If "AI" doesn't attract customers why are companies hyping it so hard? Are they just dumb or is there another factor?

Acamon@lemmy.world to Ask Lemmy@lemmy.world – 276 points –

I've seen reports and studies that show products advertised as including / involving AI are off-putting to consumers. And this matches what almost every person I hear irl or online says. Regardless of whether they think that in the long-term AI will be useful, problematic or apocalyptic, nobody is impressed Spotify offering a "AI DJ" or "AI coffee machines".

I understand that AI tech companies might want to promote their own AI products if they think there's a market for them. And they might even try to create a market by hyping the possibilities of "AI". But rebranding your existing service or algorithms as being AI seems like super dumb move, obviously stupid for tech literate people and off-putting / scary for others. Have they just completely misjudged the world's enthusiasm for this buzzword? Or is there some other reason?

76

Hype brings investment money to the table. When an emerging technology appears, you can say we are looking to develop those technologies into our existing products and you will see a bump up in your share price.

After a few years of failed products and the hype dies for the next thing you can never mention the old hype but keep the bump in share price.

Think about 5-7 years ago, Blockchain was all the hype, 5-7 before then was Machine Learning and XaaS, before that was Big Data.

Yeah, investors kind of amplify hype. When there is hype, you will have some investors investing money.
If there's investors investing money, it makes sense for other investors to try to invest first, so that their invested money gains value (the share price rises).
And then it becomes somewhat of a self-fulfilling prophecy, because suddenly you do have companies equipped with money to pursue that hype, which can feed back into the hype.

But similarly, you'll eventually reach a point where it does not live up to the inflated hype and then shareholders can just as well be extremely quick to pull out their money and amplify the crash.

Investors also know not every product will sell. So pad the bet and spread wide to increase your chances to score big.

My understanding is that a lot of venture capitalist funding is driven by gut feel and personal connection. Like, they'll tell you that they're the vanguard of the future with a vision, but most of the time they're just cliquey bros going "dude, sick" and burning money.

There's an anecdote in the book "the cold start problem" about how zoom got funding even though the guys funding it thought it was a solved problem, that a new video company wouldn't go anywhere, but the zoom guy was their bro so they gave him millions of dollars.

I feel like it's possible some future will look back at this the way we look at feudalism. Just like, that's such a bad system , why did people put up with it?

Just like, that's such a bad system , why did people put up with it?

Because hindsight is 20/20 and people had preconceptions back then that filled in the gaps, as they do right now.

The gaps are and were actually full of nonsense like "he's my buddy I'll give him money" but people expect the process to be a lot more reliable and solid, because they think they'd be more careful with that kind of money, not realising that to some millions are pocket change (and nobody is careful with pocket change) and that others gamble with other people's money and thus are a lot more cavalier.

Suits heard about this secret sauce called AI that can cut down on the need for those pesky humans that are always looking for handouts and luxuries like a living wage and benefits. The consumer will have to accept it when the only choices they're offered are varying flavors of the same shit.

(but also it doesn't work and they are, in fact, just dumb)

And then when it doesn't work, you blame other people and fire humans anyway and give yourself a raise for saving the company money. Stock prices rise.

Can confirm this is absolutely true. The number of meetings I've been in where execs are salivating...

Whereas in reality so far the payoffs are projected to be something like 2%. Not counting the costs of developing and running the AI stuff.

Attracts investors.

When people are evaluating companies, and see a company missing out on the current trend, how is that going to factor into their valuation of the stock prices?

  1. OpenAI struck gold, NVIDIA followed suit, and everyone else bought shovels hoping to get investors even though they have no plans on striking gold (developing useful AI).

  2. Would you like to buy a timeshare to the moon? If we all buy, you'll be able to sell your spot for 10x the price! Don't wait! Spots are limited!

Nvidia is the biggest shovel seller out there.

Nvidia sells the hardware (shovels), but also develops portions of the software to make it run more efficiently, like OpenAI. Nobody else but Microsoft seems to be actually developing software, though AMD is slowly working towards having comparable performance.

We kinda need to adapt the saying now. When someone finds gold, you need to sell wood and iron for all the shovel makers that will show up.

Because the boss thinks it sounds cool and doesn't want to be the only kid in the block without an AI product to sell.

Money was already spent. The hype companies were backed by big capital in their early days. Now the people who provided that capital want to cash out and they want their winnings. So you will have AI shoved down your throat on every piece of media channel those people also own. AI is a hype term that appears periodically since before the 2000s. This is nothing new. https://en.m.wikipedia.org/wiki/AI_winter

LLMs are toys that sparkle for a brief moment. Their value is laughable compared to their cost.

They believe that demand and offer in the market is an egg and chicken situation, so right now they're force feeding us the offer waiting for the demand to adapt

A lot of business people also think that AI is a "force multiplier" meaning that if they use it they can get more done in less time. Anything that can do that is basically a money printer at the business level, which is why all these execs and companies are so excited about it.

The problem is it's not or at least not reliably proven to be so. All these companies are jumping on board thinking "shove some AI in there and get 20% growth" when in reality there's no backing behind it working like that. And that's why a lot of customers are turned off, because from the consumer side, AI is just sloppy unoriginal junk. But on the business side they just see "Productivity is up" never mind that the productivity is garbage quality.

It's a symptom of shareholder-driven development.

Many companies pushing AI have had huge layoffs, and haven't launched anything worthwhile in years. Many of these companies have a metric fuck-ton of data, and already do some kind of AI (they probably have had LLM's for years too). This way, they can spend money and make it look like they're doing groundbreaking stuff to ensure shareholders are happy.

They'll continue to do stealth layoffs of people outside of AI, until the hype dies down, and they'll move to the next grift - after laying off all of their AI folks.

It's a "don't want to miss the ship" thing where companies have to invest in whatever's trending in case it becomes successful and gives you an advantage. If they wait until it's proven, they might miss a competitive advantage (having to start learning after others). In the case of AI it's even more important since the promise sounds actually useful (the summarize anything quickly bit at least), unlike, say, NFT. At least that's kind of how it got explained to me at one of my jobs.

Think like a venture investor.

A small chance of huge growth via new technology can have a big payoff. They expect most companies to fail and are more worried about missing an opportunity than losing money in a single bad investment.

Nobody is quite sure where AI technology will be in ten years, but if it's big, it's going to make people who got in early very rich. It doesn't matter that it sucks now; the web sucked in 1995, but it made people who got in (and out) at the right time very rich.

Shareholders and investors are more profitable than customers.

+1 to this one. Shareholders are the real customers these days. I'd say there's also a hefty amount of corporate peer pressure; "we'll be left behind if we don't move quick!"

I was discussing this with a friend. We came to the conclusion that "entrepreneur" means "unskilled, uneducated and unable to work" and that the harder a product is marketed, the more worthless it is.

I personally think people were "burned" by the whole NFT situation. During the NFT "hype" a year or two ago, a lot of companies were slow to get on board with releasing NFT products, and so they missed the bubble entirely. NFTs are, of course, silly but if they did take off, companies would have loved to have been part of the boon.

Fast forward to now and you have AI bros shilling AI in the same way cryptobros were shilling NFTs. However, this time it's different! They have results, they have technology. Microsoft is on board! They have fancy tech demos which are not staged at all! If you didn't have experience with the technology and limitations, you would be lead to believe that this is the same as the NFT bubble, but it's actually going to be a real technology rather than snake oil.

I think there's also the issue that it takes a long time to bring a product to market. Imagine you've spent millions developing software and hardware for your AI coffee machine or whatever and it turns out that there's no market demand. You can't really turn to your stakeholders and say "oops, we made a mistake and have to cancel this product. Sorry!", you have to finish the product and try to recoup losses where you can. That's why there's all these weird posts advertising AI products - they can't just not release a product, and AI bros might be tempted to buy it.

I also wonder if the whole AI hate is bias due to us being here on mastodonlemmy... We tend towards fairly cynical people who are critical towards new technology and corporations. Maybe actual consumers who aren't online all day and clued into the tech scene are wowwed by AI. I've certainly seen people here casually remark that they use ChatGPT and Copilot.

I also wonder if the whole AI hate is bias due to us being here on mastodonlemmyโ€ฆ

Yeah, we're cynical but we have every right to be.

I use ChatGPT, Copilot, and image generators for different things and I'm generally not on board with the blind hate because it's been nice to have an assistant that can do all these manial things. But honestly, I've gotten mixed results and don't see this tech correcting its obvious problems. The latest ChatGPT-4o release was great with its web browser, images, and speech, but it still struggles with accuracy to a tangible degree. Or worse, other companies use it for the wrong things as a cash grab to change perfectly working products. Even the applications that do seem perfect for it are not.

For example, I can't get Gemini to answer anything but simple questions about Google Docs without it getting confused and repeating the same thing. Copilot will sometimes reach conclusions wildly different from the sources it cites. ChatGPT will give you suboptimal code samples, be subtly wrong about the meaning of words in other languages, or suddenly forget part of my instructions. And now people are adding it to the fucking coffee machine for crying out loud. I'd have a different opinion if it were more accurate most of the time and genuinely useful, but using it more often only cements it in my mind as a secondary productivity tool rather than the main feature.

I hope the hype dies down and AI is seen as an afterthought enhancement rather than a stupid selling point. Anybody selling AI now looks clueless to me.

You're essentially describing FOMO. The hype bros are telling the CEOs "if you don't offer AI then your competition will, and they'll take all your customers."

You can't really turn to your stakeholders and say "oops, we made a mistake and have to cancel this product. Sorry!", you have to finish the product and try to recoup losses where you can.

You can and should. You're describing sunk cost fallacy, which is pretty close to universally understood as a terrible money vacuum of a flaw in our reasoning. (I would have made this comment if I hadn't read Quit literally yesterday, but it really is an excellent book about the value of abandoning bad decisions when new information makes it clear that they're bad decisions.) Buying time and raising expectations with a dead end nonsense tech might be better 6 months from now, but 5 years from now, being the guy that saw the writing on the wall that the continued investment was lighting money on fire will leave you better off.

LLMs have limited applications, and will in the future, but nowhere near enough to warrant the obscene amount of resource burn companies are spending to get there.

I also wonder if the whole AI hate is bias due to us being here on mastodonlemmy... We tend towards fairly cynical people who are critical towards new technology and corporations

Corporations, sure, but tech? Anti-tech people aren't early adopters of new tech products. Early adopters are just generally more aware of the actual shape of the field than people jumping on hype trains once they've already started moving.

So firstly, if you were the person running a several million dollar project which then gets cancelled, you are absolutely getting fired. If you were acting entirely in your own self interest, it would be better to make the project last as long as you can. Maybe it ends up succeding by a fluke and you keep your job?

Secondly, you're assuming that all that money just vanishes when the product is released. The product is still out there in the market, and there are still some people will buy it. If you're 80% of the way through the project, it might be worth spending the remaining 20% in order to recoup at least some of your costs.

People cancel projects and keep their jobs or get new ones all the time. Having the awareness that a project is junk is better for your employment long term than running through way more money and still failing miserably. The people with money notice, and other high level executives notice. Moonshots fail. Pull the damn plug when it's not viable.

You're not 80% of the way there. You haven't made a dent. Because the tech doesn't work and isn't capable of working. Every additional dollar you spend has a massively negative expected return. Chasing "make up your losses" on a project that's dead in the water is virtually never a valid strategy.

The things that make a company successful are not the same as the things that make executives successful

Because whilst technical people know it has limited applications (like blockchain), business people tend to fall for buzzwords easily because they don't realise a lot of the things it does were solved in other ways

A lot of people have come to realize that LLMs and generative AI aren't what they thought it was. They're not electric brains that are reasonable replacements for humans. They get really annoyed at the idea of a company trying to do that.

Some companies are just dumb and want to do it anyway because they misread their customers.

Some companies know their customer hate it but their research shows that they'll still make more money doing it.

Many people that are actually working with AI realize that AI is great for a much larger set of problems. Many of those problems are worth a ton of money; (eg. monitoring biometric data to predict health risks earlier, natural disaster prediction and fraud detection).

Many people that are actually working with AI realize that AI is great for a much larger set of problems. Many of those problems are worth a ton of money; (eg. monitoring biometric data to predict health risks earlier, natural disaster prediction and fraud detection

None of those are LLMs though, or particularly new.

You're right. They're not LLMs and they're not particularly new.

The main new part is that new techniques in AI and better hardware means that we can get better answers than we used to be able to get. Many people also realize that there's a lot of potential to develop systems that are much better at answering those questions.

So when people ask, "Why are companies investing in AI when customers hate AI." Part of the answer is that they're investing in something different than what most people think of when they hear "AI".

I see two basic reasons.

  1. it gives companies plausible argument to embed telemetry into their products. Should your TV manufacturer or coffee maker manufacturer be able to monitor every single button you press on your device? Probably not, but they would like to โ€œbecause AIโ€! Now they have an excuse to be as invasive as they want, โ€œto serve you betterโ€. The dream - for them - would be total surveillance of your habits to sell you more shit. Remember, it always comes back to money.

  2. The old adage never fails: if itโ€™s free, you are the product. Imagine AI being so pervasive, that now everywhere you look, everything you interact with can subtly suggest things. It doesnโ€™t have to be overt. But if AI can nudge the behavior of the masses to do a thing, like buy more soda, or favor one brand over another, then it has succeeded in boosting company bottom line. Sure the AI can do useful shit for you, but the true AI problem companies want to solve is โ€œsay or do the right shit to influence this consumer to buy my thingโ€. You are the target the AI is operating on. And with billions of interactions and tremendous training, it will find the optimal way to influence the masses to buy the thing.

AI is the new ad driven model.

Everything that AI touches, ends up machine learning content.

AI DJ? I now have your name your email address and every single taste you have in music. As you use the app I will gain more insight into more music that you are or might be interested in.

That roomba thats running around your house looking for socks and cables not to run over, is also image processing on everything in your house. We know how big your house is they probably know how big my TV is.

They're not just farming your email and text messages to figure out what to sell you they know at a core intimate level what you're interests are.

They're in for a rude awakening in a few years. All of this AI information gathering is a bubble. You have companies like anovo complaining that they can't afford to host a single website. All this AI training is not cheap and the return on investment is not great after the initial plunge right?

It delights me to lie to myself that they are nervous someone somewhere would pick a golden ticket with their AI application and they'd miss out. But more obvious explanation for big corpos is that they hide problematic data-mining, content appropriation, ad personalization and other stuff behind this curtain, maybe not for these crude tools alone, but to force a precedent into existence that they can do it whenever they like in the future. They make you give up your personal stonks for a shiny penny that is corporate LLM genies and they probably pay a lot to showcase their beauty at loss.

It delights me to lie to myself that they are nervous someone somewhere would pick a golden ticket with their AI application and they'd miss out

Tbf, that's exactly what happens sometimes. CVS partnered with Theranos despite the lack of evidence supporting their product. Their reasoning was that if only Walgreens partnered with Theranos and it was a success, then CBS would have been screwed

It's all a sunk-cost fallacy. They've dumped all of this money into it, so therefore they have to double down on it.

Especially if they're trying to get a bunch of money from Wall Street and other investors.

The biggest contributor being all of these companies believe they can now just lay off a bunch of workers and make up the difference with these LLMs even though they are not at all a replacement for humans.

Less workers, less people that have to pay, and more money can be funneled to the top.

And don't forget: companies are just a group of people, and they can fall for a good hyped up scam as easily anyone else.

I don't know if it's AI specific or just connected devices that people are starting to avoid. I don't want anything with the word "smart" in it. I don't want anything that requires an app, and I try to avoid things that require an account. I think people are fucking finally figuring out that all this shit is designed to spy on you for profit, and will probably eventually require a subscription if it doesn't already. It's a bullshit business model. Make a fucking decent product, and sell it to people for a flat fee, and they'll be your customers for life. Stop trying to be a data company, or a subscription company.

As for the second part of your question, companies are pushing it so hard because it's like having a money printing press. They can turn a few months worth of work into an endless stream of money.

They think theyโ€™ll get money.

Is why.

Unless a hyped-up investor gives it to them, they wonโ€™t.

If my workplace is in any way representative, it's because decisions are made by close to retirement out of touch old geezers who want to virtue signal very hard that they are not out of touch old geezers. So they push the "new thing" for lack of any actually innovative ideas of their own. Then, when the younger team members who do have some rough knowledge of the "new thing" try to explain why it might be a bad idea, they call them afraid of progress and double down on the "new thing" even harder.

Two mains reasons:

Attracting investors

Attracting talented workers by signaling they are doing technical research

Also, people working in the industry might not even use those products. They want a cool job not a cool product

AI has some useful applications, just most of them are a bit niche and/or have ethical issues so while it's worth having the tools and functionality to do things, no one can do much with them.

Like for example we pretty much have AIs that could generate really good audio books using your favourite actors voi e likeness, but it's a legal nightmare, and audio books are a niche already.

In game development being able to use AI for texture generation, rigging, animations are pretty good and can save lots of time, but it comes at the cost of jobs.

Some useful applications for end users are things like noise removal and dynamic audio enhancement AIs which can make your mic not sound like you are talking from a tunnel under a motorway when in meetings, or being able to do basic voice activation of certain tools, even spam filtering.

The whole using AI to sidestep being creative or trying to pretend to collate knowledge in any meaningful way is a bit out of grasp at the moment. Don't get me wrong it has a good go at it, but it's not actually intelligent it's just throwing out lots of nonsense hoping for the best.

It will kill us all or solve everything, step right up and place your bets! No ma'am, there is no third option to bet on, none at all I say.

I wonder how many of the people pushing it believe in some variation of Roko's Basilisk. Either that or they believe AI is going to enhance their data collection abilities; and that if everyone pushes AI together, there won't be any AI-less options and the consumer will be trapped into giving someone even more data than they already do.

No, they're pretty much just dumb. In tech, this works along hype cycles where there's gotta be some new thing all the fucking time, and it cures what ails ya and is perfect for every case. This mostly involves taking any actual merits of [new tech] and blowing them way out of proportion and context, making it the best thing since sliced bread. This invariably makes people invest because hype is more important than making sense. When the cycle for that particular tech winds down into the Trough of Disillusionment, a new one shows up.

We had IoT, Web3, and now AI. Part of it seems linked to very good salespeople pushing it onto other salespeople.

For the first two, we've seen business spinup quickly and have very aggressive arguments, backed by cash, pushed onto existing business as "the solution to everything". Only to burn down later as a gimmick nobody really cares beyond a handful of niche applications.

So far with AI there's a handful of "big name" business that pushes it as the ultimate solution for everything and are injecting ton of cash in that discourse. We just have to wait a bit if the last part of that happens. After that we'll go back to normal until the next "big thing" gets propped up.

They believe that we're on the precipice of creating agi. Agi, if invented, would be like inventing the nuclear bomb. Whoever does it first gets a massive leg up on the competition. Except in this case instead of destroying a rival with a bomb you just ask your Agi, "what's the best way to become the most profitable company in history" and it will tell you and it will be right.

Money, really. Someone thinks AI is going to make money, so everyone is trying to make money by slapping "AI" on everything to see if something sticks.

Honestly, the LLM / Generative AI tech is pretty cool and amazing that it even works, but it is still in its infancy. As one person put it, we're watching a baby (AI) take its first steps walking, but the people with money are going "Get that baby a job!"

It's nowhere near ready for daily driving / what it is advertised to do, but I believe the ones that are serious about making it work are hoping technological and programming innovation will come along that will make it more energy efficient and more accurate as it is used more.

They want to create some hype and look cool by using AI chatbots. And most normies don't care about privacy and the dangers of AI in the future, they only care about "wow I can use AI for bla.. bla.."

But they have no idea, that one day AI could take over their jobs.. and rich people like Sam Altman are getting richer, and he only pays you with UBI money some pieces of computing

https://x.com/tsarnick/status/1789107043825262706

Also, AI companies aim for government contracts and medium / big corpos.

It's just a modern version of "Fuzzy Logic"

And at this point it's also code for 'machine learning'.

Which really is just fancy statistics. Sometimes it's barely more advanced than plain regression.

businesses are hoping they can use ai to higher cheaper labor. they want to be able to higher someone who does not know how to use the relevant program but can coax a machine to produce it or can get enough information from a machine to do the job. also they hope to eliminate human trainers and such.

Venture capitalists and shareholders want to make money. Company executives want to give them eternally increasing profit. Simple as that.

It's simple. Because Microsoft can't say that their operating system is garbage and quietly wants their users to switch to Linux

Edit: someone missed the joke

So you're telling me the software developed by Microsoft is being used on Linux platforms to convince manufacturers using Linux to use Linux instead of Microsoft because Microsoft systems are bad?? Make it make sense dawg.