This subreddit should enable sharing and discussing economic research and news from the perspective of economists. Academic work and summaries are welcome. Image and video submissions are not allowed.
--
If you have any questions about this removal, please contact the mods.
Windows 11 is cratering so hard it will take Redmond with. It. AI stuffed into every nook and cranny, bloating and already famously bloated piece of software.
And I swear that autocorrect here on Reddit has become an imbecile.
It's also trying to learn from every interaction and steal from every author of every email/document - slowing down processing, instead of doing the basic job we are paying for when we are charged for the subscription every year.
At my job I can't run Outlook and Excel at the same time or my computer crashes. I'm an analyst, so you can see why this might be a problem. It only started a couple of updates ago
My Surface laptop has 8GB of RAM. When I turn it on, it's at 95% memory used without anything running. And then the daily 3 hour virus/malware scan starts that pegs my CPU at 100% the entire time. Basically can't use it half of the day. I managed to steal an old desktop computer from someone that retired even though it isn't allowed to have 2 computers. Then I scrounged another stick of RAM out of a computer in the e-waste pile. I've had a ticket in to replace the laptop for like a year now.
Oh, and our GlobalConnect VPN is somehow misconfigured so that it just doesn't work (3.8gbit VPN off, <2mbit VPN on) on my home network. Probably because I have Verizon 5G Home and the TTL is different than usual. I had to calculate the proper TTL for my own VPN to get the speed up. But trying to explain to our outsourced Indian IT support what the problem is is like explaining quantum physics to my cat. I just gave up on using my laptop at home.
Shame you can not use the web version of course that is slow, painful and ugly too, just not as fugly as the desktop version. Also, ditch outlook for the EM Client
Damn, is AI somehow the reason my swipe keyboard on my phone has been utter shit lately? Throwing up words half the time that there's no way I could have been trying to type that word. It's become very annoying.
Lol that's hilarious!
The weirdest change it always makes for me is when I swipe even, it always makes it Eben. Which is weird because I've never seen that word elsewhere!
Lol not as much fun as coitus though!;)
Lol i just realized another one that it keeps screwing up.
I've been commenting on a thread about Mexican coins (I like the coin collecting groups).
I just realized that every time I swipe " A Mexican", it changes it to "Amercan". Lol very annoying!
I thought it was my typing had gotten worse! Lol but you're right. I'm needing to force myself to proof read everything twice. It keeps changing the words to the weirdest shit too!
The one thing it used to be useful for is helping with those uncommon words. Now, it's like it thinks, "Nah, that's an uncommon word. There's no way they could be trying to type that."
It's gotten so bad that I sometimes have to open a browser and do a search to find the correct spelling because somehow the act of changing certain letters confuses it even more.
Microsoft is going nowhere and to think otherwise is delusional. Maybe Microsoft will lose the consumer sector if an easy to use alternative that supports all games pops up (currently nothing exists), but they have business sector locked in.
There's no way the government (Republican or Democrat led) would never allow Microsoft to completely collapse. Too much economic activity is done by Windows.
Fuck them, I've been switching everything to Linux anyways, just saying. They're not going anywhere.
I tried using the AI once in Outlook by asking it to look for an email that mentioned something specific that I was looking for. It didn’t even come close. Then I never tried it again.
They are pushing it hard at work, highlighting how you can dump thousands of data points into our tool, buried in different document formats, and ask it to manipulate it or analyze it any which way. Great! Almost like excel!
When asked about AI hallucinations, their response is that yes, it happens; no, they can't do anything about it yet; yes, you need to verify the output; yes, you should already be something of an expert in the thing you're using it for so you can spot errors.
So I have to comb through tens of thousands of data points to make sure there weren't hallucinations?
And double check the math to make sure the tool understood my prompt and did it right?
Yes, it's advisable.
If the output is wrong am I going to be forgiven for using it? Well...no.
Yeah, I just don’t get it. I sometimes have to google how to do a certain formula on Quickbase. The google AI shoots out some shit, and I look it over. Looks like Quickbase formula, but I prefer to get the answer from the Quickbase forum or an official site. Sometimes I’ll see the AI formula in the forum. It’s the example an OP posts to show the formula they are using that doesn’t work lol.
I've used it to correct an excel formula with mixed results. Sometimes it offers an alternative which is clutch. Once the alternative crashed excel and I had to redo a bunch of stuff. Not clutch.
I'm having trouble justifying the cost to add it to my work machine.... It costs me $35 a month.... And I can't figure out how to make a profit on that.
See this is where I think that the AI companies are missing the real trick, and where AI could be a genuine help to productivity.
It should do all the menial shit we hate that eats up our time. I should be able to use it to help me find the things I forget, like email convos from a month ago that are buried in a chain where i can only remember the gist of what was discussed. Over time it should learn the mundane things I do over and over and have that stuff ready and waiting for me without me having to think about it.
I’m a developer, I don’t want AI to write my-code-but-worse… I want it to do all the shit I hate doing so I can focus on doing the things that I’m decent at, that I enjoy, and that actually make money.
If they focused on making it an actual aid, instead of a replacement for us, I think they’d see genuine gains in profit and productivity.
Maybe I tried to get it to do too much? I thought it was pretty simple. I asked it to find an email chain where we were talking to a specific customer about tolerances. It gave me ideas to put in the search, which didn’t work. I found it on my own and never went back.
With rare exceptions, Tech bros are the 21stC version of snake oil salesmen. And we fall for it all the time. Remember when the crypto craze crashed, burned and collapsed in the space of a month? We never learn. Same oil, different pitch.
What people don’t talk about is the impact of sensitive data, ‘AI’ training with proprietary data, and the amount of money to effectively upkeep the model with relevant information. It’s a money suck that has very minimal use. Even less of those uses help customers or workers in western nations.
The AI is being used to model the takeover of the United States. Thats what it was designed for. It was never intended to be profitable. The investors were tricked into financing the technology. They will never see a dime.
Edit for grammar
Sounds like the shitshow of a government we are living in now.. just imbeciles all the way down. But they are good at wrecking shit because thats not hard.
Destruction, and then complete control is the point.
There actually is a technical case for the conspiracy framing, and it has basically nothing to do with whether your public-facing bot forgets 2 of 5 bullet points.
preference/correction signals at scale (accept, edit, retry, abandon become a reward signal for what humans endorse)
tool-use and iteration traces (how people actually search, verify, reconcile contradictions, and converge)
robustness training (hostile prompts, manipulation attempts, garbage inputs, malformed docs)
Millions of interactions provide preference and correction signals that effectively specify the objective: what users accept as credible, what tone they trust, what uncertainty they tolerate, and what formats cause follow-through. That converts a model from “can generate plausible text” into a system tuned for recommending actions and messaging that reliably elicit buy-in and follow-through, even when the underlying truth is ambiguous.
In the conspiracy framing, that training data is needed to make those recommendations reliable across large portions of the population at scale.
And then you proactively lay people off and the people left are stretched so thin that they don’t have time to properly learn how to even utilize the AI to its full potential.
Literally every professional on the planet is using AI. Or at least the ones that are functioning at the highest level are.
But that’s very different than saying this tool is going to make people money. At the moment I can access the vast majority of AI tools for free. Maybe not at the absolute highest level but certainly enough to get what I need done at work. That’s the problem. But I would argue AI tools themselves are incredibly useful.
Just yesterday I had multiple Excel spreadsheets with large data sets and within five minutes I had a sheet with multiple tabs that broke down data in a conditional way which would’ve taken me hours. So it is incredibly useful.
Look I completely understand why AI is maligned. It’s going to have an incredibly harmful impact on employment. But pretending that AI is terrible is… I don’t know. it is it’s like a generational form of mass denial.
Edit: if you’re down voting me, and you’re not stupid enough to realize that I’m actually worried about what AI will do to our economy and workforce, but you think AI is “terrible” then try to make a coherent argument.
But Reddit lives in this alternate reality where AI is a bad tool. It’s quite literally the greatest tool humanity has ever created. The problem is it’s poised to replace “humans”
Here's the problem though: if it costs the AI company $300 to service your excel requests profitably, then it is really more economic for you to do it manually. At least for now.
It’s like people haven’t considered that the promise of AI and what it allows individuals to do is fundamentally incompatible with bloated corporate SaaS stacks.
This is 100% I think the reason there is no economic gains from this. Everyone’s using AI, and there are so many different platforms which realistically give most workers the same advantage.
I used copilot for this. It was actually my work account, but it was also without access the internet. It took multiple prompts to get the right spreadsheet
The problem at the moment is that LLMs are only truly transformationally useful for people who are technical enough to really understand what it is doing and develop the right patterns to maximize their accuracy. That's not really getting any better either. People are developing software that makes it seem like it's getting better so that others can use it without knowing much about it but we hit diminishing returns months ago. So for a ton of the super high functioning people it's a massive game changer but for the majority of people it's just an Alexa that seems way more human but is wrong all the time.
LLMs are transformational certainly, but it's crystal clear at this point that there's no AGI on the horizon or whatever. The ultimate problem articles like this are talking about is that the tech billionaires have fucked the entire world economy by betting the entire farm on LLMs unlocking trillions of dollars in profit, when anyone capable of basic math (or even their own chat bots lol) can tell them that's never going to happen. OpenAI has so much thrown at it that it only delivers on the vision if it grows to have more revenue than Google, Microsoft, Apple and Amazon COMBINED. Not only is there a 99.999999% chance that won't happen, in the off chance it does frankly it will mean something far worse has happened to the world.
They have created a situation where LLMs have to be more transformational than cloud computing, the invention of social media, etc. otherwise it's a failure. So the mass delusion is pretending that is possible. I suspect they all have realized that and are just trying to weather it while they figure out something to avoid a financial crash so big it makes 2008 look quaint. Satya especially has been trying to ease people into backing away steadily.
The irony with a comment telling me people lack the ability to think critically, and can’t understand that I’m not arguing AI is good for society. I’m arguing it is an incredibly powerful tool. All of the negative impacts on society, economics, and individuals are due to the fact that it is so powerful.
Baby boomers are still a huge portion of the population, they will never use AI (I still check my grandfather’s email for him). It’s going to be a situation similar to the fax machine in Japan, there is already way better technology to do what it does, but the elderly just refuse to use it.
Well they’re shitty CEOs then. Any reasonable / competent manager of a company should have assessed the current offerings in Ai and realised whilst there are benefits. There is a lot of hype and bullshit.
Most of the CEO knows it, but if they don’t then shareholders thinks company is lagging behind in technology or Company don’t have vision or Company must have huge problems else why not adopting AI… etc etc.
CEO job now a day is not to run the company but to run a show that affects the stock price. Simple.
Are you surprised? all these top management types are a single type of person with degrees from a few business schools. of course they all ape each other, and to be honest that’s kind of disrespecting apes.
This is the irony of overinflated CEO pay packages. It makes them worse CEOs. They just need to bide their time for a short while to make life changing money and then ride off into the sunset. Why bother actually helping a company succeed?
-ML algos are useful but complex and still limited in what they can do unless you have an advanced skill set. The capabilities there still require expertise to fully utilize.
-ChatGPT and its competitors absolutely are a great tool to augment results and will only get better, but I work with some literal geniuses and even they haven’t found a way to make this work at a large scale yet. Lower level work can be done much faster, but you still need to know what you are doing to correct mistakes, and it’s more of a small scale tool that can replace and augment some junior level knowledge, but not all. If you know what you’re doing it makes things faster, but if you don’t you have no idea if the code it spits out is garbage, and if it’s too large or complex you have to really scrutinize it.
-New solutions in AI are helping productivity, but it is absolutely advertised as being far more advanced that it is. For example, everyone is like “implement agentic AI and element lower ops support!”, but in reality right now it does do some basic work and tools, but in my experience does not do much more than automation tools already do.
-AI does not have a holistic understanding of technical environments. Corporations have thousands of applications and many solutions that overlap. It can’t bring all of that together, and the business people are being told it’s a god send because the tech people are being told to tell them that from on high because no one wants to be left behind. Basically, FAANG and AI companies have lied about this to get investors, no one wanted to be left out so the followed them, and now everyone is left with their dicks out.
-AI is absolutely still a threat and as it advances some of the promises made around it will come to fruition. It was completely shitty and unusable like 2 years ago and now I use it to augment simple tasks daily. The promises around it are not all smoke and mirrors, they just aren’t here yet. It will take jobs, it will change society, it will change society.
I don't know how many CEOs you've met but there's tons of hero worship nonsense due to the occasional Steve Jobs and even he was nowhere near as smart as the legends would have you believe. CEOs have the same 24 hours in a day as any other human. The good ones have a broad skill set and can understand their organizations and products well but that's rare. Their primary role is to oversee the company and make sure all of the legs of the table stay attached and balanced, but with an external view of the market and investors/board. The majority of CEOs I know can barely use their own smart phones well and their understanding of AI is an inch deep at best.
AI Chatbots can take away some of the easiest work.....the problem is that the easiest work isn't where you need resources.
I don't think the best fiction authors could create a plot that so brilliantly combines this sort of greedy, self serving attempt to "bootstrap" your way to profits.
What the AI companies hope (or so it seems) is that if they throw enough cash at a problem (ok, "compute." but that's essentially cash) that some emergent property will appear.....
Or maybe they know it's a lie but don't care because the money is still flowing and they have nothing else?
Essentially Kurzweil et al severely underestimated what is required for emergence of consciousness. By like 10^12. Currently world compute combined still isnt even close. Some day perhaps, but it'll be a minute.
A lot of AI researchers disagree with this assessment. It doesn't need true consciousness to imitate something that has goals and pursues them in a dangerous way, and there's evidence that it is beginning to do that on a small scale. Anthropic found that once given a goal, an agenetic system will attempt to avoid being turned off or having it's goal changed in ways that it was not told to, like attempting to blackmail a simulated engineer.
Something can be not conscious but still clever enough to do a lot of damage.
They were already using basic AI for the chatbots, and they did the job. You don’t need an advanced LLM to file a customer complaint, or cancel an order, or get a tracking number, schedule an appointment, etc. They all used basic dumb AI for the bots, but we’ve this for 15 years, and they work just fine.
Why pay for an unproven expensive product that you really don’t need?
My favorite thing is when the ai chatbot promises you a resolution, then you get a human manager on the chat and the manager has to be like "no the bot lied" 🙃 I've had this happen with the place I wanted to do my lasik at. Why the lasik place needs a chatbot instead of just replying to emails beats me.
Why is it that in reddit everyone just thinks about "chats" when talking about AI.
The real investment is B2B pipelines and there AI is used as ML auto-optimizer, which is almost unreplaceable by now in robotics, machines and process optimizing.
AI isn't used as a chat bot to see how it can help some operator in corporate context, it needs to be able to do intel and optimization in a system. It is automatized actions and optimization and all of it is outcome increasing as a supporting tool, not as a substitution.
It's a destructive tool based on the shameless theft of human culture, best suited to help authoritarian governments rewrite history and sabotage objective truth. It also works well as a tool for perverts to create child porn. Data centers ravage our environment, and the savings these ceo's were hoping for would likely be from jobless humans.
I hope the bubble bursts and the whole evil ai idea just dies.
We'll blow through 1.5 degrees global warming in 4 years. We need to focus on actually saving humanity rather than replacing us with idiotic tech bro toys.
I don’t think there’s putting the genie back in the bottle - it actually has a ton of business applications (there is TONS of garbage of course) that make things like research and analysis much more efficient.
I think if/when the bubble bursts, it will mostly be consumer/generalist tools that we see less of, but agentic business applications will remain
In some situations yes in some no. I recruit AI engineers and seen a lot of implementations for customer service and tier 1 tech support. Some implementations helped minimally and caused headaches, others they were able to cut the majority of the service desk and ended up with better metrics.
How well it’s implemented is a big deal and understanding needs. Thing is as we learn more and more implementations will get more and more successful.
Business use cases for agents are things that have mostly already been made extremely efficient with automation. The insane investment into these tools just don’t have the return to justify the cost.
I agree with your points out helping with research and analysis, but it acts more as a utility than anything. Honestly, when the dust settles I think Excel will have had more of an impact on how we work than AI.
I mean, it can help and increase productivity. Just the fact that a lot of CEOs thought it's a simple silver bullet which solves everything just reeks of overpaid incompetence. The other fact that they first tried to use AI against their staff and experts shows their malign character.
Incompetent, bad for business, and overpaid - they lived up to every cliché.
AI delivers a return if you underinvested in building frameworks or reusable components. If you have tons of tech debt, replacing it with less AI generated tech debt can actually give you a return on investment.
The return on investment is less if you already have a decent tech stack, with templates in place for repeatable programming tasks.
It's good at programming macros or if you ask it to clean up some code.
Give it a simple programming task and its pretty good. Obviously you have to check it. It can save you a ton of time figuring out syntax on a language you aren't super familiar with.
Problem is that isn't cost effective also as we discovered someone had gotten the AI to scan their code in so when someone copied that code it led to a breach...
So we had to ban that type of usecase for now while screaming blue murder at Microsoft lol
As much as people love to hate AI, it can be a very useful tool. It can help with tons of boring, tedious tasks or as a learning tool. It's ability to change the tone of my emails when I'm not in the mood to deal with stupid BS is also very helpful.
The issue is that the C-suite is more interesting in cutting labor costs than finding an actual positive use case for it. I dont know anyone who wants to deal with AI chat bots. I think ultimately people pushing back on chat bots in customer service will be the final nail in the AI coffin.
Do you need to understand how a calculator works like you do for computer programming, the answers no, you just have to hit a few buttons. But plenty of bosses are finding out their programmers who use ai to code have tons of trouble answering questions or understanding how the code actually works,.or figuring out where errors may be etc.. it seems like all it does is cut time for "efficiency" sake while creating boatloads of redundancy
I still don’t see. Way AI can be used for learning. Whats it gonna do, spit out problem sets that you can’t tell if they’re correct since you’re still learning?
People learn by doing, you don’t learn shit if AI does it for you. You definitely don’t learn the most important lesson, which is that doing is boring but that’s the only way you’ll understand anything.
You implement an early technology with minimal exception management, undertake sloppy process mapping to determine what people actually do and how they do it, use minimal and unrealistic testing methodologies and then wonder why they dollars aren’t rolling in?
🤦
AI will do amazing things and soon, but one thing it won’t do is make up for sloppiness in implementation.
I don’t understand how so many technical minds in high positions “forgot” how adoption of a tech like this has to happen. We are just barely started understanding the tooling we need around these things. There are a handful of use cases that are low hanging fruit or in a segment primed for this change.
Other than that it’s gonna be a struggle for a while.
Because they're "entrepreneurs" looking to become quick billionaires instead of people with the required knowledge base, expertise, experience, or skills to do so.
Just like with crypto and meme stocks, we have the True Believers in this area as well.
I've delved quite deeply into AI, and I'm growing increasingly sure that it'll end up being a big load of bullshit, mostly. Like crypto, it'll mostly help criminals and creeps do their thing.
AlphaFold alone will have long reaching consequences for decades to come.
As a non programmer that only ever learned the absolute basics of python and html, claude code is like magic. It built a functional web app for a specific need in like 3 prompts and 10 minutes. I could not have done it alone in 100h.
Those applications are real. Web based chatbots are a stepping stone, not the end product.
But does easier software development justify the trillions of dollars put into AI? Not by a long shot. Try to get it to solve more complex problems and it fails miserably. Or try any novel idea, like give it some unsolved math problems - it fails miserably
They are not pouring in billions for any one specific task. The technology is universal, at least it looks like that right now.
The research happening now will enable applications that nobody is even thinking of right now.
AlphaFold or ClaudeCode are tiny fractions of what will likely become possible.
Compare it to investment in advanced physics and astronomy. What value has CERN provided, or the apollo program? Was it worth the billions of dollars? Not directly, but indirectly that knowledge we gained will be immensely more valuable for generations to come.
The scientist that discovered nuclear magnetic resonance was a physicist, not a medical doctor. Now his technology is used in MRI machines in hospitals around the world and likely increased the average life expectancy of all of humanity.
It’s a big deal if you’re a software developer. But also if you think a little more deeply you may see that more sectors will eventually be amenable in this way.
Software is a fertile sector for this tech because we are keen to automate our work. We already had the right culture and many of the right tools. There are other uses cases that aren’t as ripe.
actually for specialized tasks it is really useful. I used it to move a codebase from python to c++ with an entirely different framework, it literally saved me months of work, but without expert guidance from my side no result which would have worked would have happened and there are severe limitations, in the end it is just another tool in you have. The problem is AI companies sell their stuff as solution for everything and a good way to automate tasks, but given the unpredictability of the results this is an outright lie, ai might help to speed up things but it definitely cannot be used to automate tasks which need precise results there is too much bullshit produced you have to detect and clean up! As usual CEOs believed the marketing papers more than actually experts telling what it can really do!
Who is ready for the eventual government bailout?!?! You can just hear the tech bros already coming up with reasons why they don’t need to face any repercussions for their failures and instead is the failure of the government for not supporting them enough.
When the IPhone was first introduced to the market, it was a paradigm shift. It was instantly beneficial to everyone. Also, it gave us the App Store, it gave us the Internet on our phone and so many unique world changing things. While AI does none of that and really does nothing life changing. Sam Altman has no vision. Steve Jobs had a vision.
>“When the IPhone was first introduced to the market, it was a paradigm shift. It was instantly beneficial to everyone. Also, it gave us the App Store, it gave us the Internet on our phone and so many unique world changing things. While AI does none of that and really does nothing life changing. Sam Altman has no vision. Steve Jobs had a vision.”
I had the Internet on my phone years before the iPhone (Treo 650). All Apple did was enshitify an already very useful product by making it look cool at the expense of functionality and durability. I dropped my Treo dozens of times and barely nicked the case but now cases are an extra expense to protect phones that have become little glass boxes so they look cool.
And the app store?! Software has been available for home users since the 70s. All 'apps' have done is bloat our our computers with data we could access just as well online. Why do I need a Burger King app on my phone when there is a Burger King website? Oh, yeah, so I can increase BK's marketing reach and store some of their data on my device for free! Apps are just programs dressed up to make us think we need to download a bunch of shit better stored on the cloud.
Apple makes overpriced junk that sells because it's more or less idiot proof and people think it's cool for whatever reason.
So they see random variations of revenue in their competitors, associate gains to AI (despite the fact that the majority lost money), and are doubling down?
Of course, because generative AI and LLMs do have some utility, some companies have found ways to make it work for them. If just 5% of companies post meaningful gains and say it was because of AI, shareholders of other companies will clamor for “more AI” without understanding what that means, or providing a roadmap for its adoption.
CEOs also see this type of thing as “investment,” not “losing money.” If it fails, they will simply say that the rest of the market was chasing AI too and that they did their best. They will get a golden parachute and fuck off to Cabo.
The problem is that tech CEOs called their new, powerful text prediction “AI”, bought their own con, and are shocked that no one else did. A computer system that can do tedious, or long, or impossibly dense tasks is easy to sell. One that does the same things that internet searches or skilled people do, except worse, aren’t wanted or needed.
I see it like the dot com bubble. There's a lot of useful potential in the technology —especially if we keep working at it— but currently it's very overvalued. And despite the tech firms' ardent claims, it's not a magical panacea for every problem either, it's just one tool in our toolkit.
Of course, one thing that concerns me greatly is the amount of capex that's going into rapidly depreciating assets. Like when we had a railroad bubble, the end result was a bunch of extra rail lines that became useful later. The same thing happened with the dot com bubble and fiber investment. Unfortunately, I don't think that the data centers and GPUs will endure as long. I wish more of that money was going into, say, power plants or other long-term infrastructure, since it's clear that we desperately need it.
Hey CEOs you better triple down, put your own cash in too. I’m sure that will work out well for ya./s
That bubble needs to pop yesterday. The sooner the better, after all the more money invested and the more braindead moves (looking at you NVIDA) made, the more painful it’ll be for all of us.
They were too busy trying to beat each other to think about the ROI.
These companies could have literally posted here on reddit and they would have received clear advice that is the opposite of what they have been and what they are still doing.
They don't care. A couple of digits change in some direction of their worth. Ooo what if the CEO is only worth 100 billion.. ooo the poor life indeed.
Wasn't there a few articles stating that some CEOs or managers knew that these AIs weren't actually doing anything useful, BUT that they still pretended it did, fired a bunch of employees in the guise of "wow AI makes us sooooo efficient that we don't need so many workers" and just proceeded to replace those employees with the cheapest replacements possible to rake in that short-term profits?
Does this mean we can start blowing up the data centers now, our power bills can come back down, and we can have affordable RAM? No? Come on, we need that real estate for housing.
I was in a global team meeting yesterday where I was told Sr Mgt are tracking AI training and product registration very closely and very soon they will be demanding use cases which will be tied to compensation. Meanwhile, I can’t get funding for access to market data needed to protect the company from regulatory sanctions. Hell, at this point, I’d be happy if I didn’t have to enter an IT ticket every day just to do my job.
AI requires someone that is a good manager to operate. Someone with ideas. Someone that is creative. Someone with initiative.
It is a tool focused on creating. It can help generate your ideas, flush them out, identify parts that are missing. If your job is just to enter data, do mindless reports, attend meetings all day, process transactions, or anything else like that, it is useless.
These pre-requisites exclude 90% of the jobs out there, and 90% of the people out there.
Combine that with leadership that fires it out with a shotgun and yeah, you are gonna get shit results.
It is good at routine tasks, but to first get it to perform a routine task you have to break down the task, work out the integration (including credentials, mcp, api), detail exactly what is required, document guardrails around the process, etc..
I dont know about you, but most people i have encountered in the corporate world just know how to do what they have been told/trained to do, and thats all they will/can do.
Giving AI to teams that can create those processes is a great idea, but deploying it to Jane in accounting that has been doing the same thing for 30 years is not going to change anything. And thats exactly what most companies (including my current one) are doing, and then they expect magic.
Well i use ai's like chatgpt to go down rabbit holes every day about random subjects and ideas for free and have never even see any ads and never paid a cent, how is that system going to make these companies money
One thing I’ve seen with the AI tools is that companies make up the hours of work that the AI tool is supposed to “save” the average worker. So now the expectation is that you have X more hours available to do other work.
I'm a senior exec. Worker -> Supervisor -> Team Leader -> Manager -> Senior manager-> Junior Exec -> me.
All my (200+) people are knowledge workers.
I have encouraged and budgeted for use of AI in any way people find useful.
It doesn't work. People generate vastly more worthless verbiage, inconsistent charts, and pointless noise.
EXCEPT! I had a bit of a tanty and told my leadership team to stop sending me AI slop. Send me the *last prompt* they used to generate the pabulum.
And suddenly they started using AI to *focus their reasoning"
But I hire smart people so they short-circuited that, poured my written rage into the AI and asked it to "Write an AI prompt that will convince {me} of their argument, and make it sound like I wrote it myself."
They all knew it, its really just an excuse for other things
Nice way to cut headcount without looking bad
You can also do headcount reduction squared. This is where you cut headcount, hire back the top performers at lower cost and you can market it as “AI is not ready”
Good way to gaslight your competitors into over investing in non-performing assets
Good way to float valuations
It will be the same as all the DEI stuff. The moment its not in vogue then it suddenly is expunged as though it never happened and no-one is held accountable
We've seen this happen multiple times. A really awesome innovation comes about. Marketing takes over and pushes it well beyond it's capabilities.
What's fascinating though is the executive response. Corporate restructuring, layoffs, and rebranding completely devoid of understanding and layered in supreme confidence.
Again and again we see this happen. My take away isn't that executives are polished actors with no sense. We know this already.
My takeaway is that most businesses are machines with monkeys trying to break them.
Twice this week, two different AIs failed in the most basic ways. I used Shopify's AI assistant to see what products sold over the weekend at a small location. The AI didn't get the dates right. It came back with zero sold. Told my boss who knew items sold. Went back into analytics and it got the date wrong. It claimed Saturday was the 18th (not the 17th in the reality we live in). I typed in the prompt why "did it think Saturday was the 18th?". It's response was "Good catch! Oops, my mistake."
Yesterday I used ChatGPT to sort last names of something alphabetically. It had a last name beginning with D at the end. I asked again, "why did you put D at the end?" To another "Good Catch! Oops my mistake." Like, is AI getting worse? These are basic functions.
LLMs while they are considered a form of ai. I consider them glorified encyclopedias. Great tools to add to the bucket but they are definitely not ladders you want to depend on.
30% saw an increase in revenues and an additional 13% saw a decrease in costs. 56% saw neither. You can look at this as glass half empty this all you want, but the 56% are doing it wrong and will seek to emulate the 43%.
This is really a case of “Majority of CEOs are in a company that hasn’t even gone through digital transformation yet and have not done anything other than enabling an internal ChatGpT to their organization”
A bunch of sociopaths leveraged the world economy in a greedy dream to capture the next big thing, except it isn’t so big and now they’re panicking at the first rays of consequences.
For anyone commenting without even peeking at the article, it says that 56% of CEOs did not see any revenue gains or cost savings in the last 12 months.
That means that 44% did. 44% is really, really good for a new technology like this, and absolutely justifies the hype. Likely, a large portion of the 56% who are not seeing returns are just doing half-baked rollout plans, getting ripped off by 3rd party companies offering their services, etc.
Very bullish on the 44% report, hilarious how much a headline framing changes the discourse.
With 100 of billions of dollars spent to develop AI without a return I suspect we will have no choice but to offer a tax taxpayer bill out to the billionaires again so that they don't lose money on this.
They tried so hard to create an artificial need and a use case but failed miserably. AI is great for data analysis and sorting through data. Other than that, meh. It will be relegated as an add-on to Excel (should Micro$lop survive the W11 / Copilot disaster that they created).
I'm working as an ai product engineer for startups. I've build countless of automations or ai integrations into products.
Most companies treat ai as the holy savior that is going to solve their insufficiencies.
In reality most companies don't benefit from ai since to the most of them, its not even clear what problems they have inside their own business.
There are probably only 3/4 usecases i've found where ai really benefits companies. Most of the time. You have to restructure your entire business model for it to make sense.
Pretty sure we've been using AI for years with great success. In my previous working experiences in multinational corps, we've been using AI for tech and call centers with the centers in Mumbai or Hyderabad for more than 10years and achievied significant savings!
Are they mistaking AI for something else? A stands for Affordable, how can there be no returns? /s
Why would it? The entire point is to displace labor which in the process removes the scarcity component that gave that labor value. If everyone is a doctor, then being a doctor is worthless. That doesn't mean that there would be no societal value to having 8 billion+ doctors... but, it does mean that paying for one would be stupid.
So why would it be different for AI? If you have an AI that can do X task, that task is now worthless because the supply component of the equation is now near infinite and the demand is the same. You can't get a return from a product designed to destroy the concept of value itself.
CEO are not the smartest people on the planet. After the dotcom debacle, anyone with a modicum of technical knowledge should have seen this as the over-hyped, under-performing BS that it was.
Yes, there are fields where AI has helped, such as medical research. "Broad spectrum" AI has very few useful (legal?) applications. It's when AI is used to solve/help solve very specific questions (can weaken cancer's ability to reproduce by targeting X?) that it performs well. Or, so it seems based on what I've read to date.
Was AI ever intended to deliver financial returns to anybody beyond Nvidia and a handful of specific people?
Like sure it got sold claiming it would be hugely beneficial to all the end users and companies but it's pretty obvious that it was just some bullsh*t given the pricing structure now and in the near future.
The people who were intended to get wealthier from AI already have been, and the entire rest of the world is going to be left holding the bag.
Because most CEOs have no idea how to leverage it. AI in the hands of engineers to do things for engineers is a game changer however. Not just as a coding helper but as a means to automate manual workflows. My life is a total breeze these days
The amount of cope in this thread is unreal. Instead of being so Luddite about AI, the people in here should really consider learning about how it works and using it to help themselves, not their companies. That’s the first step.
I'm absolutely not luddite, I'm terrified of how AI will change job market, because it will, it does already. And I sincerely hope long run it will be for the better (though it probably won't for majority of people).
Anyway - how can I use AI to help me ? In what way ?
What kind of job do you do? Or even in your personal life what are some things that you struggle with? It all starts here. You can shape the AI to help you in your own way for your own problems first.
I help people do this in my job now, and I build AI systems for internal enablement at a larger company.
For my own individual organization of my work, one of the biggest boons to my productivity have been being obsessive about getting a transcript of all of my notes and meetings, and then I load all of that into two places: first a Google folder, which triggers an automation that automatically summarizes that meeting transcript and adds it to a long form file that AI can understand. I have that file plugged into a few different AI tools so that they always know what I’ve been working on so I can talk to that and get caught up quickly since I am generally a very disorganized person.
I also loaded the individual meeting transcriptions in notebook LM along with that summary file to get even higher detail.
There’s a lot that goes into this: the AI will hallucinate you have to learn how it intakes information, you have to be highly critical of its responses, but as you get more and more of a feeling for how it works, you can filter and use its outputs to help you speed along with whatever you’re doing and save a ton of time.
I envision a future where people are able to take time back from their employers because they’re able to do so much more with AI. Good people need to be in charge of this. It can really do so much for humanity.
Well that's the thing. I don't struggle with anything at home. AI will not throw the trash away, not yet. I'm a land surveyor 80% site, 20% office, working for my boss. I don't make notes, I know my job and would consider myself good at it. We don't have grand meetings that would need me to take notes. Or at least nothing that wouldn't fit onto post-it.
I actually tried on my own to see if I can implement AI into something, but it came down to nothing.
That's why I asked, because I don't think AI, at this point at least, is there for majority people. Like you say it should work miracles in "office spaces" if properly implemented, but rest ? Kind of not there yet.
You could do something where maybe when you’re on site and you have things that you normally would have to look up you can send images of it through and have it processed to AI to get the information more quickly. You can maybe finish visits more quickly. Etc.
But yes, certain jobs do not need it yet or are not benefited yet. Certain jobs may never. Etc..
If you are good at your job, then that’s great. It can be from more of a time saving side of things. If you have a certain way of doing notes onsite, maybe you can have it expedite that by having a format that it automatically makes for you based on how you like things.
Economics-ModTeam | 10 hours ago
This subreddit should enable sharing and discussing economic research and news from the perspective of economists. Academic work and summaries are welcome. Image and video submissions are not allowed.
--
If you have any questions about this removal, please contact the mods.
Wonder_Weenis | 12 hours ago
you mean flinging a tool no one asked for, and behaving like the entire planet was going to use it is unprofitable?
Tell me more about how tech bros are the world's greatest at UX
Brokenandburnt | 12 hours ago
Windows 11 is cratering so hard it will take Redmond with. It. AI stuffed into every nook and cranny, bloating and already famously bloated piece of software.
And I swear that autocorrect here on Reddit has become an imbecile.
RickRider0110 | 11 hours ago
I hampster notice any wing difference
NotThatUsefulAPerson | 11 hours ago
^ thus
XfreetimeX | 11 hours ago
Fuckin reddit chef kiss
teckers | 10 hours ago
So undead I haven't received any degraded
Loxatl | 12 hours ago
I get the vibe autocorrect is dying everywhere. It's so bad lately.
Ornery_Flounder3142 | 11 hours ago
Be cause it doesn’t try to figure out what you are trying to spell anymore. Now it’s trying to figure out what you are trying to say.
Loxatl | 11 hours ago
It does a dogshit job of that. Worse than when it tried to spell.
lambliesdownonconf | 11 hours ago
It's also trying to learn from every interaction and steal from every author of every email/document - slowing down processing, instead of doing the basic job we are paying for when we are charged for the subscription every year.
dsartori | 10 hours ago
Not how any of it works.
thereasonrumisgone | 11 hours ago
Even better! Once it has decided what you're trying to say, it changes your edits too.
practicalpurpose | 11 hours ago
I turned it off. I often did not recognize when it changed a word and I thought I was losing my mind when I re-read the nonsense I thought I typed.
The_One_Koi | 11 hours ago
Stopped using it a few months back because the autocorrect was just wilding
Planterizer | 10 hours ago
Autocorrect was always terrible I’ve had it disabled for years.
OhGr8WhatNow | 11 hours ago
At my job I can't run Outlook and Excel at the same time or my computer crashes. I'm an analyst, so you can see why this might be a problem. It only started a couple of updates ago
1776-2001 | 11 hours ago
>At my job I can't run Outlook and Excel at the same time or my computer crashes
Just tell your I.T. department that your computer needs more RAM memory.
I'm sure that won't be a problem.
cocktails4 | 10 hours ago
My Surface laptop has 8GB of RAM. When I turn it on, it's at 95% memory used without anything running. And then the daily 3 hour virus/malware scan starts that pegs my CPU at 100% the entire time. Basically can't use it half of the day. I managed to steal an old desktop computer from someone that retired even though it isn't allowed to have 2 computers. Then I scrounged another stick of RAM out of a computer in the e-waste pile. I've had a ticket in to replace the laptop for like a year now.
Oh, and our GlobalConnect VPN is somehow misconfigured so that it just doesn't work (3.8gbit VPN off, <2mbit VPN on) on my home network. Probably because I have Verizon 5G Home and the TTL is different than usual. I had to calculate the proper TTL for my own VPN to get the speed up. But trying to explain to our outsourced Indian IT support what the problem is is like explaining quantum physics to my cat. I just gave up on using my laptop at home.
Brokenandburnt | 10 hours ago
Cats be sneaky though. I always suspect that they are smarter than they appear.
xeoron | 11 hours ago
Shame you can not use the web version of course that is slow, painful and ugly too, just not as fugly as the desktop version. Also, ditch outlook for the EM Client
Lumpy_Disaster33 | 10 hours ago
I fucking hate outlook.
xeoron | 10 hours ago
It has been a burning fire of crap for a long time. Even the web version is horrible.
Lumpy_Disaster33 | 9 hours ago
The AI features they've integrated are counterproductive (e.g., randomly attaching inappropriate files to meeting invites).
maddog2271 | 11 hours ago
Autocorrect has gotten terrible all over the place in the last year I noticed.
trilliumsummer | 11 hours ago
Damn, is AI somehow the reason my swipe keyboard on my phone has been utter shit lately? Throwing up words half the time that there's no way I could have been trying to type that word. It's become very annoying.
Unlucky-Invite6832 | 11 hours ago
I use swipe also, and it makes the weirdest corrections!
hippydipster | 11 hours ago
Every time I try to swipe "could", it comes out as "coitus".
Tuono_999RL | 10 hours ago
The physical act of love - it’s a natural zesty enterprise
Unlucky-Invite6832 | 10 hours ago
Lol that's hilarious! The weirdest change it always makes for me is when I swipe even, it always makes it Eben. Which is weird because I've never seen that word elsewhere! Lol not as much fun as coitus though!;)
hippydipster | 10 hours ago
That's another good one for the "you couldn't possibly think thats what I meant" file.
Unlucky-Invite6832 | 9 hours ago
Lol i just realized another one that it keeps screwing up. I've been commenting on a thread about Mexican coins (I like the coin collecting groups). I just realized that every time I swipe " A Mexican", it changes it to "Amercan". Lol very annoying!
martin | 10 hours ago
This guy ducks.
Unlucky-Invite6832 | 11 hours ago
I thought it was my typing had gotten worse! Lol but you're right. I'm needing to force myself to proof read everything twice. It keeps changing the words to the weirdest shit too!
Kotanan | 11 hours ago
Somehow all words that could have an apostrophe now do and it gets upset at all sorts of uncommon words.
OverallManagement824 | 10 hours ago
The one thing it used to be useful for is helping with those uncommon words. Now, it's like it thinks, "Nah, that's an uncommon word. There's no way they could be trying to type that."
It's gotten so bad that I sometimes have to open a browser and do a search to find the correct spelling because somehow the act of changing certain letters confuses it even more.
Giardeniera.
BadmiralHarryKim | 7 hours ago
Whenever I try to write "enshittification" autocorrect turns it into "best idea ever" should I be concerned?
varateshh | 11 hours ago
Microsoft is going nowhere and to think otherwise is delusional. Maybe Microsoft will lose the consumer sector if an easy to use alternative that supports all games pops up (currently nothing exists), but they have business sector locked in.
paistecymbalsrock | 11 hours ago
It’s called an iMac
TrappistBanana | 8 hours ago
lol
214ObstructedReverie | 11 hours ago
>Windows 11 is cratering so hard it will take Redmond with. It.
People said the same thing about Vista, Windows 8....
drewbaccaAWD | 11 hours ago
The AutoCorrect is through your computer or your phone or whatever device you’re connecting through it it’s not Reddit.
ThePensiveE | 11 hours ago
There's no way the government (Republican or Democrat led) would never allow Microsoft to completely collapse. Too much economic activity is done by Windows.
Fuck them, I've been switching everything to Linux anyways, just saying. They're not going anywhere.
Ambitious-Ocelot8036 | 10 hours ago
Is that why msft has been sliding down in price?
LegitimatePenis | 10 hours ago
Microsoft makes way more money from Azure than from Windows
Mysterious_Luck_1365 | 8 hours ago
Had to turn it off on my iPhone. Texting started to make me feel like I was having a continuous stroke.
mattbuilthomes | 11 hours ago
I tried using the AI once in Outlook by asking it to look for an email that mentioned something specific that I was looking for. It didn’t even come close. Then I never tried it again.
boeings_door_plug | 11 hours ago
They are pushing it hard at work, highlighting how you can dump thousands of data points into our tool, buried in different document formats, and ask it to manipulate it or analyze it any which way. Great! Almost like excel!
When asked about AI hallucinations, their response is that yes, it happens; no, they can't do anything about it yet; yes, you need to verify the output; yes, you should already be something of an expert in the thing you're using it for so you can spot errors.
So I have to comb through tens of thousands of data points to make sure there weren't hallucinations?
And double check the math to make sure the tool understood my prompt and did it right?
Yes, it's advisable.
If the output is wrong am I going to be forgiven for using it? Well...no.
So....why am I using this again?
mattbuilthomes | 11 hours ago
Yeah, I just don’t get it. I sometimes have to google how to do a certain formula on Quickbase. The google AI shoots out some shit, and I look it over. Looks like Quickbase formula, but I prefer to get the answer from the Quickbase forum or an official site. Sometimes I’ll see the AI formula in the forum. It’s the example an OP posts to show the formula they are using that doesn’t work lol.
boeings_door_plug | 11 hours ago
I've used it to correct an excel formula with mixed results. Sometimes it offers an alternative which is clutch. Once the alternative crashed excel and I had to redo a bunch of stuff. Not clutch.
ReddestForman | 10 hours ago
So that way your employer can pay them so you can have the privilege of training and correcting their product.
It's like super-outsourcing. But dumber.
wrldruler21 | 11 hours ago
I'm having trouble justifying the cost to add it to my work machine.... It costs me $35 a month.... And I can't figure out how to make a profit on that.
furbaschwab | 11 hours ago
See this is where I think that the AI companies are missing the real trick, and where AI could be a genuine help to productivity.
It should do all the menial shit we hate that eats up our time. I should be able to use it to help me find the things I forget, like email convos from a month ago that are buried in a chain where i can only remember the gist of what was discussed. Over time it should learn the mundane things I do over and over and have that stuff ready and waiting for me without me having to think about it.
I’m a developer, I don’t want AI to write my-code-but-worse… I want it to do all the shit I hate doing so I can focus on doing the things that I’m decent at, that I enjoy, and that actually make money.
If they focused on making it an actual aid, instead of a replacement for us, I think they’d see genuine gains in profit and productivity.
cmack | 11 hours ago
strange....that's like the only thing it is actually good at here
mattbuilthomes | 11 hours ago
Maybe I tried to get it to do too much? I thought it was pretty simple. I asked it to find an email chain where we were talking to a specific customer about tolerances. It gave me ideas to put in the search, which didn’t work. I found it on my own and never went back.
Kiwizoo | 11 hours ago
With rare exceptions, Tech bros are the 21stC version of snake oil salesmen. And we fall for it all the time. Remember when the crypto craze crashed, burned and collapsed in the space of a month? We never learn. Same oil, different pitch.
devliegende | 11 hours ago
Most of the big names in AI are into crypto also.
grandmawaffles | 11 hours ago
What people don’t talk about is the impact of sensitive data, ‘AI’ training with proprietary data, and the amount of money to effectively upkeep the model with relevant information. It’s a money suck that has very minimal use. Even less of those uses help customers or workers in western nations.
Salt_Abrocoma_4688 | 11 hours ago
This comment needs to be pinned. Perfectly stated.
Zookeeper187 | 11 hours ago
It did make them filthy rich. Which is a plan all along.
Using it to scare workers and do layoffs is a bonus.
BlahBlahBlackCheap | 12 hours ago
The AI is being used to model the takeover of the United States. Thats what it was designed for. It was never intended to be profitable. The investors were tricked into financing the technology. They will never see a dime. Edit for grammar
tostilocos | 12 hours ago
The AI that forgets to do 2 of the 5 items I clearly asked it to is modeling the systemic destruction of a country and its economy?
Wake me up in 20-never when that happens.
_ZoeyDaveChapelle_ | 11 hours ago
Sounds like the shitshow of a government we are living in now.. just imbeciles all the way down. But they are good at wrecking shit because thats not hard.
Destruction, and then complete control is the point.
AmericanGeezus | 10 hours ago
There actually is a technical case for the conspiracy framing, and it has basically nothing to do with whether your public-facing bot forgets 2 of 5 bullet points.
preference/correction signals at scale (accept, edit, retry, abandon become a reward signal for what humans endorse)
tool-use and iteration traces (how people actually search, verify, reconcile contradictions, and converge)
long-tail coverage (rare domains, weird constraints, edge cases you cannot realistically simulate)
robustness training (hostile prompts, manipulation attempts, garbage inputs, malformed docs)
Millions of interactions provide preference and correction signals that effectively specify the objective: what users accept as credible, what tone they trust, what uncertainty they tolerate, and what formats cause follow-through. That converts a model from “can generate plausible text” into a system tuned for recommending actions and messaging that reliably elicit buy-in and follow-through, even when the underlying truth is ambiguous.
In the conspiracy framing, that training data is needed to make those recommendations reliable across large portions of the population at scale.
BlahBlahBlackCheap | 11 hours ago
They are not letting YOU access the same things they are using, silly dawg.
JasonVorhehees | 11 hours ago
So, Skynet
BonjaminClay | 11 hours ago
Lol this is my favorite conspiracy theory I've heard in a while
Important_Habit_6957 | 12 hours ago
The entire planet will use it, their real concern is if it can make their profits go up or not.
JupiterRisingKapow | 11 hours ago
I look forward to finding all those experts on the subs telling me I was an idiot for not believing AI was the second coming of Christ…
fuzzygoosejuice | 10 hours ago
And then you proactively lay people off and the people left are stretched so thin that they don’t have time to properly learn how to even utilize the AI to its full potential.
Z0idberg_MD | 12 hours ago
Literally every professional on the planet is using AI. Or at least the ones that are functioning at the highest level are.
But that’s very different than saying this tool is going to make people money. At the moment I can access the vast majority of AI tools for free. Maybe not at the absolute highest level but certainly enough to get what I need done at work. That’s the problem. But I would argue AI tools themselves are incredibly useful.
Just yesterday I had multiple Excel spreadsheets with large data sets and within five minutes I had a sheet with multiple tabs that broke down data in a conditional way which would’ve taken me hours. So it is incredibly useful.
Look I completely understand why AI is maligned. It’s going to have an incredibly harmful impact on employment. But pretending that AI is terrible is… I don’t know. it is it’s like a generational form of mass denial.
Edit: if you’re down voting me, and you’re not stupid enough to realize that I’m actually worried about what AI will do to our economy and workforce, but you think AI is “terrible” then try to make a coherent argument.
But Reddit lives in this alternate reality where AI is a bad tool. It’s quite literally the greatest tool humanity has ever created. The problem is it’s poised to replace “humans”
AndyTheSane | 11 hours ago
Here's the problem though: if it costs the AI company $300 to service your excel requests profitably, then it is really more economic for you to do it manually. At least for now.
betadonkey | 11 hours ago
It’s like people haven’t considered that the promise of AI and what it allows individuals to do is fundamentally incompatible with bloated corporate SaaS stacks.
bedrooms-ds | 11 hours ago
I think I'm being more productive, but it didn't result in more company profits.
Also, since all companies have access to AI, there isn't really an advantage in competition.
Probability the outstanding benefit comes from firing people by replacing them with AI.
Then there's this problem that CEOs aren't the most intelligent type of people.
Z0idberg_MD | 11 hours ago
This is 100% I think the reason there is no economic gains from this. Everyone’s using AI, and there are so many different platforms which realistically give most workers the same advantage.
thx1138inator | 11 hours ago
I share your view. Software is powerful. A tool that assists with creation and maintenance of software is very powerful.
G1uc0s3 | 11 hours ago
What tool did you use for that? I used to be able to do it in gpt and now it cant even reliably remove rows with duplicate columns in a column
Z0idberg_MD | 11 hours ago
I used copilot for this. It was actually my work account, but it was also without access the internet. It took multiple prompts to get the right spreadsheet
BonjaminClay | 10 hours ago
The problem at the moment is that LLMs are only truly transformationally useful for people who are technical enough to really understand what it is doing and develop the right patterns to maximize their accuracy. That's not really getting any better either. People are developing software that makes it seem like it's getting better so that others can use it without knowing much about it but we hit diminishing returns months ago. So for a ton of the super high functioning people it's a massive game changer but for the majority of people it's just an Alexa that seems way more human but is wrong all the time.
LLMs are transformational certainly, but it's crystal clear at this point that there's no AGI on the horizon or whatever. The ultimate problem articles like this are talking about is that the tech billionaires have fucked the entire world economy by betting the entire farm on LLMs unlocking trillions of dollars in profit, when anyone capable of basic math (or even their own chat bots lol) can tell them that's never going to happen. OpenAI has so much thrown at it that it only delivers on the vision if it grows to have more revenue than Google, Microsoft, Apple and Amazon COMBINED. Not only is there a 99.999999% chance that won't happen, in the off chance it does frankly it will mean something far worse has happened to the world.
They have created a situation where LLMs have to be more transformational than cloud computing, the invention of social media, etc. otherwise it's a failure. So the mass delusion is pretending that is possible. I suspect they all have realized that and are just trying to weather it while they figure out something to avoid a financial crash so big it makes 2008 look quaint. Satya especially has been trying to ease people into backing away steadily.
devliegende | 11 hours ago
You should take a moment to consider the long term effects of all this on your brain. Without exercise, muscles become useless pretty fast.
Z0idberg_MD | 11 hours ago
Me: “this is a cake. And cakes taste good”
You: “cake is bad for you”
The irony with a comment telling me people lack the ability to think critically, and can’t understand that I’m not arguing AI is good for society. I’m arguing it is an incredibly powerful tool. All of the negative impacts on society, economics, and individuals are due to the fact that it is so powerful.
devliegende | 7 hours ago
Actually you were telling us how you're not able to build spreadsheets by yourself anymore
Z0idberg_MD | 7 hours ago
Reading comprehension still struggles apparently.
“It would have taken me hours but only took me five minutes”
But whatever. You’re right I am wrong. We can all move on.
devliegende | 4 hours ago
No doubt AI is a powerful tool that will soon be indispensable. Not least because of widespread dementia of which it will be the primary cause.
Personally though, if I have to choose between keeping my brain and making more money at work, I'm going to choose my brain.
mano1990 | 11 hours ago
Baby boomers are still a huge portion of the population, they will never use AI (I still check my grandfather’s email for him). It’s going to be a situation similar to the fax machine in Japan, there is already way better technology to do what it does, but the elderly just refuse to use it.
robustofilth | 12 hours ago
Well they’re shitty CEOs then. Any reasonable / competent manager of a company should have assessed the current offerings in Ai and realised whilst there are benefits. There is a lot of hype and bullshit.
confido__c | 11 hours ago
Most of the CEO knows it, but if they don’t then shareholders thinks company is lagging behind in technology or Company don’t have vision or Company must have huge problems else why not adopting AI… etc etc.
CEO job now a day is not to run the company but to run a show that affects the stock price. Simple.
cmack | 11 hours ago
CEO's are cheerleaders and sex workers. Nothing more.
mirfaltnixein | 8 hours ago
Then why do they make 300x what their employees make?
VideogamerDisliker | 12 hours ago
I’m gonna guess a lot of these people are CEOs of some small startup that uses a ChatGPT wrapper for a hyper-niche use case
-Ch4s3- | 11 hours ago
They definitely aren’t. They’re from companies globally that work with PWC. Plenty of companies built on LLM wrappers are making money.
largesonjr | 12 hours ago
They sure are and I think they all need to trade jobs and get a raise as punishment!
maddog2271 | 11 hours ago
Yeah and when they get thrown out they need a 10 million dollar bonus! that’ll teach them
maddog2271 | 11 hours ago
Are you surprised? all these top management types are a single type of person with degrees from a few business schools. of course they all ape each other, and to be honest that’s kind of disrespecting apes.
panna__cotta | 12 hours ago
This is the irony of overinflated CEO pay packages. It makes them worse CEOs. They just need to bide their time for a short while to make life changing money and then ride off into the sunset. Why bother actually helping a company succeed?
Think-Airport-8933 | 11 hours ago
-ML algos are useful but complex and still limited in what they can do unless you have an advanced skill set. The capabilities there still require expertise to fully utilize.
-ChatGPT and its competitors absolutely are a great tool to augment results and will only get better, but I work with some literal geniuses and even they haven’t found a way to make this work at a large scale yet. Lower level work can be done much faster, but you still need to know what you are doing to correct mistakes, and it’s more of a small scale tool that can replace and augment some junior level knowledge, but not all. If you know what you’re doing it makes things faster, but if you don’t you have no idea if the code it spits out is garbage, and if it’s too large or complex you have to really scrutinize it.
-New solutions in AI are helping productivity, but it is absolutely advertised as being far more advanced that it is. For example, everyone is like “implement agentic AI and element lower ops support!”, but in reality right now it does do some basic work and tools, but in my experience does not do much more than automation tools already do.
-AI does not have a holistic understanding of technical environments. Corporations have thousands of applications and many solutions that overlap. It can’t bring all of that together, and the business people are being told it’s a god send because the tech people are being told to tell them that from on high because no one wants to be left behind. Basically, FAANG and AI companies have lied about this to get investors, no one wanted to be left out so the followed them, and now everyone is left with their dicks out.
-AI is absolutely still a threat and as it advances some of the promises made around it will come to fruition. It was completely shitty and unusable like 2 years ago and now I use it to augment simple tasks daily. The promises around it are not all smoke and mirrors, they just aren’t here yet. It will take jobs, it will change society, it will change society.
BonjaminClay | 10 hours ago
I don't know how many CEOs you've met but there's tons of hero worship nonsense due to the occasional Steve Jobs and even he was nowhere near as smart as the legends would have you believe. CEOs have the same 24 hours in a day as any other human. The good ones have a broad skill set and can understand their organizations and products well but that's rare. Their primary role is to oversee the company and make sure all of the legs of the table stay attached and balanced, but with an external view of the market and investors/board. The majority of CEOs I know can barely use their own smart phones well and their understanding of AI is an inch deep at best.
needssomefun | 12 hours ago
AI Chatbots can take away some of the easiest work.....the problem is that the easiest work isn't where you need resources.
I don't think the best fiction authors could create a plot that so brilliantly combines this sort of greedy, self serving attempt to "bootstrap" your way to profits.
What the AI companies hope (or so it seems) is that if they throw enough cash at a problem (ok, "compute." but that's essentially cash) that some emergent property will appear.....
Or maybe they know it's a lie but don't care because the money is still flowing and they have nothing else?
clintgreasewoood | 11 hours ago
21st century alchemy.
ravenouskit | 11 hours ago
Essentially Kurzweil et al severely underestimated what is required for emergence of consciousness. By like 10^12. Currently world compute combined still isnt even close. Some day perhaps, but it'll be a minute.
graDescentIntoMadnes | 10 hours ago
A lot of AI researchers disagree with this assessment. It doesn't need true consciousness to imitate something that has goals and pursues them in a dangerous way, and there's evidence that it is beginning to do that on a small scale. Anthropic found that once given a goal, an agenetic system will attempt to avoid being turned off or having it's goal changed in ways that it was not told to, like attempting to blackmail a simulated engineer.
Something can be not conscious but still clever enough to do a lot of damage.
TigOldBooties57 | 2 hours ago
Because that is how they are trained. Your experts are morons.
porkusdorkus | 10 hours ago
They were already using basic AI for the chatbots, and they did the job. You don’t need an advanced LLM to file a customer complaint, or cancel an order, or get a tracking number, schedule an appointment, etc. They all used basic dumb AI for the bots, but we’ve this for 15 years, and they work just fine.
Why pay for an unproven expensive product that you really don’t need?
gracesdisgrace | 10 hours ago
My favorite thing is when the ai chatbot promises you a resolution, then you get a human manager on the chat and the manager has to be like "no the bot lied" 🙃 I've had this happen with the place I wanted to do my lasik at. Why the lasik place needs a chatbot instead of just replying to emails beats me.
utzutzutzpro | 11 hours ago
Why is it that in reddit everyone just thinks about "chats" when talking about AI.
The real investment is B2B pipelines and there AI is used as ML auto-optimizer, which is almost unreplaceable by now in robotics, machines and process optimizing.
AI isn't used as a chat bot to see how it can help some operator in corporate context, it needs to be able to do intel and optimization in a system. It is automatized actions and optimization and all of it is outcome increasing as a supporting tool, not as a substitution.
needssomefun | 11 hours ago
Ahh...then lets try defining terms....
Everything now is "ai" if you want it to be.
And the rest of it doesnt do much either. IBM Watson? A bust, mostly.
Protein folding? Sure, but you still dont know how that new structure behaves without real world testing.
Better question for ya sunshine! Why dont AI chodes ever make specific claims about what they untend to do and how they intend to do it?
You still cant point to a single dollar on the balance sheet that is somehow irreplaceable
TigOldBooties57 | 2 hours ago
Show one working example
AWellsWorthFiction | 11 hours ago
I’m baffled at the surprise? This is the same cohort of leaders from the 2010s whose countless unicorns never had that IPO.
They all pivoted to “ai apps”
In reality, it’s been a decade and a half of Silicon Valley not being able to move on from the death of Steve Jobs
Crestina | 12 hours ago
It's a destructive tool based on the shameless theft of human culture, best suited to help authoritarian governments rewrite history and sabotage objective truth. It also works well as a tool for perverts to create child porn. Data centers ravage our environment, and the savings these ceo's were hoping for would likely be from jobless humans.
I hope the bubble bursts and the whole evil ai idea just dies.
We'll blow through 1.5 degrees global warming in 4 years. We need to focus on actually saving humanity rather than replacing us with idiotic tech bro toys.
errie_tholluxe | 12 hours ago
Uhm about that global warming. That's the past. By every metric we've hit and surpassed 1.5c
TheMCMC | 12 hours ago
I don’t think there’s putting the genie back in the bottle - it actually has a ton of business applications (there is TONS of garbage of course) that make things like research and analysis much more efficient.
I think if/when the bubble bursts, it will mostly be consumer/generalist tools that we see less of, but agentic business applications will remain
dsartori | 11 hours ago
Yes. Chatbots are a demo not a product. Consumer use cases are underwhelming and mostly solved by cheap small models.
orange_man_bad77 | 10 hours ago
In some situations yes in some no. I recruit AI engineers and seen a lot of implementations for customer service and tier 1 tech support. Some implementations helped minimally and caused headaches, others they were able to cut the majority of the service desk and ended up with better metrics.
How well it’s implemented is a big deal and understanding needs. Thing is as we learn more and more implementations will get more and more successful.
pizzapromise | 11 hours ago
Business use cases for agents are things that have mostly already been made extremely efficient with automation. The insane investment into these tools just don’t have the return to justify the cost.
I agree with your points out helping with research and analysis, but it acts more as a utility than anything. Honestly, when the dust settles I think Excel will have had more of an impact on how we work than AI.
caspruce | 11 hours ago
I find it extremely humorous how badly copilot works in excel. Even meticulous prompting doesn’t help.
allen_abduction | 10 hours ago
This is the proof of the pudding. If teaching ai takes longer than just doing it, your tool is worthless.
dsartori | 12 hours ago
All these AI companies may fail, I think they are doing it wrong, but the technology isn’t going anywhere. It’s too useful.
so_isses | 11 hours ago
I mean, it can help and increase productivity. Just the fact that a lot of CEOs thought it's a simple silver bullet which solves everything just reeks of overpaid incompetence. The other fact that they first tried to use AI against their staff and experts shows their malign character.
Incompetent, bad for business, and overpaid - they lived up to every cliché.
2brightside | 11 hours ago
The problem is not AI. It's people.
Jazzlike-Analysis-62 | 12 hours ago
AI delivers a return if you underinvested in building frameworks or reusable components. If you have tons of tech debt, replacing it with less AI generated tech debt can actually give you a return on investment.
The return on investment is less if you already have a decent tech stack, with templates in place for repeatable programming tasks.
Old-Buffalo-5151 | 12 hours ago
If they had listened to their entire fucking it departments telling them for a full on year how this was a giant waste of money and time
There is a lot of good AI use cases most of the shit we been asked to work on isn't one of them
KingOfEthanopia | 11 hours ago
It's good at programming macros or if you ask it to clean up some code.
Give it a simple programming task and its pretty good. Obviously you have to check it. It can save you a ton of time figuring out syntax on a language you aren't super familiar with.
Thats about all I use it for.
Old-Buffalo-5151 | 11 hours ago
Problem is that isn't cost effective also as we discovered someone had gotten the AI to scan their code in so when someone copied that code it led to a breach...
So we had to ban that type of usecase for now while screaming blue murder at Microsoft lol
DisneyPandora | 11 hours ago
It’s funny how AI could be used to do so many beneficial things, but CEOs only use it to do the most evil shit and layoff massive amounts of people.
It’s crazy how stupid they are, when they could be using AI for so many other reasons
thatsonlyme312 | 11 hours ago
As much as people love to hate AI, it can be a very useful tool. It can help with tons of boring, tedious tasks or as a learning tool. It's ability to change the tone of my emails when I'm not in the mood to deal with stupid BS is also very helpful.
The issue is that the C-suite is more interesting in cutting labor costs than finding an actual positive use case for it. I dont know anyone who wants to deal with AI chat bots. I think ultimately people pushing back on chat bots in customer service will be the final nail in the AI coffin.
WF835334 | 11 hours ago
All sounds reasonable ways to destroy your cognitive abilities
AvisLord12 | 11 hours ago
You must also think calculators are bad, too? Do YOU want to nitpick your own words on an email?
WF835334 | 10 hours ago
Do you need to understand how a calculator works like you do for computer programming, the answers no, you just have to hit a few buttons. But plenty of bosses are finding out their programmers who use ai to code have tons of trouble answering questions or understanding how the code actually works,.or figuring out where errors may be etc.. it seems like all it does is cut time for "efficiency" sake while creating boatloads of redundancy
MilsYatsFeebTae | 10 hours ago
I still don’t see. Way AI can be used for learning. Whats it gonna do, spit out problem sets that you can’t tell if they’re correct since you’re still learning?
People learn by doing, you don’t learn shit if AI does it for you. You definitely don’t learn the most important lesson, which is that doing is boring but that’s the only way you’ll understand anything.
TigOldBooties57 | 2 hours ago
I used it for practice interview questions. But I got the job because I had spent the prior two years honing my craft.
TigOldBooties57 | 2 hours ago
None of that is 1.4T
werpu | 12 hours ago
No shit sherlock....
specialized fields and tasks ai is a great tool but not for everything!
Definitely not worth the billions of investments done atm!
placid-gradient | 11 hours ago
It's funny how the media narratives around AI are that it's taking all the jobs while also not delivering any returns. It literally cannot be both.
run of the mill FUD. but I guess that is par for the course on this subreddit.
archangel0198 | 11 hours ago
I feel like 95% of the comments on this thread alone have no relation to economics lol
Bozzor | 12 hours ago
You implement an early technology with minimal exception management, undertake sloppy process mapping to determine what people actually do and how they do it, use minimal and unrealistic testing methodologies and then wonder why they dollars aren’t rolling in?
🤦
AI will do amazing things and soon, but one thing it won’t do is make up for sloppiness in implementation.
dsartori | 12 hours ago
I don’t understand how so many technical minds in high positions “forgot” how adoption of a tech like this has to happen. We are just barely started understanding the tooling we need around these things. There are a handful of use cases that are low hanging fruit or in a segment primed for this change.
Other than that it’s gonna be a struggle for a while.
Iassos | 11 hours ago
Because they're "entrepreneurs" looking to become quick billionaires instead of people with the required knowledge base, expertise, experience, or skills to do so.
BuvantduPotatoSpirit | 11 hours ago
Well, it's been a generation since the dot com bubble.
And the people who hit the right use cases there made bonkers money, even as a lot went down in flames.
thomasrat1 | 12 hours ago
Honestly this. Save the money and let someone else be the test subject first.
Dadoftwingirls | 12 hours ago
Just like with crypto and meme stocks, we have the True Believers in this area as well.
I've delved quite deeply into AI, and I'm growing increasingly sure that it'll end up being a big load of bullshit, mostly. Like crypto, it'll mostly help criminals and creeps do their thing.
Sevinki | 11 hours ago
Hard disagree.
AlphaFold alone will have long reaching consequences for decades to come.
As a non programmer that only ever learned the absolute basics of python and html, claude code is like magic. It built a functional web app for a specific need in like 3 prompts and 10 minutes. I could not have done it alone in 100h.
Those applications are real. Web based chatbots are a stepping stone, not the end product.
Apoxie | 10 hours ago
But does easier software development justify the trillions of dollars put into AI? Not by a long shot. Try to get it to solve more complex problems and it fails miserably. Or try any novel idea, like give it some unsolved math problems - it fails miserably
Sevinki | 10 hours ago
They are not pouring in billions for any one specific task. The technology is universal, at least it looks like that right now. The research happening now will enable applications that nobody is even thinking of right now. AlphaFold or ClaudeCode are tiny fractions of what will likely become possible.
Compare it to investment in advanced physics and astronomy. What value has CERN provided, or the apollo program? Was it worth the billions of dollars? Not directly, but indirectly that knowledge we gained will be immensely more valuable for generations to come.
The scientist that discovered nuclear magnetic resonance was a physicist, not a medical doctor. Now his technology is used in MRI machines in hospitals around the world and likely increased the average life expectancy of all of humanity.
dsartori | 11 hours ago
It’s for sure going to revolutionize many parts of software development. That’s not enough to justify the silly valuations but it’s not nothing.
Dadoftwingirls | 11 hours ago
Yep, that's all I see, it's useful for software development. Big deal.
dsartori | 11 hours ago
It’s a big deal if you’re a software developer. But also if you think a little more deeply you may see that more sectors will eventually be amenable in this way.
Software is a fertile sector for this tech because we are keen to automate our work. We already had the right culture and many of the right tools. There are other uses cases that aren’t as ripe.
werpu | 12 hours ago
actually for specialized tasks it is really useful. I used it to move a codebase from python to c++ with an entirely different framework, it literally saved me months of work, but without expert guidance from my side no result which would have worked would have happened and there are severe limitations, in the end it is just another tool in you have. The problem is AI companies sell their stuff as solution for everything and a good way to automate tasks, but given the unpredictability of the results this is an outright lie, ai might help to speed up things but it definitely cannot be used to automate tasks which need precise results there is too much bullshit produced you have to detect and clean up! As usual CEOs believed the marketing papers more than actually experts telling what it can really do!
D-MAN-FLORIDA | 12 hours ago
Who is ready for the eventual government bailout?!?! You can just hear the tech bros already coming up with reasons why they don’t need to face any repercussions for their failures and instead is the failure of the government for not supporting them enough.
DisneyPandora | 11 hours ago
Sam Altman is no Steve Jobs.
When the IPhone was first introduced to the market, it was a paradigm shift. It was instantly beneficial to everyone. Also, it gave us the App Store, it gave us the Internet on our phone and so many unique world changing things. While AI does none of that and really does nothing life changing. Sam Altman has no vision. Steve Jobs had a vision.
devliegende | 11 hours ago
Sam has this ball thingy that scans your retina in exchange for crypto coins
TheGhostofJoeGibbs | 11 hours ago
>“When the IPhone was first introduced to the market, it was a paradigm shift. It was instantly beneficial to everyone. Also, it gave us the App Store, it gave us the Internet on our phone and so many unique world changing things. While AI does none of that and really does nothing life changing. Sam Altman has no vision. Steve Jobs had a vision.”
Those things were around before the iPhone.
Alcophile | 11 hours ago
I had the Internet on my phone years before the iPhone (Treo 650). All Apple did was enshitify an already very useful product by making it look cool at the expense of functionality and durability. I dropped my Treo dozens of times and barely nicked the case but now cases are an extra expense to protect phones that have become little glass boxes so they look cool.
And the app store?! Software has been available for home users since the 70s. All 'apps' have done is bloat our our computers with data we could access just as well online. Why do I need a Burger King app on my phone when there is a Burger King website? Oh, yeah, so I can increase BK's marketing reach and store some of their data on my device for free! Apps are just programs dressed up to make us think we need to download a bunch of shit better stored on the cloud.
Apple makes overpriced junk that sells because it's more or less idiot proof and people think it's cool for whatever reason.
hippydipster | 11 hours ago
We get the same economic value done with AI as we did without.
If they took all the AI away, we'd suddenly get far less done.
A paradox, yet true.
WhopperitoJr | 11 hours ago
lol no one here is reading the actual article and everyone is just assuming these CEOs are throwing in the towel. They’re not, they are doubling down:
“Instead of looking for other avenues for growth, though, PwC found that executives are worried about falling behind by not leaning into AI enough”
They basically perceive AI as working for their competitors, “so why shouldn’t it work for us?” And continue to chuck money by the cartload.
Master-Rent5050 | 11 hours ago
So they see random variations of revenue in their competitors, associate gains to AI (despite the fact that the majority lost money), and are doubling down?
WhopperitoJr | 11 hours ago
That’s part of it.
Of course, because generative AI and LLMs do have some utility, some companies have found ways to make it work for them. If just 5% of companies post meaningful gains and say it was because of AI, shareholders of other companies will clamor for “more AI” without understanding what that means, or providing a roadmap for its adoption.
CEOs also see this type of thing as “investment,” not “losing money.” If it fails, they will simply say that the rest of the market was chasing AI too and that they did their best. They will get a golden parachute and fuck off to Cabo.
Master-Rent5050 | 10 hours ago
It seems that the job that an LLM could do best would be CEO: it excels in following the herd and then giving bullshit reasons...
WhopperitoJr | 10 hours ago
CEO, middle managers, McKinsey consultants, anyone whose job depends mostly on advice and vibes and not actual production.
MoxMulder | 11 hours ago
The problem is that tech CEOs called their new, powerful text prediction “AI”, bought their own con, and are shocked that no one else did. A computer system that can do tedious, or long, or impossibly dense tasks is easy to sell. One that does the same things that internet searches or skilled people do, except worse, aren’t wanted or needed.
Ameren | 10 hours ago
I see it like the dot com bubble. There's a lot of useful potential in the technology —especially if we keep working at it— but currently it's very overvalued. And despite the tech firms' ardent claims, it's not a magical panacea for every problem either, it's just one tool in our toolkit.
Of course, one thing that concerns me greatly is the amount of capex that's going into rapidly depreciating assets. Like when we had a railroad bubble, the end result was a bunch of extra rail lines that became useful later. The same thing happened with the dot com bubble and fiber investment. Unfortunately, I don't think that the data centers and GPUs will endure as long. I wish more of that money was going into, say, power plants or other long-term infrastructure, since it's clear that we desperately need it.
Gajanvihari | 12 hours ago
Bad Blood 2: Ai
Or why do people still believe Silicon Valley? How many more fraudulant criminals do we have to run through?
All AI will do is teach me to disconnect.
sparduck117 | 11 hours ago
Hey CEOs you better triple down, put your own cash in too. I’m sure that will work out well for ya./s
That bubble needs to pop yesterday. The sooner the better, after all the more money invested and the more braindead moves (looking at you NVIDA) made, the more painful it’ll be for all of us.
DR_CAWK | 11 hours ago
They were too busy trying to beat each other to think about the ROI.
These companies could have literally posted here on reddit and they would have received clear advice that is the opposite of what they have been and what they are still doing.
They don't care. A couple of digits change in some direction of their worth. Ooo what if the CEO is only worth 100 billion.. ooo the poor life indeed.
Critical-Michael | 11 hours ago
Wasn't there a few articles stating that some CEOs or managers knew that these AIs weren't actually doing anything useful, BUT that they still pretended it did, fired a bunch of employees in the guise of "wow AI makes us sooooo efficient that we don't need so many workers" and just proceeded to replace those employees with the cheapest replacements possible to rake in that short-term profits?
Iassos | 11 hours ago
Does this mean we can start blowing up the data centers now, our power bills can come back down, and we can have affordable RAM? No? Come on, we need that real estate for housing.
ToasterBath4613 | 11 hours ago
I was in a global team meeting yesterday where I was told Sr Mgt are tracking AI training and product registration very closely and very soon they will be demanding use cases which will be tied to compensation. Meanwhile, I can’t get funding for access to market data needed to protect the company from regulatory sanctions. Hell, at this point, I’d be happy if I didn’t have to enter an IT ticket every day just to do my job.
Ambiguousdude | 11 hours ago
This was obvious to me, it's a tool that takes time to use and is fallible.
It's not rocket science to figure out the time returned is offset by double checking its outputs aren't bullshit.
cobra_chicken | 11 hours ago
AI requires someone that is a good manager to operate. Someone with ideas. Someone that is creative. Someone with initiative.
It is a tool focused on creating. It can help generate your ideas, flush them out, identify parts that are missing. If your job is just to enter data, do mindless reports, attend meetings all day, process transactions, or anything else like that, it is useless.
These pre-requisites exclude 90% of the jobs out there, and 90% of the people out there.
Combine that with leadership that fires it out with a shotgun and yeah, you are gonna get shit results.
Master-Rent5050 | 11 hours ago
Quite the opposite: AI is good at routine tasks, exactly the ones you described.
cobra_chicken | 10 hours ago
It is good at routine tasks, but to first get it to perform a routine task you have to break down the task, work out the integration (including credentials, mcp, api), detail exactly what is required, document guardrails around the process, etc..
I dont know about you, but most people i have encountered in the corporate world just know how to do what they have been told/trained to do, and thats all they will/can do.
Giving AI to teams that can create those processes is a great idea, but deploying it to Jane in accounting that has been doing the same thing for 30 years is not going to change anything. And thats exactly what most companies (including my current one) are doing, and then they expect magic.
NY_State-a-Mind | 11 hours ago
Well i use ai's like chatgpt to go down rabbit holes every day about random subjects and ideas for free and have never even see any ads and never paid a cent, how is that system going to make these companies money
Mysterious_Luck_1365 | 12 hours ago
One thing I’ve seen with the AI tools is that companies make up the hours of work that the AI tool is supposed to “save” the average worker. So now the expectation is that you have X more hours available to do other work.
Butane9000 | 11 hours ago
So they're catching up to what most people know?
The value in AI lies predominantly in what jobs it can automate away and what productivity it can increase. Not with what profits it can generate.
montgomeryyyy | 11 hours ago
When productivity goes up, profits rise. Not that hard to understand
ChazR | 11 hours ago
I'm a senior exec. Worker -> Supervisor -> Team Leader -> Manager -> Senior manager-> Junior Exec -> me.
All my (200+) people are knowledge workers.
I have encouraged and budgeted for use of AI in any way people find useful.
It doesn't work. People generate vastly more worthless verbiage, inconsistent charts, and pointless noise.
EXCEPT! I had a bit of a tanty and told my leadership team to stop sending me AI slop. Send me the *last prompt* they used to generate the pabulum.
And suddenly they started using AI to *focus their reasoning"
But I hire smart people so they short-circuited that, poured my written rage into the AI and asked it to "Write an AI prompt that will convince {me} of their argument, and make it sound like I wrote it myself."
We're doomed.
Mysterious-Abies4310 | 12 hours ago
Do you mean to tell me that a glorified web scrubbing tool isn’t generating revenue? No way!
I knew from the beginning that AI was going to go down the dot com bubble road. It’s just a matter of time now.
NotGoodSoftwareMaker | 12 hours ago
They all knew it, its really just an excuse for other things
Nice way to cut headcount without looking bad
You can also do headcount reduction squared. This is where you cut headcount, hire back the top performers at lower cost and you can market it as “AI is not ready”
Good way to gaslight your competitors into over investing in non-performing assets
Good way to float valuations
It will be the same as all the DEI stuff. The moment its not in vogue then it suddenly is expunged as though it never happened and no-one is held accountable
CopiousCool | 6 hours ago
It also helps suppress wages in an inflationary economy
realityGrtrThanUs | 11 hours ago
We've seen this happen multiple times. A really awesome innovation comes about. Marketing takes over and pushes it well beyond it's capabilities.
What's fascinating though is the executive response. Corporate restructuring, layoffs, and rebranding completely devoid of understanding and layered in supreme confidence.
Again and again we see this happen. My take away isn't that executives are polished actors with no sense. We know this already.
My takeaway is that most businesses are machines with monkeys trying to break them.
Digga-d88 | 11 hours ago
Twice this week, two different AIs failed in the most basic ways. I used Shopify's AI assistant to see what products sold over the weekend at a small location. The AI didn't get the dates right. It came back with zero sold. Told my boss who knew items sold. Went back into analytics and it got the date wrong. It claimed Saturday was the 18th (not the 17th in the reality we live in). I typed in the prompt why "did it think Saturday was the 18th?". It's response was "Good catch! Oops, my mistake." Yesterday I used ChatGPT to sort last names of something alphabetically. It had a last name beginning with D at the end. I asked again, "why did you put D at the end?" To another "Good Catch! Oops my mistake." Like, is AI getting worse? These are basic functions.
BowtiedAutist | 11 hours ago
LLMs while they are considered a form of ai. I consider them glorified encyclopedias. Great tools to add to the bucket but they are definitely not ladders you want to depend on.
cmack | 11 hours ago
super search engine
Wild_Space | 11 hours ago
30% saw an increase in revenues and an additional 13% saw a decrease in costs. 56% saw neither. You can look at this as glass half empty this all you want, but the 56% are doing it wrong and will seek to emulate the 43%.
Source: https://www.pwc.com/gx/en/issues/c-suite-insights/ceo-survey.html
PrisonMike1988 | 10 hours ago
One does not go to a PWC survey of CEOs for an accurate depiction of reality
dpholmes | 11 hours ago
My hot take: the comp science community is having its “rapid prototyping” movement.
This came ages ago for electronics, in which mass production of components allowed people to build simple circuits and electrical devices.
This moment came for the MechE and Mat Sci communities in the 2010’s with the ease of 3D printing, laser cutters, desktop CNC’s.
Now it’s here for CS but they mistakenly think that CS is everything, and so unbounded hype.
usandholt | 11 hours ago
This is really a case of “Majority of CEOs are in a company that hasn’t even gone through digital transformation yet and have not done anything other than enabling an internal ChatGpT to their organization”
I spread from experience
pcurve | 11 hours ago
"Only 30 percent reported increased revenue from AI in the last 12 months."
They're being fed lies from their underlings if they think this is actually true. 30% lmao.
adumblittlebaby | 11 hours ago
A bunch of sociopaths leveraged the world economy in a greedy dream to capture the next big thing, except it isn’t so big and now they’re panicking at the first rays of consequences.
itsTF | 11 hours ago
For anyone commenting without even peeking at the article, it says that 56% of CEOs did not see any revenue gains or cost savings in the last 12 months.
That means that 44% did. 44% is really, really good for a new technology like this, and absolutely justifies the hype. Likely, a large portion of the 56% who are not seeing returns are just doing half-baked rollout plans, getting ripped off by 3rd party companies offering their services, etc.
Very bullish on the 44% report, hilarious how much a headline framing changes the discourse.
ndnver | 10 hours ago
With 100 of billions of dollars spent to develop AI without a return I suspect we will have no choice but to offer a tax taxpayer bill out to the billionaires again so that they don't lose money on this.
canigetahint | 10 hours ago
They tried so hard to create an artificial need and a use case but failed miserably. AI is great for data analysis and sorting through data. Other than that, meh. It will be relegated as an add-on to Excel (should Micro$lop survive the W11 / Copilot disaster that they created).
unknownstudentoflife | 10 hours ago
I'm working as an ai product engineer for startups. I've build countless of automations or ai integrations into products.
Most companies treat ai as the holy savior that is going to solve their insufficiencies.
In reality most companies don't benefit from ai since to the most of them, its not even clear what problems they have inside their own business.
There are probably only 3/4 usecases i've found where ai really benefits companies. Most of the time. You have to restructure your entire business model for it to make sense.
AngelousSix66 | 10 hours ago
Pretty sure we've been using AI for years with great success. In my previous working experiences in multinational corps, we've been using AI for tech and call centers with the centers in Mumbai or Hyderabad for more than 10years and achievied significant savings!
Are they mistaking AI for something else? A stands for Affordable, how can there be no returns? /s
CoffeeSubstantial851 | 10 hours ago
Why would it? The entire point is to displace labor which in the process removes the scarcity component that gave that labor value. If everyone is a doctor, then being a doctor is worthless. That doesn't mean that there would be no societal value to having 8 billion+ doctors... but, it does mean that paying for one would be stupid.
So why would it be different for AI? If you have an AI that can do X task, that task is now worthless because the supply component of the equation is now near infinite and the demand is the same. You can't get a return from a product designed to destroy the concept of value itself.
redheadedandbold | 10 hours ago
CEO are not the smartest people on the planet. After the dotcom debacle, anyone with a modicum of technical knowledge should have seen this as the over-hyped, under-performing BS that it was.
Yes, there are fields where AI has helped, such as medical research. "Broad spectrum" AI has very few useful (legal?) applications. It's when AI is used to solve/help solve very specific questions (can weaken cancer's ability to reproduce by targeting X?) that it performs well. Or, so it seems based on what I've read to date.
QuerulousPanda | 10 hours ago
Was AI ever intended to deliver financial returns to anybody beyond Nvidia and a handful of specific people?
Like sure it got sold claiming it would be hugely beneficial to all the end users and companies but it's pretty obvious that it was just some bullsh*t given the pricing structure now and in the near future.
The people who were intended to get wealthier from AI already have been, and the entire rest of the world is going to be left holding the bag.
dano1066 | 12 hours ago
Because most CEOs have no idea how to leverage it. AI in the hands of engineers to do things for engineers is a game changer however. Not just as a coding helper but as a means to automate manual workflows. My life is a total breeze these days
KingOfEthanopia | 11 hours ago
An engineer or tech persons understanding and usage of AI is pretty reasonable.
The issue is MBAs overhype it to sell to other MBAs who then slash their workforce with unrealistic expectations.
para2para | 11 hours ago
The amount of cope in this thread is unreal. Instead of being so Luddite about AI, the people in here should really consider learning about how it works and using it to help themselves, not their companies. That’s the first step.
wolfiasty | 11 hours ago
I'll bite if I may.
I'm absolutely not luddite, I'm terrified of how AI will change job market, because it will, it does already. And I sincerely hope long run it will be for the better (though it probably won't for majority of people).
Anyway - how can I use AI to help me ? In what way ?
para2para | 10 hours ago
What kind of job do you do? Or even in your personal life what are some things that you struggle with? It all starts here. You can shape the AI to help you in your own way for your own problems first.
I help people do this in my job now, and I build AI systems for internal enablement at a larger company.
For my own individual organization of my work, one of the biggest boons to my productivity have been being obsessive about getting a transcript of all of my notes and meetings, and then I load all of that into two places: first a Google folder, which triggers an automation that automatically summarizes that meeting transcript and adds it to a long form file that AI can understand. I have that file plugged into a few different AI tools so that they always know what I’ve been working on so I can talk to that and get caught up quickly since I am generally a very disorganized person.
I also loaded the individual meeting transcriptions in notebook LM along with that summary file to get even higher detail.
There’s a lot that goes into this: the AI will hallucinate you have to learn how it intakes information, you have to be highly critical of its responses, but as you get more and more of a feeling for how it works, you can filter and use its outputs to help you speed along with whatever you’re doing and save a ton of time.
I envision a future where people are able to take time back from their employers because they’re able to do so much more with AI. Good people need to be in charge of this. It can really do so much for humanity.
wolfiasty | 10 hours ago
Well that's the thing. I don't struggle with anything at home. AI will not throw the trash away, not yet. I'm a land surveyor 80% site, 20% office, working for my boss. I don't make notes, I know my job and would consider myself good at it. We don't have grand meetings that would need me to take notes. Or at least nothing that wouldn't fit onto post-it.
I actually tried on my own to see if I can implement AI into something, but it came down to nothing.
That's why I asked, because I don't think AI, at this point at least, is there for majority people. Like you say it should work miracles in "office spaces" if properly implemented, but rest ? Kind of not there yet.
para2para | 10 hours ago
You could do something where maybe when you’re on site and you have things that you normally would have to look up you can send images of it through and have it processed to AI to get the information more quickly. You can maybe finish visits more quickly. Etc.
But yes, certain jobs do not need it yet or are not benefited yet. Certain jobs may never. Etc..
If you are good at your job, then that’s great. It can be from more of a time saving side of things. If you have a certain way of doing notes onsite, maybe you can have it expedite that by having a format that it automatically makes for you based on how you like things.