AI and the Economy

[Pexbo]

I’d like to keep this thread hypothetical although I appreciate that’s virtually impossible and the usual AI skeptics will be in here to tell us that ChatGPT gets things wrong and AI has no real world applications - continue use the other AI thread for that.

This is more philosophical and ideally is discussed on this premise:

Sam Altman is rumoured to be briefing Congress on an AI breakthrough where they have achieved a level of AGI, with a PHD level super agent. In short, it’s an AI which can replace a skilled technical human to do complex tasks.

Mark Zuckerberg believes they will be replacing all mid-level software engineers with AI starting this year.

The likelihood is that this isn’t going to be limited to Software development and will could be a disruptor like we have not seen since the Industrial Revolution. Lawyers, Architects, Chemists, Data Scientists, Media editors, content creators the list goes on and on.


So if middle to upper earners are suddenly made redundant in their droves, what happens to the economy? Who is going have the money to actually buy the products AI is making?

 

[Massive Spanner]

I’d like to keep this thread hypothetical although I appreciate that’s virtually impossible and the usual AI skeptics will be in here to tell us that ChatGPT gets things wrong and AI has no real world applications - continue use the other AI thread for that.

This is more philosophical and ideally is discussed on this premise:

Sam Altman is rumoured to be briefing Congress on an AI breakthrough where they have achieved a level of AGI, with a PHD level super agent. In short, it’s an AI which can replace a skilled technical human to do complex tasks.

Mark Zuckerberg believes they will be replacing all mid-level software engineers with AI starting this year.

The likelihood is that this isn’t going to be limited to Software development and will could be a disruptor like we have not seen since the Industrial Revolution. Lawyers, Architects, Chemists, Data Scientists, Media editors, content creators the list goes on and on.


So if middle to upper earners are suddenly made redundant in their droves, what happens to the economy? Who is going have the money to actually buy the products AI is making?

I don’t think that’s what Zuckerberg actually said (replacing people) but I guess we can all draw our own conclusions.

I think a lot of the big tech companies are already preparing for it, you can see it by the fact that they never hire grads or do internships anymore. I honestly think our generation could be the last to have long term engineering careers (and that’s being optimistic) and in a few years experienced people like me probably won’t even be writing much code, just planning and working with customers and such, almost like PMs.

Tech is in a weird bubble right now and it’s definitely going to burst. The salaries are obscene and the companies are going to do everything they can to stop it.

Ai writing code that’s reviewed by other Ai could truly be end times though, I do think it’s incredibly dangerous path to go down, but if it kills Meta and that cnut at least that’s a plus.

 

[Revan]

I don’t think that’s what Zuckerberg actually said (replacing people) but I guess we can all draw our own conclusions.

I think a lot of the big tech companies are already preparing for it, you can see it by the fact that they never hire grads or do internships anymore. I honestly think our generation could be the last to have long term engineering careers (and that’s being optimistic) and in a few years experienced people like me probably won’t even be writing much code, just planning and working with customers and such, almost like PMs.

Tech is in a weird bubble right now and it’s definitely going to burst. The salaries are obscene and the companies are going to do everything they can to stop it.

Ai writing code that’s reviewed by other Ai could truly be end times though, I do think it’s incredibly dangerous path to go down, but if it kills Meta and that cnut at least that’s a plus.

Companies are hiring interns though. Right now, I have 6 interns, and one starting in a couple of months. I know lots of FAANG companies who are getting interns. But I agree that the numbers are lower than a few years ago. And the entry barrier is getting higher.

I think your post is very logical but cannot help but think that the last part to really make AI is going to be a bit harder than what people are thinking. Or at least, that is the hopium I am selling to myself.

Saying that, I mostly agree with your general point but think that the timeline will be a bit longer. Similarity, I wouldn’t advice people to study computer science right now, albeit, I hope there will be computer science for a long time, but maybe quite a bit different.

Ultimately, I think there are going to be lots of unknown unknowns going ahead that are going to affect the economy in unimaginable ways. And that, assuming that a super intelligence is not going to just kill us in the first place. But again, not this decade.

 

[Telsim]

Good timing on this thread, considering the Chinese just obliterated SV and put significant doubt on the valuation of certain companies and their demands for monstrous investment. Nvidia already sliding down and insiders report it's a pretty panicky environment within OpenAI and the rest. Needless to say what could happen if investors lose faith and the thing comes crashing down. It will be interesting to see what the response will be, unless they just shut down chip shipments to China, which will at least put an end to the "free market" narrative.

 

[Red Star One]

Good timing on this thread, considering the Chinese just obliterated SV and put significant doubt on the valuation of certain companies and their demands for monstrous investment. Nvidia already sliding down and insiders report it's a pretty panicky environment within OpenAI and the rest. Needless to say what could happen if investors lose faith and the thing comes crashing down. It will be interesting to see what the response will be, unless they just shut down chip shipments to China, which will at least put an end to the "free market" narrative.

Yeah great timing indeed, reading the news from China and seeing it’s impact on the stock market, I’d guess it will be Sam Altman that gets replaced first, not the mid-upper level lawyers and architects.

 

[Superden]

and the UK govt has just turned down furthering the pilot studies on using AI in some govt functions. Theres a surprise, the idea that AI is a magic pill to solve everything is just madness, especially when on one side you have luddites in the blob and on the other snake oil salesman promising the earth to make lots of quick bob.

from the guardian today
>>>Ministers have shut down or dropped at least half a dozen artificial intelligence prototypes intended for the welfare system, the Guardian has learned, in a sign of the headwinds facing Keir Starmer’s effort to increase government efficiency.
Pilots of AI technology to enhance staff training, improve the service in jobcentres, speed up disability benefit payments and modernise communication systems are not being taken forward, freedom of information (FoI) requests reveal.
Officials have internally admitted that ensuring AI systems are “scalable, reliable [and] thoroughly tested” are key challenges and say there have been many “frustrations and false starts” <<<

 

[neverdie]

Saying that, I mostly agree with your general point but think that the timeline will be a bit longer. Similarity, I wouldn’t advice people to study computer science right now, albeit, I hope there will be computer science for a long time, but maybe quite a bit different.

The irony is that the STEM stuff is that gets hit the hardest because its logic and math. Ai can do that without a problem. I say this because the old attitude about Arts degrees. Other than Doctors, Teachers (arts), and some other sectors, AI will have a massive impact (they'll use AI with doctors but won't be replacing them imo).

 

[Simbo]

Is this like when they hyped up the Tesla share price to absurd levels through pure rumours and stories?

 

[neverdie]

Is this like when they hyped up the Tesla share price to absurd levels through pure rumours and stories?

No, from what I've read, and I'm not in the industry or anything, it is inevitable. For AI to do advanced typed coding, however long it takes, "by itself" (and checked by other AIs), this will eliminate a lot of jobs.

 

[Laurencio]

I'm generally sceptical about "break throughs" that are supposedly miles ahead of companies in Silicon Valley.

AI is part bubble, part reality, the bubble part will burst at some point and maybe Deep Seek is the trigger for that to happen. Long term NVIDIA should be fine.

 

[Revan]

No, from what I've read, and I'm not in the industry or anything, it is inevitable. For AI to do advanced typed coding, however long it takes, "by itself" (and checked by other AIs), this will eliminate a lot of jobs.

If the AI will be able to do advanced coding so it replaces good software engineers, I think by default it would be able to replace lots of other office jobs.

If coders are replaced, then from high paying jobs, I guess only doctors are safe.

 

[neverdie]

I think by default it would be able to replace lots of other office jobs.

Yep, it's just there will be massive conversations around areas like chemists and teachers where people are not comfortable with having their kids being taught be an AI or their meds decided that way either.

But it cuts a lot of jobs, you're right.

 

[Beachryan]

Can only really speak to my experience - though it should be reasonably relevant: the actual AI applications are being chased, rather than surfacing organically. Every group CEO has to be out there telling the market they're all in on AI or else their shareprice takes a tumble. Then that trickles down to business areas, and we get things like an annual objective to come up with ways to use AI to 'improve' our workflows, when really it's just about saying we use AI, regardless of suitability.

To date, the LLMs have been impressive, and there are definitely actual, practical applications of them one can do right now, but I honestly cannot see a world where they improve to a point where they'd be replacing that many positions. I think they make me a better employee (particularly helping with my rusty coding) but there is still creativity in more white collar roles than we'd imagine, as well as accountability concerns.

So some jobs - paralegals, lower level accountants, document-review stuff - those could well go the way of the translator. But I don't think it'll replace vast swathes of industry. Personally think it's all waaaaay overblown at present.

 

[altodevil]

I for one am glad that this once in a generation opportunity to drastically improve work/life balance and living conditions has arrived with Trump/Musk at the helm

 

[langster]

I think people using AI to make wannabe gangster movies is the way forward and will definitely drive the economy forward at a level not seen before.

Ya get me bruv?

 

[adexkola]

There's a lot of snake oil branded as AI in the marketplace.

 

[Revan]

There's a lot of snake oil branded as AI in the marketplace.

I have a friend who knows nothing about AI, who made an AI company last year and now is close to closing 5-7M in funding. The AI there is literally an API call to OpenAI/Claude.

When I asked him why he was making an AI company, he said ‘because AI now is a buzz word’. But kudos to him, for a lot of people this is a goldmine opportunity.

 

[4bars]

Can only really speak to my experience - though it should be reasonably relevant: the actual AI applications are being chased, rather than surfacing organically. Every group CEO has to be out there telling the market they're all in on AI or else their shareprice takes a tumble. Then that trickles down to business areas, and we get things like an annual objective to come up with ways to use AI to 'improve' our workflows, when really it's just about saying we use AI, regardless of suitability.

To date, the LLMs have been impressive, and there are definitely actual, practical applications of them one can do right now, but I honestly cannot see a world where they improve to a point where they'd be replacing that many positions. I think they make me a better employee (particularly helping with my rusty coding) but there is still creativity in more white collar roles than we'd imagine, as well as accountability concerns.

So some jobs - paralegals, lower level accountants, document-review stuff - those could well go the way of the translator. But I don't think it'll replace vast swathes of industry. Personally think it's all waaaaay overblown at present.

I Know nothing about AI and the only thing I hear is that will kill a lot of jobs. Isn't more feasible that instead of replacing a lot of jobs, AI will be a tool for the existing jobs and boost its productivity and that a coder (as an example) would be able to do what 2 people does thanks to the AI assistance?

Wouldn't that kill a lot of jobs? or the AI helping productivity is way far away to assist to convert 2 jobs to 1 or 3 jobs to 1?

 

[Ultimate Grib]

AI is the new dotcom bubble. It will burst and cause a wave of economic destruction but it will change the world.

 

[Mr Pigeon]

As long as this stops companies from sprinkling AI on everything, like me sprinkling MSG into every meal I make, then I'll be happy. Hopefully it also stops my heart palpitations and constant thirst.

 

[yumtum]

I can't see how we can become efficient enough at storing and accessing data for AI (automation) to replace every top job.

They're starting to build data centres next to swimming pools to heat them up! It uses 500ml of water for every 5-50 Chat-GPT searches and 90% of the world's data was created between 2021-2024.

Earth overshoot day was on August 1st in 2024, the earliest that day has been was July 29th 2019 (it was down to 20th of August during covid), we'll likely soar past that soon enough, we're pillaging the earth and storing and accessing massive amounts of data is going to extrapolate this massively

It's not sustainable.

 

[4bars]

I can't see how we can become efficient enough at storing and accessing data for AI (automation) to replace every top job.

They're starting to build data centres next to swimming pools to heat them up! It uses 500ml of water for every 5-50 Chat-GPT searches and 90% of the world's data was created between 2021-2024.

Earth overshoot day was on August 1st in 2024, the earliest that day has been was July 29th 2019 (it was down to 20th of August during covid), we'll likely soar past that soon enough, we're pillaging the earth and storing and accessing massive amounts of data is going to extrapolate this massively

It's not sustainable.

In 2 years AI will tea h us go to built a warp drive and create an atmosphere in mars so they can send us there while billiomaires stays on earth

 

[adexkola]

I have a friend who knows nothing about AI, who made an AI company last year and now is close to closing 5-7M in funding. The AI there is literally an API call to OpenAI/Claude.

When I asked him why he was making an AI company, he said ‘because AI now is a buzz word’. But kudos to him, for a lot of people this is a goldmine opportunity.

I don't think I'd be able to live with pulling that off

 

[Reapersoul20]

I am training interns at the moment.

AI is creating a generation of illiterate programmers and coding jobs are still pretty safe IMO. People underestimate how much of the world is ran on computers now. For every job being "replaced" by AI, more will open up.

AI has been, and will continue to be disruptive, but it'll just change priorities - not sack all the mid-level engineers

 

[Ultimate Grib]

@Revan @nickm @Balljy Looks like China isn't as far behind as you all thought in the AI race.

 

[Scandi Red]

When I asked him why he was making an AI company, he said ‘because AI now is a buzz word’. But kudos to him, for a lot of people this is a goldmine opportunity.

"The traysh AI business is a gooooldmine!"

I think IASIP is due an AI episode where Dennis loses his mind.

 

[Kaos]

I am training interns at the moment.

AI is creating a generation of illiterate programmers and coding jobs are still pretty safe IMO. People underestimate how much of the world is ran on computers now. For every job being "replaced" by AI, more will open up.

AI has been, and will continue to be disruptive, but it'll just change priorities - not sack all the mid-level engineers

I agree with this. I see this with junior developers (or whats left of them), most of them heavily leverage libraries and frameworks and understand very little on DSA principles.

 

[Pexbo]

I agree with this. I see this with junior developers (or whats left of them), most of them heavily leverage libraries and frameworks and understand very little on DSA principles.

Coming from a ComSci degree background I had the opposite issue as a junior developer. I’d try to build everything from scratch.

I soon learnt that the wheel has already been invented and finding the right libraries and frameworks to leverage is generally the correct approach. There’s a lot to be said for the critical analysis of the options and actually understanding why the correct one is the correct one though.

 

[Edgar Allan Pillow]

The response from Google AI, which does seem a balanced outlook.

What this and the Tech company leadership glosses over are the short-term adverse impact. Short Team can be a generation long (30-50 years) and the impact in practical terms would be more significant than what these companies predict (or are willing to disclose).

AI Overview

Artificial intelligence (AI) can both create new jobs and make some traditional roles obsolete. The impact of AI on unemployment depends on the time scale, with short-term effects being negative and long-term effects being positive.

Short-term impact

• Automation: AI can automate routine tasks, which can lead to short-term job losses
• Clerical support: Clerical support workers, such as receptionists, secretaries, and customer service workers, may be particularly affected by AI

Medium-term impact

• Industry restructuring
  As industries restructure, the positive and negative effects of AI on employment balance out
  
  
• New jobs
  New job roles may emerge as AI frees up employees to focus on more complex work

Long-term impact

• Job creation
  AI can create new jobs, such as social media managers and cybersecurity specialists
  More efficient work environment
  AI can make businesses more efficient and productive, which can lead to job creation and expansion

Overall impact

• Job market transformation: AI can transform the job market, impacting workers of all skill levels
  More fulfilling work: AI can free up employees to focus on more complex work, which can lead to a more fulfilling work experience

 

[buchansleftleg]

I Know nothing about AI and the only thing I hear is that will kill a lot of jobs. Isn't more feasible that instead of replacing a lot of jobs, AI will be a tool for the existing jobs and boost its productivity and that a coder (as an example) would be able to do what 2 people does thanks to the AI assistance?

Wouldn't that kill a lot of jobs? or the AI helping productivity is way far away to assist to convert 2 jobs to 1 or 3 jobs to 1?

You're right currently AI is good at doing tasks that are a pain for Humans to do and it can do them quicker than a Human can. So it is finding it's way into thinks like summarising the content of a call / conversation or a selection of product reviews. It can do that quite well as long as you don't want it in absolute real time. It's generally too slow / inaccurate for you to depend on it and remove the real person.

So it will shave a bit of time off the amount of time it takes someone to write up a conversation, but you still need the person to provide the oversight in most cases. So in conversations that don't matter so much or those that get very busy...like contacting a retail store it will free up more staff to take calls because the AI summarises that "4bars rang to amend trainer order from green to red Adidas Mondial size 9" quicker than the agent could type it.

However if you rang the doctor looking to raise an allergic reaction to some medicine an AI summary of that conversation would still need a lot of checking, even if accuracy levels got as high as 95% so the "time saved" by auto-summary it is reduced because every detail is potentially important.

Real-time replacement of the person answering your call by AI would take a huge amount of investment and is currently much too slow because it has to be carefully authored to not start giving out random information/ be encouraged to break it's programming rules. it's responses will tend to be way too vague and be like a horoscope rather than providing the sort of factual assistance a person might be looking for.

So yes, as you say, it will be a productivity boost in certain areas but expect a whole load of bullshit projects trying to add in AI aspects to gain traction, funding or publicity.

 

[4bars]

You're right currently AI is good at doing tasks that are a pain for Humans to do and it can do them quicker than a Human can. So it is finding it's way into thinks like summarising the content of a call / conversation or a selection of product reviews. It can do that quite well as long as you don't want it in absolute real time. It's generally too slow / inaccurate for you to depend on it and remove the real person.

So it will shave a bit of time off the amount of time it takes someone to write up a conversation, but you still need the person to provide the oversight in most cases. So in conversations that don't matter so much or those that get very busy...like contacting a retail store it will free up more staff to take calls because the AI summarises that "4bars rang to amend trainer order from green to red Adidas Mondial size 9" quicker than the agent could type it.

However if you rang the doctor looking to raise an allergic reaction to some medicine an AI summary of that conversation would still need a lot of checking, even if accuracy levels got as high as 95% so the "time saved" by auto-summary it is reduced because every detail is potentially important.

Real-time replacement of the person answering your call by AI would take a huge amount of investment and is currently much too slow because it has to be carefully authored to not start giving out random information/ be encouraged to break it's programming rules. it's responses will tend to be way too vague and be like a horoscope rather than providing the sort of factual assistance a person might be looking for.

So yes, as you say, it will be a productivity boost in certain areas but expect a whole load of bullshit projects trying to add in AI aspects to gain traction, funding or publicity.

Thanks for the answer. It makes sense on what I though. I use AI for work and some times mathematical problems are obviously off. Like if I have a 1 meter diameter pipe and I want to fit 4 equally sized pipes inside the 1 meter diameter pipe, what is the maximum diameter of the 4 pipes (And I put parameters of not interjecting between them or the main pipe and blah blah blah). And it answers 0.5 meters each which at mathematical naked eye is impossible as the radius is 0.5. I asked again telling that the prior answer is incorrect and should be less than 0.5m and it agrees and after some other calculations it reach the conclusion of 0.707m which is ever dumber

If chatGPT can't calculate this, after many trials, parameters and conclusions and agreeing that should be smaller than 0.5m and recalculating to 0.707m, I don't know how useful can be in other areas where the naked eye can't even see if it is right or wrong. You can't trust it yet

 

[Revan]

@Revan @nickm @Balljy Looks like China isn't as far behind as you all thought in the AI race.

In software, of course not. There is nothing mystical about ChatGPT et al., and every company with good resources should be able to reach similar LLMs.

On hardware, China is nowhere close to TSMC when it comes to chips or Nvidia when it comes to AI GPUs. They will reach there but not for another half a decade though.

 

[afrocentricity]

 

[jderbyshire]

I worry for my brother in law - he has spent the last five years doing a coding course (python, java, etc)

He's in the final year of his course. Will there be career opportunities for him? Or will it be replaced with AI?

 

[tenpoless]

I worry for my brother in law - he has spent the last five years doing a coding course (python, java, etc)

He's in the final year of his course. Will there be career opportunities for him? Or will it be replaced with AI?

It's much harder to get an entry level job nowadays as a software engineer. Mainly because of the abundance of talents while the demand has been getting lower after the startup booms 5-6 years ago - investors are more conservative now and not everyone can get into top tech companies. It used to be that if you're half decent chances are there's a startup somewhere that you can join and they need a lot of applications to be built.

So the biggest competitors aren't even AI... it's another people including those from other countries (in some parts of asia they're cheaper and skilled enough to do the job). However I'd argue AI can't replace a good software engineer, at least not yet. And a good software engineer uses AI these days. So for aspiring SEs, they shouldn't see AI as a competition but as a helper for them to get better, to grow, and to find solutions faster. What took you 3 days of Googling can be done in 30 mins with GPT 4. Whatever happens in the future, be the person who works with AI to your own advantage than the other way around. I've been coding for almost 10 years now and started from last year I've been using AI. As a result my yearly performances had been getting better. I completed projects faster and that's what business wants and happy to pay you for - AI or no AI.

Do I worry i will get replaced? nope. Nobody knows for sure if it can reach the point of not even a very skilled SE is needed to build proper and secure apps. So I just take whatever advantage I can get now than worrying about the future. Making money while AIs do most of the tedious/repetitive codes for me. I can live with that.

 

[Zehner]

It's much harder to get an entry level job nowadays as a software engineer. Mainly because of the abundance of talents while the demand has been getting lower after the startup booms 5-6 years ago - investors are more conservative now and not everyone can get into top tech companies. It used to be that if you're half decent chances are there's a startup somewhere that you can join and they need a lot of applications to be built.

So the biggest competitors aren't even AI... it's another people including those from other countries (in some parts of asia they're cheaper and skilled enough to do the job). However I'd argue AI can't replace a good software engineer, at least not yet. And a good software engineer uses AI these days. So for aspiring SEs, they shouldn't see AI as a competition but as a helper for them to get better, to grow, and to find solutions faster. What took you 3 days of Googling can be done in 30 mins with GPT 4. Whatever happens in the future, be the person who works with AI to your own advantage than the other way around. I've been coding for almost 10 years now and started from last year I've been using AI. As a result my yearly performances had been getting better. I completed projects faster and that's what business wants and happy to pay you for - AI or no AI.

Do I worry i will get replaced? nope. Nobody knows for sure if it can reach the point of not even a very skilled SE is needed to build proper and secure apps. So I just take whatever advantage I can get now than worrying about the future. Making money while AIs do most of the tedious/repetitive codes for me. I can live with that.

And then there's also the fact that even if the AI is capable of writing the code, prompting is almost an art form. You need to be very precise with the formulation of your requirements which again requires to think them through. I don't know many product owners who are capable of this if any. Especially since many details only become apparent when you are actually implementing their ideas and many issues are solved by thoughtful engineers without them even knowing. Understanding how code works helps you enormously with prompting. As you said, in the midterm it will primarily increase the productivity of coders as it minimizes the time you have to spend on research.

 

[Gambit]

AI is decimating the creative industry sector. I'm currently adjusting myself and my work to incorporate it more and more. If you have any young people wanting to get into creative industries I'd strongly advise against it.

 

[Massive Spanner]

I know very strong, experienced software engineers who were let go and can't even get an interview right now, the industry is fecked. Mad that just a few years ago companies would have been throwing everything they could to get people like them. Feels like there's no industry out there as volatile as tech.

 

[tenpoless]

I know very strong, experienced software engineers who were let go and can't even get an interview right now, the industry is fecked. Mad that just a few years ago companies would have been throwing everything they could to get people like them. Feels like there's no industry out there as volatile as tech.

And that was the start. They overhired and later had to correct their mistakes by laying people off. And they have also caused more people wanting to be a SE due to perks like WFA, high salary, etc. very sexy back then but not even relevant anymore unless if you work for top tech companies. So now we have an oversupply of engineers in tech. I've seen Linked In job posting that gets filled by 100+ candidates within hours.

 

[Pexbo]

AI is decimating the creative industry sector. I'm currently adjusting myself and my work to incorporate it more and more. If you have any young people wanting to get into creative industries I'd strongly advise against it.

I work in a creative tech house and our design team are all UX first. We had a network of illustrators which we have worked with for 10+ years and I was speaking to the head of design a couple of weeks ago and he said they haven’t contracted anything to them in the last 18 months because they use AI instead now. The only assets we have received from external sources are from brands who control their IP.

I know very strong, experienced software engineers who were let go and can't even get an interview right now, the industry is fecked. Mad that just a few years ago companies would have been throwing everything they could to get people like them. Feels like there's no industry out there as volatile as tech.

Honestly it’s fecking scary. I might be naive but I think my job is relatively for the next year or two but beyond that who knows? Even in a secure role the glut of out of work developers is only going to make our salary trajectory shite.

We had a AI crisis workshop last week lead by our head of tech and he basically said that we are in a good position being a small tech house because there’s very little red tape. We can start playing with these technologies now and offering AI solutions to our customers and begin to capitalise on it today while the bigger companies (hopefully) waste time setting up their policies and teams and are slower picking up new technologies as it drops.