How are you reacting to AI and it's influence on your job?

[Skipple]

I'm just curious to how everyone is handling to steady and increasing progression of AI and its eventual influence on the everyday life of your white-collar working individual.

 

I think to an extent, it's safe to assume that well trained AGI (models that have equivalent reasoning power of specially trained individuals) will be here for better or worse in the next 3-10 years... depending on growth, efficancy, power, and other variables. 

 

I'm trying not to draw too many conclusions, but as someone who's livelyhood is highly dependent on software development and data analysis (I work in the ITSM space) it's hard not to look at this ever increasing march towards AI dominance. Imagining a future where my job is eliminated is not hard to envision. While I'm not the first, I'm on the list.

 

Its something that keeps me up at night. I've seriously considered leaving my, fairly lucrative, job and becoming an apprentice in a trade, such as electrical work, to simply buy myself more time and allow society to adjust to the eventual economic overhawl that will be needed.

 

Am I being dramatic? Perhaps.

Unrealistic? I don't think so. 

 

I'm having a difficult time processing  through the day to day pretending like there isn't a lion over my shoulder that's going to destroy day to day life as we all know it in a matter of a few short years. 

 

How are you are yours reacting to this change? 

[dizmo]

White collar*

Very different meaning to white color.

[05032-Mendicant-Bias]

1 hour ago, Skipple said:

Its something that keeps me up at night. I've seriously considered leaving my, fairly lucrative, job and becoming an apprentice in a trade, such as electrical work, to simply buy myself more time and allow society to adjust to the eventual economic overhawl that will be needed.

You do you, I know people that have left IT to become farmers, but it's not because they think AI is coming for them.

 

Historically, when a new tool has emerged  to dramatically lower the marginal cost of production, demand has gone up, not down.

 

Historically there has been disruption in the current batch of workers in the field, that lose their job and need to reskill or move to other fields.

 

Historically, everyone, everywhere is better off, including the people still working in the field that got the better tool. it's something called induced demand.

 

It's a trend unbroken since the discovery of fire.

 

When camera displaced portrait artists, it unlocked photography, and there are uncountably more photographers than ever were portrait artists.

 

Or when photoshop and blender replaced rotoscopers, it lead to a vast increase in animation quality and quantity.

 

Or when increasingly better calculators replaced rooms full of human doing calculations. Nobody today has the job of just doing math operation and there are more clerks than ever.

 

Agricolture instead has a seen a vast reduction in workforce, those people moved to industry. Then industry has seen a vast reduction in workforce, those people moved into service. It's anybody's guess if we'll see a vast reduction in service work, or if we'll generate more service work.

 

On the industry side, there still are people employed in industry. It's just they no longer screw a bolt 14 hours a day, but spend eight hours monitoring the machines, keeping them maintained and fed, a much better job. I do work in industry, a field that has seen multiple waves of automation, and no machine is taking my job anytime soon, despite machine taking industry jobs for a century. As a matter of fact, more machines, mean more work to be done! I welcome cheaper motors, cheaper electronics and cheaper robots. Clients and suppliers that did not have robots, now are considering them. lowering cost mean more application now become cheaper with robots.

 

Personally, I suspect that lowering the marginal cost of doing computer programs, will vastly increaase the amount of programming needed. After all, software is increasingly inside everything, and it's easy to see a new wave of generic automation. Automated road maintenance, automated homes, automated grocery deliveries, etc... SOMEONE has to write all that code, machines still lack intent, and will lack intent for the foreseeable future.

 

 

[LogicalDrm]

I just started testing effects of MS Copilot for my work. So far, I'm between unimpressed and deciding its worth for my employers to pay €27/mo for the license. Currently ways to use:

• Make Python scrips for me
• Geocode addresses for me (I could do this with ArcGIS Online credits already. But our org is already using them in phase that will not last until the end of license term)

I don't do word processing or mass reading, so I don't need those capabilities. I also noticed that using it to google for me is worse than me doing googling and reading results old-fashioned ways. As example, I tried to troubleshoot issue where some of our new laptops cannot connect to database while others can. It did gave help, but very surface level stuff I already knew weren't the issue. Like how to connect, not why it wasn't working. 

[manikyath]

any business that's replacing people with AI is doing so out of a complete disregard for quality... and IMO that's not the sort of place you want to work for anyways.

 

as for how i react to this... we have a bunch of machines at work that are supposed to do the most mundane things for us, and most of them dont work reliable enough to not have a person standing next to them to fix them every 5 minutes... so i'm not worried AI is gonna change anything if even tasks that dont require any intelligence still require the intuition of a skilled human being.

[PDifolco]

1 hour ago, 05032-Mendicant-Bias said:

You do you, I know people that have left IT to become farmers, but it's not because they think AI is coming for them.

 

Historically, when a new tool has emerged  to dramatically lower the marginal cost of production, demand has gone up, not down.

 

Historically there has been disruption in the current batch of workers in the field, that lose their job and need to reskill or move to other fields.

 

Historically, everyone, everywhere is better off, including the people still working in the field that got the better tool. it's something called induced demand.

 

It's a trend unbroken since the discovery of fire.

 

When camera displaced portrait artists, it unlocked photography, and there are uncountably more photographers than ever were portrait artists.

 

Or when photoshop and blender replaced rotoscopers, it lead to a vast increase in animation quality and quantity.

 

Or when increasingly better calculators replaced rooms full of human doing calculations. Nobody today has the job of just doing math operation and there are more clerks than ever.

 

Agricolture instead has a seen a vast reduction in workforce, those people moved to industry. Then industry has seen a vast reduction in workforce, those people moved into service. It's anybody's guess if we'll see a vast reduction in service work, or if we'll generate more service work.

 

On the industry side, there still are people employed in industry. It's just they no longer screw a bolt 14 hours a day, but spend eight hours monitoring the machines, keeping them maintained and fed, a much better job. I do work in industry, a field that has seen multiple waves of automation, and no machine is taking my job anytime soon, despite machine taking industry jobs for a century. As a matter of fact, more machines, mean more work to be done! I welcome cheaper motors, cheaper electronics and cheaper robots. Clients and suppliers that did not have robots, now are considering them. lowering cost mean more application now become cheaper with robots.

 

Personally, I suspect that lowering the marginal cost of doing computer programs, will vastly increaase the amount of programming needed. After all, software is increasingly inside everything, and it's easy to see a new wave of generic automation. Automated road maintenance, automated homes, automated grocery deliveries, etc... SOMEONE has to write all that code, machines still lack intent, and will lack intent for the foreseeable future.

 

 

I'd lean towards the same view, looking back history to previous work (r)evolutions

However I'm a bit more pessimistic as AI is coming hard and fast, too fast for people to adapt, and with not that much new opportunities created

If you look at the 19th century industrial revolution, people had to transfer from agricultural work to industry, but it was quite easy, neither required skill nor education, and there was a ton of jobs available

If you look at the first computer/IS revolution, it took place over like 40 years, from the 60s and their mainframes that only existed in a few big corps, to the 2000s and Internet, people have had the time to evolve/learn and transfer from industry to services

But now ? In less than 10 years AI will wreck the job landscape, most people won't have time to movbe on, and move on to where ?

In our era stakeholders take like 90% of any productivity gains, so if people lose their jobs, who will buy their stuff ? What services can be created to replace lost jobs ?

I'm pretty skeptical... 

[jaslion]

1 hour ago, manikyath said:

any business that's replacing people with AI is doing so out of a complete disregard for quality... and IMO that's not the sort of place you want to work for anyways.

 

as for how i react to this... we have a bunch of machines at work that are supposed to do the most mundane things for us, and most of them dont work reliable enough to not have a person standing next to them to fix them every 5 minutes... so i'm not worried AI is gonna change anything if even tasks that dont require any intelligence still require the intuition of a skilled human being.

This the moment they cut people for ai replacement you know quality is going to go down.

 

Since im a it consultant I see this over and over and over and over again.

 

I refuse to use generative ai out of principle as well as how this garbage is eroding understanding reading, critical thinking, problem solving ans learning ability of so many people that use it.

[PDifolco]

1 hour ago, manikyath said:

any business that's replacing people with AI is doing so out of a complete disregard for quality... and IMO that's not the sort of place you want to work for anyways.

 

as for how i react to this... we have a bunch of machines at work that are supposed to do the most mundane things for us, and most of them dont work reliable enough to not have a person standing next to them to fix them every 5 minutes... so i'm not worried AI is gonna change anything if even tasks that dont require any intelligence still require the intuition of a skilled human being.

Was the same with manufacturing going to India/China, didn't stop it... We're just going from AI -Actual Indians- to AI  

[05032-Mendicant-Bias]

1 hour ago, PDifolco said:

In our era stakeholders take like 90% of any productivity gains, so if people lose their jobs, who will buy their stuff ? What services can be created to replace lost jobs ?

Unfortunately this is a society issue, and in general, technology cannot solve society's problem.

 

The problem here is that productivity has gone up, worker wage has not. What has gone up is the income of the owner class. The problem is resource distribution.

The golden age in the USA and what they look fondly at, the american dream and the middle class, was brought about by strong worker unions, international dollar, and post war largesse. If you look what changed, it's the USA president Ronald Reagan 80s policy of deregulation and erosion of worker unions. And since the USA was the de facto economic center of the world, we all got dragged down with them.

 

The solution is as easy, as it is hard to implement. Working class people needs to demand that productivity gains are shared more fairly, with taxation on automation, taxation on capital, and taxation on top earners. It's done the same way it was done back then: protests and elections. It's a problem that has nothing to do with technology, and everything to do with curtailing greed.

  

32 minutes ago, PDifolco said:

Sure wealth distribution isn't a tech problem, but new techs leading to less workers make the problem much worse, it's a compound effect

Technology increasing productivity has no downside. You make more with less. If it was distributed fairly, we could even be working fewer days a week a week and keep everything else the same.

 

Technology is our ladder out of the problem we created for ourselves.

 

A pessimist scenario is a dystopia, where the rich own the automated factories, and retain a core of people and literally everyone else starve, as they aren't needed by the automation owners.

 

An optimist scenario is an utopia, where the fruits of the automated factories are shared fairly with citizens and citizens and everyone can have high standard of living, with living and working being fully decoupled. It's a paradigm shift. 

 

The owner class is lobbying to get control of all automation.

 

It's up to us working class to demand the fruits of labour be shared more fairly. it can b done, it was done by our forefathers when they e.g. obtained the 40h workweek and other concessions. The owner class fought the working class tooth and nail to prevent such concessions, e.g. with the pinkertons and other corporate militia. 

 

The united states in particular is regressing very fast. They now have company towns, again. Company scrips, again. Government founded by regressive taxation, again. Eroded worker protections, again. The USA owner class clearly wants a repeat of the glided age. But the world is increasingly less centered around the USA. What happens there isn't necessarly going to bound us all.

 

It's funny because I make fun of communism, I deeply believe communism is not possible because centralization communism demands inevitably give raise to dictatorships, and the utopian scenario is very much communism. A friend of mine is red and calls it "automated luxury communism" and it may very well be possible as long as power is not centralized and democracy is maintained.

 

What I envision as the victory scenario for everyone is just a strong socialist democracy with strong social nets and unions, something that was achieved before. the only difference is that machines increase productivity much more than in the past, and the same work results in higher standard of living. Europe has much stronger social nets than the USA. It is possible for resources to be spread more equitably.

 

An obvious solution if productivity increases, instead of keeping full time at 40h/week, you reduce the full time requirement as productivity goes up. Imagine an extreme scenario where productivity goes 40X, instead of having 1/40 of people employed 40h/week and getting 39/40 of unemployement, you get the same people working 1h/week. Or, those people are taxed so that everyon receive 38/40 worth of UBI and the one working gets 80/40 work of pay. Increasing productivity allows for so many combination where everyone everywhere is better off. It only requires for a fraction of the productivity gain to be redistributed with taxes/social programs.

[PDifolco]

Just now, 05032-Mendicant-Bias said:

Unfortunately this is a society issue, and in general, technology cannot solve society's problem.

 

The problem here is that productivity has gone up, worker wage has not. What has gone up is the income of the owner  class.

The golden age in the USA and what they look fondly at, the american dream and the middle class, was brought about by strong worker unions, international dollar, and post war largesse. If you look what changed, it's the USA president Ronald Reagan 80s policy of deregulation and erosion of worker unions. And since the USA was the de facto economic center of the world, we all got dragged down with them.

 

The solution is as easy, as it is hard to implement. Working class people needs to demand that productivity gains are shared more fairly, with taxation on automation, taxation on capital, and taxation on top earners. It's a problem that has nothing to do with technology, and everything to do with curtailing greed.

 

Sure wealth distribution isn't a tech problem, but new techs leading to less workers make the problem much worse, it's a compound effect

[Millios]

34 minutes ago, PDifolco said:

Was the same with manufacturing going to India/China, didn't stop it... We're just going from AI -Actual Indians- to AI  

Same downwards spiral, different level of how down it is 
Labour moved towards sweat shops not for quality but simply maximizing mass production and profits (same with AI now)

With AI they just maximise the bottom dollar rather than quality even more unlike the industrial revolution where they got almost everything (mass production, consistancy, profits) up without too many drawbacks simply cause we went from physical human limits to intelectual ones (as in how advanced machinery and tools got) and now its the moral ones which I ,personally, have little to no hope that they will be maintained cause shareholders care about their dimes and not their impact on the rest of the world

[PDifolco]

11 minutes ago, Millios said:

Same downwards spiral, different level of how down it is 
Labour moved towards sweat shops not for quality but simply maximizing mass production and profits (same with AI now)

With AI they just maximise the bottom dollar rather than quality even more unlike the industrial revolution where they got almost everything (mass production, consistancy, profits) up without too many drawbacks simply cause we went from physical human limits to intelectual ones (as in how advanced machinery and tools got) and now its the moral ones which I ,personally, have little to no hope that they will be maintained cause shareholders care about their dimes and not their impact on the rest of the world

Shareholders only care about profit for sure, but for that they still need clients, and if half of the world is on social care on in the streets it won't work anymore

[Eigenvektor]

4 minutes ago, PDifolco said:

Was the same with manufacturing going to India/China, didn't stop it... We're just going from AI -Actual Indians- to AI  

Or from AI to Actual Indians… https://medium.com/@divyanshbhatiajm19/builder-ais-1-5b-collapse-how-700-engineers-pretended-to-be-ai-and-got-away-with-it-for-years-2a0d35fc4af8

 

 

[Millios]

5 minutes ago, PDifolco said:

Shareholders only care about profit for sure, but for that they still need clients, and if half of the world is on social care or in the streets it won't work anymore

Small caveat

Plenty of ppl will spend whatever little money they have for the latest hot trend/item. I've seen plenty of ppl struggling with putting food on the table but rest assured they have the new newest iphone, they have brand (nike, adidas etc)  clothing and all the "frugal" items you can imagine

The shareholders I truly do thing will lead a cycle of "squeeze them dry -> let them recover a bit->rinse and repeat" especially with the way companies like apple or microsoft are behaving but thats my opinion from what I've personally seen

[05032-Mendicant-Bias]

18 minutes ago, Millios said:

The shareholders I truly do thing will lead a cycle of "squeeze them dry -> let them recover a bit->rinse and repeat" especially with the way companies like apple or microsoft are behaving but thats my opinion from what I've personally seen

This is Cyberpunk. it's is a possibility, even likely, it doesn't even need more technology, just less worker protections. The lore has lotteries for penicillin, and the protections there are there just because any more inequality would hit the bottom line.

 

Another dystopia scenario is more like Elysium. The rich don't really need other, if everything is automated. But this scenario needs more technology to happen, a lot more. And often rich are also narcisist, and want other people to adore them, I consider it fairly unlikely.

 

The various apocalypse scenarios are fairly unlikely as well. Humanity so far has proven good at avoiding catastrophe at the last second. E.g. with Cuba's crisis. We are so widespread, and with so much excess resources, that wiping us out is a great undertaking. Even if kicked back to the stone age, we'll just rebuild in a matter of centuries.

 

The fictional matrix scenario and terminator scenarios are fairly unlikely. The odds are stacked really against the machines here.

 

We get scant few utiopia scenario in fiction because they don't make for good writings if everyone, everywhere is fed and happy. Star Trek is one, Earth is basically paradise, and I find it more likely than the worst dystopias.

[PDifolco]

47 minutes ago, 05032-Mendicant-Bias said:

This is Cyberpunk. it's is a possibility, even likely, it doesn't even need more technology, just less worker protections. The lore has lotteries for penicillin, and the protections there are there just because any more inequality would hit the bottom line.

 

Another dystopia scenario is more like Elysium. The rich don't really need other, if everything is automated. But this scenario needs more technology to happen, a lot more. And often rich are also narcisist, and want other people to adore them, I consider it fairly unlikely.

 

The various apocalypse scenarios are fairly unlikely as well. Humanity so far has proven good at avoiding catastrophe at the last second. E.g. with Cuba's crisis. We are so widespread, and with so much excess resources, that wiping us out is a great undertaking. Even if kicked back to the stone age, we'll just rebuild in a matter of centuries.

 

The fictional matrix scenario and terminator scenarios are fairly unlikely. The odds are stacked really against the machines here.

 

We get scant few utiopia scenario in fiction because they don't make for good writings if everyone, everywhere is fed and happy. Star Trek is one, Earth is basically paradise, and I find it more likely than the worst dystopias.

Gotta get some missile launcher implants then

[Hideki Ryuga]

I'm pivoting careers out of software. I think before it was interesting and you could get paid a lot for your abilities and education, but it's honestly felt like people have flooded into my field that have no business being here, and that's includes tech bros and their wishful thinking about how their models will replace me, and the experts who believe they're ushering in humanity's great new future by advancing these models without regards to ethics.

 

The only real saving grace for me I guess is that everyone has convinced themselves that GPTs are human-like.

 

With how integrated software is in everything, I've always wondered why you wouldn't need a license to do it professionally like other engineers.

[TetraSky]

The most AI we have at my work, is an automated ticketing system through Microsoft Teams for IT.

That's all. And even that is rudimentary at best.

 

I don't see AI make a big stink at my workplace anytime soon. Except maybe for some office jobs that frankly should not even exist if not for nepotism.

 

Not only are we unionized, but we are in manufacturing(not cars). Machines are there to supplement workers. Not to take their jobs.

That said, my particular job is to fix and calibrate these machines as an electronics technician. So the more the better for my job security. 

[Ha-Satan]

I work in an academic library and my job involves physically handling the items, so I don't see AI being able to take my job any time soon. I'm honestly more worried about losing my job simply because everybody gets their information from online sources instead of wanting physical items. (This has the added downside of making everyone extremely dependent on the good will of a few mega-corporations for the distribution of academic works, which is terrible. My employers have recently made the insanely stupid decision that we can remove and discard any journals in the collection that are available online through JSTOR, which is so mind-bogglingly idiotic that I don't know where to begin.)

 

I am getting a BSIT in order to diversify my skillset. I'm fully aware that IT is also a field that's under threat from AI and outsourcing, which is why I'm trying to focus my skills towards the types of work that require a real, in-person employee (like printer maintenance).

 

But anyway, more generally I feel like the whole AI job apocalypse concern is maybe not as significant as you might think, simply because I think we are going to run into other problems sooner rather than later, like resources issues. Yeah, I'm a little worried about AI taking er jerbs, but what really keeps me up at night is water and fuel shortages.

[Skipple]

8 hours ago, Hideki Ryuga said:

With how integrated software is in everything, I've always wondered why you wouldn't need a license to do it professionally like other engineers.

Excuse me sir, do you have a license for that there Python 3 script? 

[gentlemanspot]

Sometimes I wonder if the same feeling was true about calculators 

[Hideki Ryuga]

5 hours ago, Skipple said:

Excuse me sir, do you have a license for that there Python 3 script? 

I mean I get that you're joking but I am serious.

 

Doing something in your garage or at home is one thing, but it's entirely different when you're given a database of people's sensitive information.

[Kisai]

23 hours ago, Skipple said:

 

I think to an extent, it's safe to assume that well trained AGI (models that have equivalent reasoning power of specially trained individuals) will be here for better or worse in the next 3-10 years... depending on growth, efficancy, power, and other variables. 

 

There will be no AGI models this century and anyone telling you otherwise is selling you snake oil.  Without some giant leap in material sciences for computer chips (eg optical chips) we've reached the point where expending more power doesn't lead to meaningful increases in performance. It would not surprise me if the next 10 years chips get less powerful on general purpose tasks due to demands to make them safer for portable devices.

 

Like you can not tell me having a 1200w laptop is ever going to be a possibility. They top out at 240w for a reason. 

 

Meanwhile a desktop can have a 600w pull for the CPU and the GPU separately. Hence, we're at the top. There is nothing more we can get by putting more power into the CPU and GPU and still have it safe to operate as a consumer device.

 

Cloud based AI models are already on the decline as privacy concerns make people shun these AI cloud services. 

 

What I expect is that TTS, ASR, Machine Translation, body landmark, face landmark, motion capture, object identification will all be local models, and current LLM's will be discarded in favor of subject matter expert models that don't hallucinate... once they stop trying to make LLM's be a worse version of wikipedia.

 

[GMUDuckman]

Software Engineer here, I think a large problem right now is that those who don't truly understand AI are demanding AI solutions. My company (about 350 people) recently mandated "AI Champions" for each practice area and the AI Champions had to come up with an AI strategy for their practice area. My colleague and I have been pushing more automation at our company for years and have usually been met with, "we don't want to automate people out of their jobs" and "there isn't a demand for it." Now, the clients ARE demanding AI, though they don't really know what they want and as a result my company is scrambling to keep up. Thanks to all this AI buzzword and AI washing, my colleague and I are now trying to explain how what most people are talking about are LLMs and that there are other AI techniques that could serve our work better. I was voluntold to be the AI Champion for the Tech Craft Center (combined practice area of Software/WebDev/IT), and during my strategy meeting I essentially told them our strategy was to evaluation various AI solutions and ensure they are bringing value and data security to the practice areas proposing the use of various AI tools. Leadership countered with "but how is your team going to use AI", to which I responded that we have been evaluating various coding LLMs and they have so far come up short. So far, I've seen anywhere from non-working code using functions that never existed, to working code that is using deprecated functions with security holes. The most success I've had with any generative model is claud and just using that for basically an autocomplete of a function I was starting to write, no significant time savings. For IT you do NOT want generative AI up in in your infrastructure. The fact I can put the same prompt into the same model and get different responses each time is the opposite of what you need for stability in your IT infrastructure. 

 

I think my thought got away from me there for a bit, maybe I should have used AI to write a better post. New to the community, but long time LTT watcher/purchaser.

[Skipple]

10 hours ago, Kisai said:

Without some giant leap in material sciences for computer chips (eg optical chips) we've reached the point where expending more power doesn't lead to meaningful increases in performance. It would not surprise me if the next 10 years chips get less powerful on general purpose tasks due to demands to make them safer for portable devices.

I don't see AI performance scaling linearly with compute. Efficiencies and AI performance appears to continue to improve exponentially despite the slow improvement in silicon. Perhaps we will hit a wall, and I hope we do for our own sake, I just haven't seen the wall yet.