AI Isn’t Making Much Money

[erek]

Ai = Bitcoin buzz imo

“There’s also a longer-term worry that AI could dent the software-as-a-service business altogether. Most computer applications are charged per user, so revenue growth is premised on corporate customers continuing to hire.

But one of the main pitches of AI is efficiency – you need fewer customer service reps to serve the same user base, for example. Not all companies will reinvest labor savings into hiring more people. Job losses will hurt future growth for software based on “seats,” the number of workers using it.

Kash Rangan, an analyst at Goldman Sachs, posed the question to Adobe’s management during an earnings call last week. “Will generative AI be so good that it’s the end of the creative process — so we don’t need creative folks,” he asked, noting it was a major point of debate for investors.


Since AI will make software easier to use, it will expand the customer base, replied Chief Executive Officer Shantanu Narayen. As for hardware and infrastructure getting all the attention in this wave of AI spending, the veteran Adobe CEO said that customer focus will eventually swing back to use of AI within applications, or else all that investment in chips and servers wouldn’t have been worth it.”

Source: https://www.bloomberg.com/news/news...struggle-to-make-money-from-boom-in-ai-demand

 

[LukeTbk]

So you disagree with the article you posted that predict dramatic change and the promise that AI will be deflationary ? That it is a false buzz and bad article ?

 

[DukenukemX]

This is what I've been saying in that nobody has yet found a way to make this profitable. This is largely because everyone expects this to take peoples jobs, which hasn't happened. At least, not yet. The problem is that AI isn't accurate enough to make it worth ones wild. A lot of AI models are built to always answer a question, regardless how correct it is. A lot graphics revolving around it are still going to have continuity issues, because again it's not very good at it. At some point it'll get there, but it isn't now. You're more likely to see individuals or small companies make use of this technology, while more bigger corporations sit back and wait to see if it's time to jump in. We've come a long way from Smith eating spaghetti, but we're still a good ways off from it looking 100% realistic.


View: https://youtu.be/XQr4Xklqzw8?si=qrdWd74LmM_kaspP

 

[emphy]

This is what I've been saying in that nobody has yet found a way to make this profitable. This is largely because everyone expects this to take peoples jobs, which hasn't happened. At least, not yet. The problem is that AI isn't accurate enough to make it worth ones wild. A lot of AI models are built to always answer a question, regardless how correct it is. A lot graphics revolving around it are still going to have continuity issues, because again it's not very good at it. At some point it'll get there, but it isn't now. Your more likely to see individuals or small companies make use of this technology, while more bigger corporations sit back and wait to see if it's time to jump in. We've come a long way from Smith eating spaghetti, but we're still a good ways off from it looking 100% realistic.


View: https://youtu.be/XQr4Xklqzw8?si=qrdWd74LmM_kaspP

As I may have mentioned before: the biggest impact of currently hyped "ai" will be that "big data" will be running on users' devices instead of companies' servers. Given the huge numbers involved, even a tiny improvement over current analytics are going to increase the profit margins significantly, with the added advantages of being "privacy friendly" and the cost of processing the data being payed for by the users.

That juicy 24hour/day data will be used, not to mention the potential benefit of timing one's "advertisement" to the precise moment that there is the highest chance of a resulting sale. The only surprise here was that microsoft beat google at going for that cookie jar first.

 

[Tsumi]

It's a powerful assist tool, not a replacement for someone making decisions. For the next few years at least, all it's going to do is replace grunt work labor. Someone still needs to be in charge of deciding what the correct output should look like.

 

[LukeTbk]

This is largely because everyone expects this to take peoples jobs, which hasn't happened.

This is even a different question, AI could take people jobs with open source and easy to run models, there a world were AI is deflationary (and you need it to survive not to make more money that before but less) where the giant winner is the customer. If everyone has access to it and everyone do more with less employee, usually competition push cost down, you do not make more profit than before, you did it to not close.

 

[LukeTbk]

It's a powerful assist tool, not a replacement for someone making decisions. For the next few years at least, all it's going to do is replace grunt work labor. Someone still needs to be in charge of deciding what the correct output should look like.

Hard to predict for me, in software for example, there are so much to do it could be larger output than replacement.

There a lot of possible, AI get so good it destroy our business scenario that they are talking about, turning complex expensive stuff in commodity, but that still quite far away I think.

 

[DukenukemX]

It's a powerful assist tool, not a replacement for someone making decisions. For the next few years at least, all it's going to do is replace grunt work labor. Someone still needs to be in charge of deciding what the correct output should look like.

It's the other way around. The grunt labor work isn't going anywhere. The guy lifting boxes won't be replaced by AI, but the desk jockeys will be replaced. If you're an artist then AI will scare you. There's a number of artists online right now that hate AI so much that they now advertise that they didn't use AI.

As I may have mentioned before: the biggest impact of currently hyped "ai" will be that "big data" will be running on users' devices instead of companies' servers. Given the huge numbers involved, even a tiny improvement over current analytics are going to increase the profit margins significantly, with the added advantages of being "privacy friendly" and the cost of processing the data being payed for by the users.

That juicy 24hour/day data will be used, not to mention the potential benefit of timing one's "advertisement" to the precise moment that there is the highest chance of a resulting sale. The only surprise here was that microsoft beat google at going for that cookie jar first.

I think having AI run native and expecting to siphon data from it, is going to be a problem. There's going to be enough backlash that some businesses might try and avoid their workers using these devices where AI is ran locally. Look at the backlash Microsoft has received with Copilot and Recall. There's no black tape to put on a web cam this time.

 

[M76]

Corporations use AI as if you brought a smartphone to a thus far isolated tribe of humans. They'll think it is magic, and when the battery runs out they start using it as a potato peeler.

 

[paradoxical]

It's the other way around. The grunt labor work isn't going anywhere. The guy lifting boxes won't be replaced by AI, but the desk jockeys will be replaced. If you're an artist then AI will scare you. There's a number of artists online right now that hate AI so much that they now advertise that they didn't use AI.

The guys lifting boxes are going away pretty soon, and AI isn't even need to displace them. Cobots are getting cheaper and cheaper; the main reason they aren't already replacing guys unloading trailers pallets is because the safety regulation requires robots working around humans to either be walled off in a protective cell OR moving capped at a slow speed. But this is going to change soon with computer vision (not ML based, just normal algorithms); my company was engaged to use our embedded hardware to accelerate a company's camera-based safety solution which interfaces with robot's PLCs to slow them down when a human enters a certain radius. They are going for safety certification in cooperation with some large companies and if it gets certified you can expect to see the speed limitations removed from cobots (collaborative robots) in which case the ROI on them will vastly outpace that of human workers lifting boxes. It will probably take a couple years to get safety certification, but it's coming - make no mistake.

 

[sharknice]

It's the other way around. The grunt labor work isn't going anywhere. The guy lifting boxes won't be replaced by AI, but the desk jockeys will be replaced. If you're an artist then AI will scare you. There's a number of artists online right now that hate AI so much that they now advertise that they didn't use AI.

I think having AI run native and expecting to siphon data from it, is going to be a problem. There's going to be enough backlash that some businesses might try and avoid their workers using these devices where AI is ran locally. Look at the backlash Microsoft has received with Copilot and Recall. There's no black tape to put on a web cam this time.

My wife is an actual professional artist and she loves AI. She uses it to do her job faster, and got her whole company's marketing and art departments updated to adobe firefly. They haven't layed anyone off, they just do a better job because they have more time to spend on other details.

Most of the AI art hate that I've seen online has been from SJWs that aren't artists and don't even know any artists they claim to be fighting for. They think a professional artist is someone that sells images of lewd anime girls on patreon.

 

[Tsumi]

It's the other way around. The grunt labor work isn't going anywhere. The guy lifting boxes won't be replaced by AI, but the desk jockeys will be replaced. If you're an artist then AI will scare you. There's a number of artists online right now that hate AI so much that they now advertise that they didn't use AI.

I think having AI run native and expecting to siphon data from it, is going to be a problem. There's going to be enough backlash that some businesses might try and avoid their workers using these devices where AI is ran locally. Look at the backlash Microsoft has received with Copilot and Recall. There's no black tape to put on a web cam this time.

There is absolutely grunt work in things like art. Shading, rotoscoping, 3D tracking, and more. Most artists hate doing that grunt work because it's repetitive and takes time away from the creative portion.

AI art still has no soul, as it were. You can get a pretty image but it won't give you exactly what you want, unlike commissioning a human to create art for you. People that are okay with the AI art they get were never going to pay much for commissioned work anyways so it's hardly a loss there.

 

[LukeTbk]

There is absolutely grunt work in things like art.

And law, and medecine, and everywhere.

The plumber and nurse will not be replaced anytime soon, maybe personnal assistant will do part of what they do, but box lifting it has started a long time ago but like pointed out robots security issue slowed things down, if the aging population and work shortage continue to explode has expected in Corea-Germany-Japan-China, it could regain a push and go fast, China workforce population could from a billions in 2010 to under 400 millions by 2100, the amount of wanted work for the giant retired population with the low amount of worker should push a lot of it

 

[DukenukemX]

The guys lifting boxes are going away pretty soon, and AI isn't even need to displace them. Cobots are getting cheaper and cheaper; the main reason they aren't already replacing guys unloading trailers pallets is because the safety regulation requires robots working around humans to either be walled off in a protective cell OR moving capped at a slow speed. But this is going to change soon with computer vision (not ML based, just normal algorithms); my company was engaged to use our embedded hardware to accelerate a company's camera-based safety solution which interfaces with robot's PLCs to slow them down when a human enters a certain radius. They are going for safety certification in cooperation with some large companies and if it gets certified you can expect to see the speed limitations removed from cobots (collaborative robots) in which case the ROI on them will vastly outpace that of human workers lifting boxes. It will probably take a couple years to get safety certification, but it's coming - make no mistake.

The problem is that cobots require very expensive hardware while desk jockeys do not. Not that automation won't replace both these type of jobs, but that AI software is not only going to be cheap but sometimes free. You'll see people with college education losing jobs before the guy moving boxes.

My wife is an actual professional artist and she loves AI. She uses it to do her job faster, and got her whole company's marketing and art departments updated to adobe firefly. They haven't layed anyone off, they just do a better job because they have more time to spend on other details.

They will eventually lay off people if people using AI tools are able to get things done faster. They have not yet done so, but they eventually will.

Most of the AI art hate that I've seen online has been from SJWs that aren't artists and don't even know any artists they claim to be fighting for. They think a professional artist is someone that sells images of lewd anime girls on patreon.

The hate comes from the fact that anyone can type in a prompt to get results that are comparable to that of a professional artist. A lot of artists make their money off commissions and tools like Adobe FireFly basically took their jobs.

 

[Okatis]

This is largely because everyone expects this to take peoples jobs, which hasn't happened. At least, not yet.

I mean, given the post seems to be about generative output, I have a whole section of saved web pages where this has happened. Illustrators being replaced by generated art for books, article graphics using generative images instead of even stock images (this can be seen all over the web now), customer service, writers and editors (see: Microsoft firing them from MSN, also a Gizmodo division). There's even a business that has been running for more than a year for corporate headshots using generative models instead of a real photographer (and it's cheap).

There's a certain threshold below which people are apparently content with the quality/price/speed of using generative models instead of humans and it's already been happening. As the quality increases it'll inevitably impact more.

Edit: just recalled another: 2D animation matte painters being replaced at one studio. Someone I chatted with directly was impacted.

Edit 2: Hackaday has begun using generated graphics for articles, despite their usual illustrator Joe Kim. It seems unlikely it's merely some stock image that happens to be generated since the thumbnail for the article seen on the home page is a square, different variant of the same prompt (ordinarily when Hackaday use stock images they use one and just crop it).

 

[westrock2000]

Our species does seem to have a really good track record of finding innovative ways to get rid of jobs, yet requiring more to be created as a result. We’ve been innovating people’s jobs out of existence for thousands of years. Yet we still all have jobs. Maybe this time will be different.

 

[sharknice]

They will eventually lay off people if people using AI tools are able to get things done faster. They have not yet done so, but they eventually will.

The hate comes from the fact that anyone can type in a prompt to get results that are comparable to that of a professional artist. A lot of artists make their money off commissions and tools like Adobe FireFly basically took their jobs.

Your first point could be true. It depends how much a company wants to balance the quality of the work and how much time they let someone spend on it. Where my wife works they have more things they want done than they have people with time to do it, and pretty much always will. So AI is simply letting them get more done.


Most professional artists work for companies doing marketing materials, packaging, signage, brochures, website, etc. Very few people make money simply selling pictures like what you're describing in point 2. The person who lost their job to image genertion is figmant of an SJW's imagination.

 

[Decko87]

Your first point could be true. It depends how much a company wants to balance the quality of the work and how much time they let someone spend on it. Where my wife works they have more things they want done than they have people with time to do it, and pretty much always will. So AI is simply letting them get more done.


Most professional artists work for companies doing marketing materials, packaging, signage, brochures, website, etc. Very few people make money simply selling pictures like what you're describing in point 2. The person who lost their job to image genertion is figmant of an SJW's imagination.

I actually mostly agree with this as someone who works in the games industry and a developer. I don't see AI anywhere close to replacing our artists, that's including environment art, modeling, concept art, and animations. It does some stuff that helps them a bit, but for the most part it's pretty lack luster for making a polished product. I think it's the same thing for programming, because I know what I want from GPT I can prompt it in a way that will give me a correct answer I can use as a starting point and build my systems / feature around. It's not always exactly what I want and I edit everything a lot but it does speed us up a quiet a bit. The issue is again, people asking it to one shot make you a very complicated program that it hasn't been trained on is not going to happen. But if you ask it about specific patterns or implementations it can build 80-90 percent of what you need. You just need to be able to go through and correct it so it fits what you're trying to do because they don't actually understand anything.

Here's an example: I'm building a spawning system for my game the other week, I have a new pattern in mind that's data oriented and I want to use an old school java builder pattern to aggregate the enemy's functionality, model, animations, etc. I also want to instantiate the models asynchronously in a pooling system I was working on. So I wanted to use a builder pattern, a bit of dependency injection, and C# tasks or the jobs system in Unity. I also had a ton of initialization that would need to be done to inject dependencies into each enemy when they're spawned. That's not something an AI is going to do without the user being able to ask a series of questions, and review the code it's putting out to make sure it's good enough or efficient enough for what you're trying to do. Maybe it will some day, I'm not really sure how if a trained professional isn't asking the questions, but maybe it does at some point. I've seen some of the self prompting AI Agent stuff, and it's a cool concept but I'm not sure it's there yet. I could be totally wrong and dead broke in 2027 or something, who knows.


It's been a big productivity boost for our studio overall, but we're still over budget all of the time and trying to speed ourselves up to get more done any way we can. I think there's a reasonable chance we may have had to hire another programmer without it, but it's unlikely it's that much of a speed up. That or we'd just have to push our deadline out a few more months.


The real interesting part of the article to me is that they seem to understand mass layoffs are bad for the economy, anyone who's taken an economics course understands the velocity of money. As soon you have mass layoffs it causes problems down the line, and if those jobs aren't coming back it'll be a massive drain on everyone except the very wealthy. You can automate as much as you want away, but even for the large corporations eventually their balance sheets are going to have major problems as the work force declines.

 

[michalrz]

I have ChatGPT 3.5 (the one with additional chromosomes) opened in a tab at all times. It's great having a kind of an assistant to answer all sorts of silly questions.
Just today it told me how to fix a wi-fi development board thingie, the fix worked, can't complain.

Yes, it also came up with this flower of a design:

But generally we meet half-way and it's a productive relationship. I google way less nowadays.

 

[DukenukemX]

Our species does seem to have a really good track record of finding innovative ways to get rid of jobs, yet requiring more to be created as a result. We’ve been innovating people’s jobs out of existence for thousands of years. Yet we still all have jobs. Maybe this time will be different.

There's a few caveats to this. One is that we haven't seen anything like this before in history. When in the past thousands of years have you seen a person able to talk to a machine and be tricked into thinking it's human? It's gotten to the point where thieves are now using AI to automate phone calls and ask people for 2 step verification codes. The second issue here is that automation has usually taken jobs that were simple to do, but now with AI it can actually do more complicated tasks. It can draw art, sound very human, and even help program. It currently does them imperfectly, but that will likely change. The final issue is that AI is aimed at people who work in front of computers. The factory jobs were already taken by automation, and when that happened then people moved to desk jobs. Once desk jobs are gone, then what? Sex slavery to the wealthy? Become a professional ball washer?

 

[LukeTbk]

One is that we haven't seen anything like this before in history.

The ability to go from AI is a money-hoax about to burst and nothing new versus if-else algorithm of the 80s, to nothing like this it could destroy jobs from a day to the other is a bit impressive.

 

[A I]

I can confirm, I am indeed broke.

 

[Lakados]

This is even a different question, AI could take people jobs with open source and easy to run models, there a world were AI is deflationary (and you need it to survive not to make more money that before but less) where the giant winner is the customer. If everyone has access to it and everyone do more with less employee, usually competition push cost down, you do not make more profit than before, you did it to not close.

Advances in OCR and AI-assisted document imaging could be a huge thing for the medical industry, it should allow them to do a far better job of digitizing documents than they could previously.
AI has a strong possibility to make a lot of brainless data entry jobs go away, same with a lot of searching jobs, you can pay a group of 20 people to pour through legal journals for precedents and such or you can ask an AI to do it and what might take a team of people a full night the AI can pull up in a few hours.
AI has a lot of advantages as a tool, but its a long ways away from replacing people outright.

 

[paradoxical]

The problem is that cobots require very expensive hardware while desk jockeys do not. Not that automation won't replace both these type of jobs, but that AI software is not only going to be cheap but sometimes free. You'll see people with college education losing jobs before the guy moving boxes.

The problem is that people require very expensive salaries and robots do not. The problem is that people require workers comp and can sue you, and robots do not. The problem is that people moving boxes just don't show up half the time and when they do they are high; robots don't get high and always show up. The problem is that people legally require breaks, and robots do not.

The reason a company is paying my company to accelerate this stuff is because the ROI on cobots is way, way better than people if the safety regulations allow them to move faster.

 

[Lakados]

The problem is that people require very expensive salaries and robots do not. The problem is that people require workers comp and can sue you, and robots do not. The problem is that people moving boxes just don't show up have the time and when they do they are high; robots don't. The problem is that people legally require breaks, and robots do not.

The reason a company is paying my company to accelerate this stuff is because the ROI on cobots is way, way better than people if the safety regulations allow them to move faster.

Warehousing and Agriculture are getting massive AI overhauls, it's crazy the new stuff going on there.

 

[Decko87]

Warehousing and Agriculture are getting massive AI overhauls, it's crazy the new stuff going on there.

That sounds cool, what does it mean exactly?

 

[sharknice]

Warehousing and Agriculture are getting massive AI overhauls, it's crazy the new stuff going on there.

It's crazy new stuff, but historically it isn't anything crazy new.

Look at the history of agriculture. Now about 1% of the population works in agriculture, but it used to be nearly everyone. Technology has shrunk the amount of people involved. Sometimes slowly, sometimes rapidly with things like the cotton gin.

Sewing machines made it so a single person can be 100X faster. Sewing machines were actually banned places, tailors formed mobs and destroyed them, etc.

Same for numerous other industries.

It was a huge net positive for the world, food became cheaper, clothing became cheaper, etc., the people negatively affected move on.

 

[paradoxical]

That sounds cool, what does it mean exactly?

We recently came back from this trade show: https://www.automateshow.com/

The entire show was about how you can get rid of people with affordable, effective robots. They had robots that could stack boxes in pallets, unload tractor trailors, count things, etc. Basically, everything is getting automated to an astounding degree and stuff that is repeatable (like stacking boxes, picking product, etc) is the easiest to automate. In a lot of ways it's easier to automate this well than white collar jobs like customer service which are more open ended. ML is being used to handle a lot of pathing stuff right now; some of the robots have millions of ways they can move (each time you have a joint with a degree of freedom it exponentially increases the possibilities) and ML is starting to get good enough where it can optimally path in real-time based on sensor data inputs.

 

[MavericK]

Seems to be making a ton of money for companies like Nvidia.

 

[DukenukemX]

The ability to go from AI is a money-hoax about to burst and nothing new versus if-else algorithm of the 80s, to nothing like this it could destroy jobs from a day to the other is a bit impressive.

You act like it can't be both. It's like the dot com bubble when everyone invested into websites but the market did crash because only a small percentage of websites were actually profitable. The internet is still clearly a thing, and it's shutting down brick and mortar stores. Amazon took jobs away from retail but then redirected jobs to warehouses and distribution. Taxi’s and food delivery are being replaced by Uber and Lyft. For AI to be profitable it must take away jobs. People have certainly attempted to do this with lawers, driving, and writing articles. You could see it as tools that make it easier for one to code, make art, and voice acting but ultimately it pushes towards the idea of firing people. If AI makes you more productive then eventually you'll be forced to produce what is effectively 2 cows... I mean two people. The only reason this won't work so well in profitability is that everyone and their grandma is making their own AI model. Especially if and when the open source tools work just as well. Again, another reference to the dot com boom, but who here hasn't made their own website?

 

[Nasty_Savage]

AI is an elaborate cut and paster. Let me know when it develops a water powered car i can fill with my neighbors garden hose.

 

[DukenukemX]

The problem is that people require very expensive salaries and robots do not. The problem is that people require workers comp and can sue you, and robots do not. The problem is that people moving boxes just don't show up half the time and when they do they are high; robots don't get high and always show up. The problem is that people legally require breaks, and robots do not.

The reason a company is paying my company to accelerate this stuff is because the ROI on cobots is way, way better than people if the safety regulations allow them to move faster.

Again, the robots cost tens of thousands to buy and even then they do have limits. A company dealing with someone who makes $40k a year isn't as much as a problem as the guy in accounting who's making three figures. Who would a corporation focus more on replacing? Bob sitting in front of a computer that can probably be replaced by software, or Rodriguez the man who moves boxes? The three figure salary guy is much easier to replace and far more tempting. Don't get me wrong, the guy moving boxes will lose his job but don't think he's alone or even first.

View: https://youtu.be/0MCMB53jlIk?si=fVpBDzHWxtbEF8Kq

 

[Lakados]

That sounds cool, what does it mean exactly?

“Temporary Foreign Workers” are too much of a PITA for the industry to deal with correctly, so they are turning to drones.

Automated equipment is capable of picking fruit and berries 24/7 with minimal damage to the plants. Drones fly overhead and identify sick plants and either call in for treatment or culling. Pests and their signs are identified and dealt with on a plant by plant basis minimizing pesticides.

Automated tillers and water sensors level the fields and spot water as needed to keep pooling and waste to a minimum.
Pair all that with vertical farming techniques and it’s crazy what’s happening there.

Warehouses are all about inventory management and identification. Automated forklifts and self retrieving racks, drones moving around on 24/7 fetching and boxing quests. Amazon working on tech for automatically loading and unloading trucks etc.

Pair that with the new self driving all electric semi’s and bam.

For anybody familiar with the trucking industry it’s incredibly predatory and very reliant on foreign workers. Amazon and lots of others looking to cut the human component. Self driving semi’s are making a lot of runs from California to Nevada already.

 

[Lakados]

Seems to be making a ton of money for companies like Nvidia.

Nvidia has been working on visual identification and processing tech for a decade and change. That forms the backbone of pretty much every automated or self driving robot out there.
The Jetson/Orin boards exist for a reason.

 

[paradoxical]

Again, the robots cost tens of thousands to buy and even then they do have limits.

A company dealing with someone who makes $40k a year

Exactly, lmao. Both people and robots cost tens of thousands of dollars. Except a robot costs that once and humans are tens of thousands of dollars FOREVER and EACH YEAR with all kinds of problems including usage laws, liability, benefits, performance, wage increases, etc, etc, etc.

isn't as much as a problem as the guy in accounting who's making three figures.

Thank god you're not in accounting, but that's also probably why you don't understand ROI. Nobody in a company makes "three figures."

Who would a corporation focus more on replacing? Bob sitting in front of a computer that can probably be replaced by software, or Rodriguez the man who moves boxes? The three figure salary guy is much easier to replace and far more tempting. Don't get me wrong, the guy moving boxes will lose his job but don't think he's alone or even first.

View: https://youtu.be/0MCMB53jlIk?si=fVpBDzHWxtbEF8Kq

They will both get replaced eventually, but a lot of what Bob probably does has nuance to it that AI isn't advanced enough to do yet. I can't replace my finance people with AI, for example. Even my customer service folks I can't really get rid of; the cases are too complicated for AI.

But a dude who literally has one repetitive function - like removing a box from a pallet and stacking it in a corner? That guy is immediately screwed. But go ahead, don't believe me even though I have toured many 3PL warehouses and watched live as robots have slowly taken over their fulfillment centers while the finance people and office workers still have jobs.

 

[LukeTbk]

You act like it can't be both. It's like the dot com bubble when everyone invested into websites but the market did crash because only a small percentage of websites were actually profitable. The internet is still clearly a thing, and it's shutting down brick and mortar stores.

It can be incredible transformative technology (if people loose jobs by definition it is) while who will make money where from it is unknown (obviously who could known that ?), which is not all how you sound when you say it is all made up else-if.

For AI to be profitable it must take away jobs.

Or can make the same amount of people do more in industry where there is not a finite amount people would consume if they cheap enough and exist (almost all of them).

Excel added accounting jobs because the amount of accounting people loved to have if it is was that quick and that cheap to get explosed by order of magnitude, ATM augmented the amount of bank teller because they became profitable salesmen of product instead of a cost to make cheap transaction. Maybe it turn out we do not have more video game and they get cheaper to make (less jobs) or they become all GTA 6 level of complicated instead of being all simple except a couple of them with the same amount of people making them.

The only reason this won't work so well in profitability is that everyone and their grandma is making their own AI model. Especially if and when the open source tools work just as well. Again, another reference to the dot com boom, but who here hasn't made their own website?

The cheaper AI get the better and more transformative and more warranted the hype would be, the less money people make out of it who care ? Computers became incredibly cheap, people made more money from them than when they were expensive, because the number of things computer did (and the volume) go through the roff as it get cheap. Maybe all the service and software will be commodity and only the Dell-AMD-Nvidia and other in the hardware stack will make a lot of money, maybe, maybe not.

 

[Lakados]

Exactly, lmao. Both people and robots cost tens of thousands of dollars. Except a robot costs that once and humans are tens of thousands of dollars FOREVER and EACH YEAR with all kinds of problems including usage laws, liability, benefits, performance, wage increases, etc, etc, etc.



Thank god you're not in accounting, but that's also probably why you don't understand ROI. Nobody in a company makes "three figures."



They will both get replaced eventually, but a lot of what Bob probably does has nuance to it that AI isn't advanced enough to do yet. I can't replace my finance people with AI, for example. Even my customer service folks I can't really get rid of; the cases are too complicated for AI.

But a dude who literally has one repetitive function - like removing a box from a pallet and stacking it in a corner? That guy is immediately screwed. But go ahead, don't believe me even though I have toured many 3PL warehouses and watched live as robots have slowly taken over their fulfillment centers while the finance people and office workers still have jobs.

Back in the day I was sold out as an efficiency expert. Contract stipulated I would find my fee in inefficiencies or you would get a discount. Nobody got the discount…. I gutted accounting departments most notoriously “night auditing” I hated that job….

But even then Computers were taking their jobs with ease, all it took was changes to procedures and education on better reporting techniques.

I don’t do that job now, but if it was true then, it’s still true now, just more complicated to implement.

 

[paradoxical]

Back in the day I was sold out as an efficiency expert. Contract stipulated I would find my fee in inefficiencies or you would get a discount. Nobody got the discount…. I gutted accounting departments most notoriously “night auditing” I hated that job….

But even then Computers were taking their jobs with ease, all it took was changes to procedures and education on better reporting techniques.

I don’t do that job now, but if it was true then, it’s still true now, just more complicated to implement.

Accounting is an interesting one that we struggled with when we acquired a company that had 120k+ transactions a year through multiple channels. What I found was that traditional accountants are pretty much completely incapable of understanding ecomm accounting (specifically with how inventory flows in and out of a business when it is complex assemblies, complex kitting, virtual bundles, assigning proper COGS/3pl fulfillment/import duties based on proper accrual accounting principles, etc), but all the AI tools are not yet quite there and can really make a mess of things. I wasted a LOT of money going through various full-time accounting people, supposedly specialized e-com accounting forms, controllers, etc. Ultimately, I found that someone that is tech-savvy and good with systems level thinking can actually vastly reduce the need for a traditional accounting department by leveraging new tools to automate a lot of what used to have to be manual journal entries. I think nowadays if the founder is tech-savvy and understands accounting (just doens't have the time to do it) and can have a talented person own a lot of the integrations (most of which are no-code), you can probably get away with a single controller and just an AP/AR person until you are north of $25mm in revenue, combined with a fractional CFO to handle any thorny M+A or tax stuff. Obvious if the CEO is illiterate in finance and can't pitch in with supervision the need for a more senior CFO or VP Finance will occur sooner.

 

[Decko87]

“Temporary Foreign Workers” are too much of a PITA for the industry to deal with correctly, so they are turning to drones.

Automated equipment is capable of picking fruit and berries 24/7 with minimal damage to the plants. Drones fly overhead and identify sick plants and either call in for treatment or culling. Pests and their signs are identified and dealt with on a plant by plant basis minimizing pesticides.

Automated tillers and water sensors level the fields and spot water as needed to keep pooling and waste to a minimum.
Pair all that with vertical farming techniques and it’s crazy what’s happening there.

Warehouses are all about inventory management and identification. Automated forklifts and self retrieving racks, drones moving around on 24/7 fetching and boxing quests. Amazon working on tech for automatically loading and unloading trucks etc.

Pair that with the new self driving all electric semi’s and bam.

For anybody familiar with the trucking industry it’s incredibly predatory and very reliant on foreign workers. Amazon and lots of others looking to cut the human component. Self driving semi’s are making a lot of runs from California to Nevada already.

Remind me never to drive on the higj way between Cali and Nevada if that's true. I've seen the automatic weeding truck that uses lasers to kill weeds. I didn't know it went that far.

 

[DukenukemX]

Exactly, lmao. Both people and robots cost tens of thousands of dollars. Except a robot costs that once and humans are tens of thousands of dollars FOREVER and EACH YEAR with all kinds of problems including usage laws, liability, benefits, performance, wage increases, etc, etc, etc.

I really doubt these robots are as reliable as McDonalds ice cream machines, which needs repairs often to the point where it became an internet meme. Sometimes waiting months before a repair person can get to the damn things, and likely the only company who built them are able to repair them, thanks to the DMCA. More than likely they'll be paying for extended warranty because nobody is going to wait a day for these robots to get back up if something goes wrong. Assuming they keep the robots for many years as newer and better models will likely push companies to upgrade.

Thank god you're not in accounting, but that's also probably why you don't understand ROI. Nobody in a company makes "three figures."

I live in the greater part of NJ, and three figures is just what everyone makes here. I personally know two accountants who made that much, and one of them got laid off recently who made $200k. Take a guess why that person lost their job? The other one works at BD and she fears that her company maybe looking into automation as well.

They will both get replaced eventually, but a lot of what Bob probably does has nuance to it that AI isn't advanced enough to do yet.

Keep telling yourself that.

I can't replace my finance people with AI, for example. Even my customer service folks I can't really get rid of; the cases are too complicated for AI.

The reason you can't get rid of customer service is because AI would be just as capable as the automated systems companies already have in place. I can't talk to AI to get a refund or to find out why my order hasn't shipped.

But a dude who literally has one repetitive function - like removing a box from a pallet and stacking it in a corner? That guy is immediately screwed.

It's repetitive to you because for a human to do it is easy, but for a robot there's a lot going on. It's like driving where for a human it's no big deal, but so far nobody has made a self driving car that doesn't accidentally kill old woman crossing the road, or hits parked cars.

But go ahead, don't believe me even though I have toured many 3PL warehouses and watched live as robots have slowly taken over their fulfillment centers while the finance people and office workers still have jobs.

I wonder if there was any tech layoffs this year. There's certainly been a decline in warehouse jobs but again it's not a unique problem to physical labor. Let me put it like this, 75% of organizations want to adopt AI. Who's job is exactly safe from this?