This study found that between 2022-2024, there was a negative correlation between "jobs with high AI exposure" (i.e., tech jobs) and % change in wages. According to them, software engineering has both the biggest wage decline and most AI exposure.
There's a much more reasonable explanation here.. tech jobs had the most wage growth during COVID and this was a pullback. The time frame they used (2022-2024) was also pre-coding agents, where GPT-3.5/GPT-4 were frontier models.
> There's a much more reasonable explanation here.. tech jobs had the most wage growth during COVID and this was a pullback.
More concretely, companies do layoffs, visa workers have to find another job in 60 days. Companies like amazon try to scavenge and lowball the shit out of them. This is happening across the industry.
There's consolidation, acquisitions, and layoffs happening across industry. It's just a big game of musical chairs
At the end of the day, the real threat of AI is degradation of businesses, eventually leading to extinction of those of them still absolutely convinced AI will replace everybody. The real AI adoption is extremely low, but many business leaders are (willfully?) blind. Eventually, the pendulum will swing in other direction and there will be enormous demand for human labour. It will happen when businesses finally start to realize that AI is ruining their businesses.
If the companies find a new way to make their products cheaper to produce and their services cheaper to provide - will it lead to lower prices as well?
I think it’s more useful to apply OP’s question to about real, established businesses that already price products at the market rate. Not companies riding an unsustainable speculative bubble and playing pricing games in a desperate attempt to attract and retain users who they can later upsell or raise prices on to show “growth” and keep getting investor money.
Right, whats happening here is the first sales pitch to CEOs: "AI will let you fire everyone" is not panning out (duh)... so for the next trick, "AI will at least let you pay them less" is a natural pivot. Both are absurd. Just let the hypecycle die off in peace already.
We've seen this play out in history multiple times, already, so it's pretty ignorant to ignore the evidence.
Carpenters aren't being paid less because we have framing nailguns and pre-fabricated trusses. Instead their skillset had to evolve. In fact, smaller crews can do larger projects... which doesn't mean fire your crewmembers, it means a builder company can take on more projects in parallel, leading to bigger profits through volume.
"AI"s are useful tools. The skills in a post-AI world are different. Lower wages will be experienced by people that can't evolve, that's all.
AI is no longer the nail gun, that was auto complete with gpt 3.5. now it's like you say: "Build me a new roof" and an hour later it has materialized somehow. Not sure many carpenters would survive this.
Eventually AI will be able to just bootstrap solid security, payment, accounts, DB management, backups, customer support services that automatically link to the correct department or sub-system
I think the environment and landscape will change and different skills will be more valuable for developers
But I think the existing skill-sets developers use can mostly be automated away
If you were really good at your job as a contract software engineer in the past, you would basically just automate yourself out of a job and have to find a new one anyway. This just accelerates it.
> Eventually AI will be able to just bootstrap solid security, payment, accounts, DB management, backups, customer support services that automatically link to the correct department or sub-system
You are describing a templating system. It's useful to create projects from scratch, but that's not what software engineers do 99% of the time.
What software engineers do 99% of the time is modifying existing services to adapt to changing requirements that more often than not are grossly unspecified.
Also, a key part of a software engineer's work is serve as a point of contact with non software engineers to discuss what services do and plan work.
> If you were really good at your job as a contract software engineer in the past, you would basically just automate yourself out of a job and have to find a new one anyway. This just accelerates it.
I don't think this assumption holds water. At most your efficiency improved as short contracts to patch random customer projects now can meet the definition of done (maintenance be damned) in shorter iterations.
However, ironically we are also seeing the occurrence of consultancy firms specialized in providing software engineering services as contract engineers that are explicitly marketed as solutions to problems introduced by vibe-coding.
This study found that between 2022-2024, there was a negative correlation between "jobs with high AI exposure" (i.e., tech jobs) and % change in wages. According to them, software engineering has both the biggest wage decline and most AI exposure.
There's a much more reasonable explanation here.. tech jobs had the most wage growth during COVID and this was a pullback. The time frame they used (2022-2024) was also pre-coding agents, where GPT-3.5/GPT-4 were frontier models.
> There's a much more reasonable explanation here.. tech jobs had the most wage growth during COVID and this was a pullback.
More concretely, companies do layoffs, visa workers have to find another job in 60 days. Companies like amazon try to scavenge and lowball the shit out of them. This is happening across the industry.
There's consolidation, acquisitions, and layoffs happening across industry. It's just a big game of musical chairs
At the end of the day, the real threat of AI is degradation of businesses, eventually leading to extinction of those of them still absolutely convinced AI will replace everybody. The real AI adoption is extremely low, but many business leaders are (willfully?) blind. Eventually, the pendulum will swing in other direction and there will be enormous demand for human labour. It will happen when businesses finally start to realize that AI is ruining their businesses.
If the companies find a new way to make their products cheaper to produce and their services cheaper to provide - will it lead to lower prices as well?
It will. We saw this a lot with tvs, appliances, etc that are now manufactured where it is a lot less money and now they cost a lot less money.
Interestingly it didn't seem to happen with cars
It probably did in China. In EU/US we don't benefit from these low prices for protectionist reasons.
They are not much cheaper, but they are packed with stupid features no one wants
Not if it can lead to higher profit margins instead.
No, there has been a big push from big tech like Meta for more and more regulations to make sure competition couldn’t keep up
Hahahhahahahhahha….you serious?
Doesn't seem insane. OpenAI made Luna more efficient and dropped prices 80% this week.
I think it’s more useful to apply OP’s question to about real, established businesses that already price products at the market rate. Not companies riding an unsustainable speculative bubble and playing pricing games in a desperate attempt to attract and retain users who they can later upsell or raise prices on to show “growth” and keep getting investor money.
It will happen as long as there are competitors
It's probably both, but either way it means a worse life for people who work for a living
Right, whats happening here is the first sales pitch to CEOs: "AI will let you fire everyone" is not panning out (duh)... so for the next trick, "AI will at least let you pay them less" is a natural pivot. Both are absurd. Just let the hypecycle die off in peace already.
We've seen this play out in history multiple times, already, so it's pretty ignorant to ignore the evidence.
Carpenters aren't being paid less because we have framing nailguns and pre-fabricated trusses. Instead their skillset had to evolve. In fact, smaller crews can do larger projects... which doesn't mean fire your crewmembers, it means a builder company can take on more projects in parallel, leading to bigger profits through volume.
"AI"s are useful tools. The skills in a post-AI world are different. Lower wages will be experienced by people that can't evolve, that's all.
> Just let the hypecycle die off in peace already.
I so wish that would happen so we can all find a more sustainable middle ground, but the trillion or so dumped into it says otherwise.
AI is no longer the nail gun, that was auto complete with gpt 3.5. now it's like you say: "Build me a new roof" and an hour later it has materialized somehow. Not sure many carpenters would survive this.
it's not really comparable IMO
Eventually AI will be able to just bootstrap solid security, payment, accounts, DB management, backups, customer support services that automatically link to the correct department or sub-system
I think the environment and landscape will change and different skills will be more valuable for developers
But I think the existing skill-sets developers use can mostly be automated away
If you were really good at your job as a contract software engineer in the past, you would basically just automate yourself out of a job and have to find a new one anyway. This just accelerates it.
> Eventually AI will be able to just bootstrap solid security, payment, accounts, DB management, backups, customer support services that automatically link to the correct department or sub-system
You are describing a templating system. It's useful to create projects from scratch, but that's not what software engineers do 99% of the time.
What software engineers do 99% of the time is modifying existing services to adapt to changing requirements that more often than not are grossly unspecified.
Also, a key part of a software engineer's work is serve as a point of contact with non software engineers to discuss what services do and plan work.
> If you were really good at your job as a contract software engineer in the past, you would basically just automate yourself out of a job and have to find a new one anyway. This just accelerates it.
I don't think this assumption holds water. At most your efficiency improved as short contracts to patch random customer projects now can meet the definition of done (maintenance be damned) in shorter iterations.
However, ironically we are also seeing the occurrence of consultancy firms specialized in providing software engineering services as contract engineers that are explicitly marketed as solutions to problems introduced by vibe-coding.
[dead]