using AIs in such areas appears.
No, but the AIs that already exist that can use calculators and other tools to solve problems are AI. They might not yet be AGI yet but as new systems using LLMs and other techniques develop it shouldn’t take long until they can identify a problem, calculate a solution and act on it by themselves.
But they aren’t identifying problems, they’re algorithms. These aren’t true “AIs” and never will be. They’re just designed to give a response based on input, like a calculator. They might get better at giving that response, but the way the current tech is headed they won’t ever actually create an “AI” just a mechanical Turk designed to look like it passes the Turing test.
Here is a simple example of what AI can already do. It might not be possible to do it all yet in a sinlge system yet, but with current tech as it gets deploy the following should be possible.
*First AI system receives a message from someone that there is a job to be done somewhere. By using it’s “intelect” and a program it sends a robot to solve the problem.
*The robot is an AI system that has AIs that identify the enviroment around the robot, another that works with programs to creat a route, another that actually drives the robot with human like movements on which it was trained.
*The robot finds a path until the problem and then it visually identifies what is what, a 3D model of the enviroment is made and an AI identifies a solution all based on the videos it was trained and the things that all robots have experienced.
*Then the AI that moves the robot interacts with the enviroment in the way that the was deemed to solve the problem and the AIs that were looking at the problem make sure it is fixed.
All these things can already be done by LLMs/AIs to varying degrees and looking at the progress from this year alone it seems likely that all of these things will be doable soon (this decade) and then it’s just a matter of producing more AIs/robots.
But that isn’t AI. That’s my point. Is it automation? Yes. Absolutely. But it isn’t thinking for itself, it is following instructions. It has no ability to deal with problems outside of the areas it is “trained” in. It can’t invent solutions to problems, only follow the instructions for solving a problem it has been given. It doesn’t “think” any more than any other program does.
I’m writing a visual novel game right now. In this game, the characters interact with the player, they respond to questions, have their own preferences and make their own decisions. But no one would think a video game character is actually thinking for themselves. They’re just lines of code made to emulate a conversation with a person.
And you should be very cautious about accepting any claims of “this amazing sci-fi tech will be here within the decade” because the newspapers have been making those claims for as long as we’ve had sci-fi. Someone else in this thread mentioned AI being hyped up for the past 60 years, and that entire time it’s been “just around the corner.” For 60 years. Don’t fall for marketing hype and buzzwords. Are robotics getting more advanced? Yes, they are. Automation is increasing as well. These are not the same thing as actual artificial intelligence though.
But it isn’t thinking for itself, it is following instructions.
And humans don’t follow instructions??
Our body and our enviroment are constantly prompting us to do things and we act on it. We need to eat so we will find a way to keep that need satisfied. We want to have fun and will try to find something fun. If it’s too hot or cold we will try to deal with it.
It doesn’t look diferent from a robot in a factory, for exemple, that always wants to do a job, finds a job by investigating its enviroment with photo recognition AI system or by being asked by someone or something to do a job, and then it finds its way to the job and tries to think of a solution to solve it based on how it has done other jobs before and if it doesn’t work it tries to come up with new solotions just like a human would.
They may be able to match and pass humans on specific tasks, but most actual jobs require an element of being able to balance workloads, solve problems and think analytically, humans will still have to play a role for a long time, though we will gradually grow more and more dependent on AI which may(probably will) lead to a few layoffs
From what I’m seeing it seems possible that by 2024-25 there will be quite a lot of unemployment (at least in some countries) due to AIs being able to do more and more jobs or at least massively increasing productivity to the point companies using AI would have a massive advantage, getting rid of the companies still hiring large ammounts of people.
Also, as robots would take longer to appear this unemployment would start with office and WFH workers cascading to less demand for jobs that would support these people and thus to a lot of competition for jobs and ever smaller salaries compounding the situation and leading to massive social changes (with hopefully the people realizing how good an idea taking the means of production is).
Maybe not 2024-2025. Large language models and bots that copy art from the internet are cool and all but they’re not yet capable of matching humans and probably won’t for quite a while.
In some cases they can match and even pass humans, as there have been some research showing it, and even in the cases where it can’t yet do so, having workers using AI could still increase productivity to the point that humans aren’t as necessary for completing the jobs. And if the big corporations don’t start using it startups are likely to arrive to do such jobs, like callcenters which already have AI systems that can do it to some extent and might be fully automated by the time the infrastructure for using AIs in such areas appears.
A calculator can perform mathematical tasks much faster than a human, does this mean a calculator is an AI?
No, but the AIs that already exist that can use calculators and other tools to solve problems are AI. They might not yet be AGI yet but as new systems using LLMs and other techniques develop it shouldn’t take long until they can identify a problem, calculate a solution and act on it by themselves.
But they aren’t identifying problems, they’re algorithms. These aren’t true “AIs” and never will be. They’re just designed to give a response based on input, like a calculator. They might get better at giving that response, but the way the current tech is headed they won’t ever actually create an “AI” just a mechanical Turk designed to look like it passes the Turing test.
Here is a simple example of what AI can already do. It might not be possible to do it all yet in a sinlge system yet, but with current tech as it gets deploy the following should be possible.
*First AI system receives a message from someone that there is a job to be done somewhere. By using it’s “intelect” and a program it sends a robot to solve the problem.
*The robot is an AI system that has AIs that identify the enviroment around the robot, another that works with programs to creat a route, another that actually drives the robot with human like movements on which it was trained.
*The robot finds a path until the problem and then it visually identifies what is what, a 3D model of the enviroment is made and an AI identifies a solution all based on the videos it was trained and the things that all robots have experienced.
*Then the AI that moves the robot interacts with the enviroment in the way that the was deemed to solve the problem and the AIs that were looking at the problem make sure it is fixed.
All these things can already be done by LLMs/AIs to varying degrees and looking at the progress from this year alone it seems likely that all of these things will be doable soon (this decade) and then it’s just a matter of producing more AIs/robots.
But that isn’t AI. That’s my point. Is it automation? Yes. Absolutely. But it isn’t thinking for itself, it is following instructions. It has no ability to deal with problems outside of the areas it is “trained” in. It can’t invent solutions to problems, only follow the instructions for solving a problem it has been given. It doesn’t “think” any more than any other program does.
I’m writing a visual novel game right now. In this game, the characters interact with the player, they respond to questions, have their own preferences and make their own decisions. But no one would think a video game character is actually thinking for themselves. They’re just lines of code made to emulate a conversation with a person.
And you should be very cautious about accepting any claims of “this amazing sci-fi tech will be here within the decade” because the newspapers have been making those claims for as long as we’ve had sci-fi. Someone else in this thread mentioned AI being hyped up for the past 60 years, and that entire time it’s been “just around the corner.” For 60 years. Don’t fall for marketing hype and buzzwords. Are robotics getting more advanced? Yes, they are. Automation is increasing as well. These are not the same thing as actual artificial intelligence though.
And humans don’t follow instructions??
Our body and our enviroment are constantly prompting us to do things and we act on it. We need to eat so we will find a way to keep that need satisfied. We want to have fun and will try to find something fun. If it’s too hot or cold we will try to deal with it.
It doesn’t look diferent from a robot in a factory, for exemple, that always wants to do a job, finds a job by investigating its enviroment with photo recognition AI system or by being asked by someone or something to do a job, and then it finds its way to the job and tries to think of a solution to solve it based on how it has done other jobs before and if it doesn’t work it tries to come up with new solotions just like a human would.
They may be able to match and pass humans on specific tasks, but most actual jobs require an element of being able to balance workloads, solve problems and think analytically, humans will still have to play a role for a long time, though we will gradually grow more and more dependent on AI which may(probably will) lead to a few layoffs