Scenario 1: People will just adjust like every other technological advancement period and do different jobs that don't exist yet.
Scenario 2: We switch to a deflationary system and the price of goods and services drop to 0 as well with labor.
1 - doubtful. I see something more akin to Hunger Games happening. Rich get richer and dominate the poor. Middle class becomes non-existent
2 - So, Star Trek?
@hurryinghoosiers DOWN GOES HICKORY. DOWN GOES HICKORY
Biden vs. Trump on immigration and border security, in 12 charts - The Washington Post
hardly. keep proving yourself to be an imbecile though mcmurt
My wife is a programmer and has been using AI to help out with code changes.
A summary of her experience with Cursor (AI code generator) and CodeBot (their code "checker"):
1) The simpler, more repetitive the task, the better the AI code. So if you want it to create code that would take a field and index it up in value by one notch and you need to do that for 1000 fields (that you normally would have to write a piece of code for every field), it is a fantastic time saver.
2) With that said, every scrap of code that the AI generates needs to be reviewed. The code it generates is still a time saver, but it usually needs to be tweaked. Part of coding is nuance and style. You can have two different paragraphs of code that do the same thing but are organized completely different and that varies company to company and probably is a result of the first programmer writing it 30 years before and everybody follows that style to remain consistent.
3) My wife has written code, had the Codebot check it, Codebot found an issue, recommended a change to fix it (which she implemented), ran the Codebot check again, and the Codebot came back saying that the section of code that it had previously just recommended is wrong and needs to be changed. The moral of the story here is that coders coming out of college now are still going to be needed, it's just that they are going to need to be fluent in understanding how AI can be used in conjunction with their job so that they can do the job as efficiently as possible.
4) What companies are going to find out quickly is that you can save a buck now by going with AI to replace the entry-level programmers, but 10 years from now when their mid-level experience programmers have become managers (or moved on), you are going to be left with a vacuum of nobody able to take over the programming duties (that AI still can't handle) because you didn't have entry-level programmers who have gained experience and can jump to the mid-level that is necessary (the ones who understand the whole picture of the code library and understand how to improve it with new features that customers want). AI is still not imaginative.
I understand what you're saying but it's like buggy drivers in the early 1900's. Sure, there's going to be overlap but ChatGPT and the other LLMs are only a few years old. They will improve and dramatically so. They will be come better than human programmers at at least at 90% of the work. And much like the buggy drivers being replaced by trucks, AI will become more efficient and cheaper.
We are going to live through a monumental change in how humans organize themselves (well, if I live another 30 ish years - fingers crossed). Nobody in human history will experience as much change as we will over the next 10-15 years. I say this looking at the investments being made by the largest tech comopanies in the world (and China). They aren't stupid and they don't chase dreams with that kind of money. It's real, AGI will be achieved. I just hope this race doesn't start a war.
mgmt too. Even c-suite.
once AGI is achieved the changes seen in the Indistrial Revolution that took two or three generations to solve will occur in a decade.
Nah, someone has to make the decisions and take the fall. There will always be room for MGMT, it just doesn't need plebs to produce reporting.
@larsiu I’m out over my skis here bc I have zero tech knowledge but I am an expert in reading people. I think these tech guys pushing this narrative might be disingenuous. Pretending we are on the brink of these changes is self-serving as it helps procure funding, contracts, and boosts valuations.
Personally this is of no moment to me as for once I’m perfectly situated. That will coincide with the age of my retirement and I’m an owner of multiple businesses. My daughter, with God awful grades, will be likely coming out of her first divorce, and since she’s a cute kid I trust she’ll be a cute adult so should do well in the marriage game and attendant dissolution agreement, which she can couple with UBI until she lures hubby number 2
My wife is a programmer and has been using AI to help out with code changes.
A summary of her experience with Cursor (AI code generator) and CodeBot (their code "checker"):
1) The simpler, more repetitive the task, the better the AI code. So if you want it to create code that would take a field and index it up in value by one notch and you need to do that for 1000 fields (that you normally would have to write a piece of code for every field), it is a fantastic time saver.
2) With that said, every scrap of code that the AI generates needs to be reviewed. The code it generates is still a time saver, but it usually needs to be tweaked. Part of coding is nuance and style. You can have two different paragraphs of code that do the same thing but are organized completely different and that varies company to company and probably is a result of the first programmer writing it 30 years before and everybody follows that style to remain consistent.
3) My wife has written code, had the Codebot check it, Codebot found an issue, recommended a change to fix it (which she implemented), ran the Codebot check again, and the Codebot came back saying that the section of code that it had previously just recommended is wrong and needs to be changed. The moral of the story here is that coders coming out of college now are still going to be needed, it's just that they are going to need to be fluent in understanding how AI can be used in conjunction with their job so that they can do the job as efficiently as possible.
4) What companies are going to find out quickly is that you can save a buck now by going with AI to replace the entry-level programmers, but 10 years from now when their mid-level experience programmers have become managers (or moved on), you are going to be left with a vacuum of nobody able to take over the programming duties (that AI still can't handle) because you didn't have entry-level programmers who have gained experience and can jump to the mid-level that is necessary (the ones who understand the whole picture of the code library and understand how to improve it with new features that customers want). AI is still not imaginative.
And that's for coding, which has standards and basis. The garbage I get from these LLMs when it comes to trying to do a novel or unique task is wild.
I was trying to create a sports card collection database for myself. I uploaded thousands of card images over time and both Gemini and ChatGPT started to make up additional cards to add to the collection. These things have some benefit, but after spending hours on end, I was able to do it myself in a similar amount of time with more accuracy because of the errors.
@hurryinghoosiers No content Hickory. As someone else said just stop posting. If you can’t address the links cited you’re only here for turbulence
@jdb They were doing that with cases. Lawyers were using it and AI was making up cases that didn’t exist.
@jdb They were doing that with cases. Lawyers were using it and AI was making up cases that didn’t exist.
That was always a flaw I saw in Asimov's 3 Laws. It is based upon the concept that computers can't lie.
That is absolutely not the case (or at least not the way that think of what constitutes lying).
@jdb They were doing that with cases. Lawyers were using it and AI was making up cases that didn’t exist.
it's 3-4 years old. a toddler. AGI is the hinge. The investments being made by smart people aren't a dick measuring contest. The service industry today that exists in telephony, email, and those attendant systems will disappear. That is an enormous swath of the country. Then more and more white collar jobs that today require college degrees will be consumed.
Most of low to mid level white collar mgmt today is work distribution, adminisration, and QA. Buh bye. Oh, but those super smart people that find efficiencies, surely they'll be around forever right? Uh, no.
I'll stand by it. Largest disruption to human society since the wheel.
Something to also bear in mind is that there is going to be an upper limit to AI capability soon based upon power generation.
You remember how everybody had their panties in a wad about how EV's were never going to be able to replace the ICE cars because the power grid couldn't handle the additional load?
For reference, the power needed to support all the EV's on the road today is the same as what 10 AI data centers require. There are currently 4000 AI data centers in this country and there are plans for doubling that amount in the next 5-10 years.
It is coming to the point where you will need the complete output from a single nuclear power plant to operate the larger AI data centers (1 gigawatt).
AI is not going to get to the level that larsIU is talking about until we figure out a way to get a massive boost to our power grid. We're talking 20 to 25 nuclear reactors in every state. That is not happening in the next 30 years.
@squeakyclean And the public won’t go for that spike on their utility bills - essentially subsidizing them
@jdb They were doing that with cases. Lawyers were using it and AI was making up cases that didn’t exist.
The investments being made by smart people aren't a dick measuring contest.
Sure they are. Why do you think they are all competing against each other instead of working with each other? Musk is constantly bragging about this, Altman about that.
@jdb They were doing that with cases. Lawyers were using it and AI was making up cases that didn’t exist.
it's 3-4 years old. a toddler. AGI is the hinge. The investments being made by smart people aren't a dick measuring contest. The service industry today that exists in telephony, email, and those attendant systems will disappear. That is an enormous swath of the country. Then more and more white collar jobs that today require college degrees will be consumed.
Most of low to mid level white collar mgmt today is work distribution, adminisration, and QA. Buh bye. Oh, but those super smart people that find efficiencies, surely they'll be around forever right? Uh, no.
I'll stand by it. Largest disruption to human society since the wheel.
Mid-level managers aren't real managers. The C-Suite has to remain intact so when shit goes wrong, the Board and shareholders have people to blame. I'm sure there will be more and more suits lobbed against AI creators (Open, Google, Musk), but I find it hard that courts will rule they will be liable beyond a limited amount.
I don't disagree with your premise that it will be disruptive, but I am skeptical of how far they can go, even with more years under their belts. I'm sure there will be other generations that I cannot fathom yet that will be the real killers.
Something to also bear in mind is that there is going to be an upper limit to AI capability soon based upon power generation.
You remember how everybody had their panties in a wad about how EV's were never going to be able to replace the ICE cars because the power grid couldn't handle the additional load?
For reference, the power needed to support all the EV's on the road today is the same as what 10 AI data centers require. There are currently 4000 AI data centers in this country and there are plans for doubling that amount in the next 5-10 years.
It is coming to the point where you will need the complete output from a single nuclear power plant to operate the larger AI data centers (1 gigawatt).
AI is not going to get to the level that larsIU is talking about until we figure out a way to get a massive boost to our power grid. We're talking 20 to 25 nuclear reactors in every state. That is not happening in the next 30 years.
That problem is already being worked on though. I nearly invested in a nuclear power generation company that puts minireactors in/alongside Data Centers to alleviate the grid problem.