Companies are going all-in on artificial intelligence right now, investing millions or even billions into the area while slapping the AI initialism on their products, even when...
Companies are going all-in on artificial intelligence right now, investing millions or even billions into the area while slapping the AI initialism on their products, even when doing so seems strange and pointless.
Heavy investment and increasingly powerful hardware tend to mean more expensive products. To discover if people would be willing to pay extra for hardware with AI capabilities, the question was asked on the TechPowerUp forums.
The results show that over 22,000 people, a massive 84% of the overall vote, said no, they would not pay more. More than 2,200 participants said they didn't know, while just under 2,000 voters said yes.
someone tried to sell me a fucking AI fridge the other day. Why the fuck would I want my fridge to "learn my habits?" I don't even like my phone "learning my habits!"
Hi Zron, you seem to really enjoy eating shredded cheese at 2:00am! For your convenience, we’ve placed an order for 50lbs of shredded cheese based on your rate of consumption. Thanks!
Know when you're about to put groceries in so it makes the fridge colder so the added heat doesn't make things go bad.
Know when you don't use it and let it get a tiny bit warmer to save a teeny bit of power. (The vast majority of power is cooling new items, not keeping things cold though.)
Tell you where things are?
Ummm... Maybe give you an optimized layout of how to store things?
So I can see what you like to eat, then it can tell your grocery store, then your grocery store can raise the prices on those items. That's the point. It's the same thing with those memberships and coupon apps. That's the end goal.
And it would improve your life zero. That is what is absurd about LLM’s in their current iteration, they provide almost no benefit to a vast majority of people.
All a learning model would do for a fridge is send you advertisements for whatever garbage food is on sale. Could it make recipes based on what you have? Tell it you want to slowly get healthier and have it assist with grocery selection?
Exactly, it’s entirely about extra monetization. They all think in terms of hype and money, never in terms of life improvement.
I’d actually love AI to control something like a home assistant setup by learning how I like things and predicting change (mind you I still need to get it set up at all). But most people don’t even want a smart home.
Make something that makes the unpleasant parts of life easier and people will be happy with it
it doesn't seem all that hard to make, as long as you don't mind the severely reduced flexibility in capacity and glass bottles shattering against each other at the bottom
I want AI in my fridge for sure. Grocery shopping sucks. Forgetting how old something was sucks. Letting all the cool out to crawl around to see what I have sucks.
I want my fridge to be like the Sims, just get deliveries or pickup the order. Fill it out and get told what ingredients I have. Bonus points if you can just tell me what recipes I can cook right now, even better if I can ask for time frame.
That would be sick!
Still not going to give ecorp all of my data or put some half back internet of stings device on my WiFi for it. But it would be cool.
Ye, that'd be sick! and that's also not what was being sold! this fridge did none of that. What exactly made it "AI" I didn't bother to find out, but I work in IT. I guarantee it wasn't this. Also, not convinced I want my fridge to be able to spend my money for me. I want to be able to have a Ramen month if I need/want
Absolutely this. There IS a scenario in which I would love a "smart" or "AI" fridge, but it's gotta be damn impressive to even be worth my time.
It needs to know everything in my fridge, how long it's been there and it's expiration date, and I want it to build grocery lists for me based on what is low, and let me know ahead of time that I should use something up that's going bad soon. Bonus points if it recommends some options for how to do that based on my tastes. And I want to do this without having to manually input or remove everything.
But we're still SO far from being able to do this reliably, let alone at any kind of acceptable price point, and yet fridge makers keep shoving out dumb fridges with a screen on them and calling them "smart". I hate it.
I'm still pissed about the fact that I can't buy a reasonably priced TV that doesn't have WiFi. I should never have left my old LG Plasma bolted to the wall of my previous house when I sold it. That thing had a fantastic picture and doubled as a space heater in the winter.
The even crazier part to me is some chip makers we were working with pulled out of guaranteed projects with reasonably decent revenue to chase AI instead
We had to redesign our boards and they paid us the penalties in our contract for not delivering so they could put more of their fab time towards AI
This is one of those weird things that venture capital does sometimes.
VC is is injecting cash into tech right now at obscene levels because they think that AI is going to be hugely profitable in the near future.
The tech industry is happily taking that money and using it to develop what they can, but it turns out the majority of the public don't really want the tool if it means they have to pay extra for it. Especially in its current state, where the information it spits out is far from reliable.
I don't want it outside of heavily sandboxed and limited scope applications. I dont get why people want an agent of chaos fucking with all their files and systems they've cobbled together
NDA also legally prevent you from using this forced garbage too. Companies are going to get screwed over by other companies, capitalism is gonna implode hopefully
I have to endure a meeting at my company next week to come up with ideas on how we can wedge AI into our products because the dumbass venture capitalist firm that owns our company wants it. I have been opting not to turn on video because I don’t think I can control the cringe responses on my face.
Back in the 90s in college I took a Technology course, which discussed how technology has historically developed, why some things are adopted and other seemingly good ideas don't make it.
One of the things that is required for a technology to succeed is public acceptance. That is why AI is doomed.
Someone did a demo recently of AI acceleration for 3d upscaling (think DLSS/AMDs equivilent) and it showed a nice boost in performance. It could be useful in the future.
I think it's kind of a ray tracing. We don't have a real use for it now, but eventually someone will figure out something that it's actually good for and use it.
We have plenty of real uses for ray tracing right now, from blender to whatever that avatar game was doing to lumen to partial rt to full path tracing, you just can't do real time GI with any semblance of fine detail without RT from what I've seen (although the lumen sdf mode gets pretty close)
although the rt cores themselves are more debatably useful, they still give a decent performance boost most of the time over "software" rt
One of our helpdesk told me about his amazing idea for our software the other day.
"We should integrate AI into it..."
"Right? And have it do what?"
"Uh, I don't know"
This from the same man who came up with an idea for orange juice pumped directly into your home, and you pay with crypto.
And the scary thing is, I can imaging these things coming out of the mouths of people in actual positions of power, where laughing at them might actually get people fired...
who came up with an idea for orange juice pumped directly into your home
That maybe not as cool, but pneumatic city-wide mail system would be cool. Too expensive and hard to maintain, not even talking about pests and bacteria which would live there, but imagine ordering a milkshake with some fries and in 10 minutes hearing "thump", opening that little door in the wall of your apartment and seeing a package there (it'll be a mess inside though).
What I do mind is the software running on my PC sending all my personal information and screenshots and keystrokes to a corporation that will use all of it for profit to build user profile to send targeted advertisement and can potentially be used against me.
I would pay for a power efficient AI expansion card. So I can self host AI services easily without needing a 3000€ gpu that consumes 10 times more than the rest of my pc.
I would consider it a reason to upgrade my phone a year earlier than otherwise. I don’t know what ai will stick as useful, but most likely I’ll use it from my phone, and I want there to be at least a chance of on-device ai rather than “all your data are belong to us” ai
AI for IT companies is looking more and more like 3D was for movie industry
All fanfare and overhype, a small handful of examples that do seem a solid step forward with millions others that are just a polished turd. Massive investment for something the market has not demanded
Well, NPU are not in pair with modern GPU. General GPU has more power than most NPUs, but when you look at what electricity cost, you see that NPU are way more efficient with AI tasks (which are not only chatbots).
This is yet another dent in the “exponential growth AGI by 2028” argument i see popping up a lot. Despite what the likes of Kurzweil, Musk, etc would have you believe, AI is severely overhyped and will take decades to fully materialise.
You have to understand that most of what you read about is mainly if not all hype. AI, self driving cars, LLM’s, job automation, robots, etc are buzzwords that the media loves to talk about to generate clicks. But the reality is that all of this stuff is extremely hyped up, with not much substance behind it.
It’s no wonder that the vast majority of people hate AI. You only have to look at self driving cars being unable to handle fog and rain after decades of research, or dumb LLM’s (still dumb after all this time) to see why. The only real things that have progressed quickly since the 80s are cell phones, computers, etc. Electric cars, self driving cars, stem cells, AI, etc etc have all not progressed nearly as rapidly. And even the electronics stuff is slowing down soon due to the end of Moore’s Law.
There is more to AI than self driving cars and LLMs.
For example, I work at a company that trained a deep learning model to count potatoes in a field. The computer can count so much faster than we can, it’s incredible. There are many useful, but not so glamorous, applications for this sort of technology.
I think it’s more that we will slowly piece together bits of useful AI while the hyped areas that can’t deliver will die out.
Machine vision is absolutely the most slam dunk "AI" does work and has practical applications. However it was doing so a few years before the current craze. Basically the current craze was driven by ChatGPT, with people overestimating how far that will go in the short term because it almost acts like a human conversation, and that seemed so powerful .
Idk robots are absolutely here and used. They’re just more Honda than Jetsons. I work in manufacturing and even in a shithole plant there are dozens of robots at minimum unless everything is skilled labor.
I think manufacturers need to get a lot more creative about simplified computing. The RPi Pico's GPIO engine is powerful yet simple, and a good example of what is possible with some good application analysis and forethought.
Problem for the big market is that it's hardly profitable. In fact make things too easily multipurpose and you undercut your specialized devices opportunities. Why buy a smart device for 500 dollars that requires a monthly subscription when you could get a 100 dollar device with a popular preload of a solution on it?
Like when the WRT54G came out in the day and OpenWRT basically drove Cisco to buy out Linksys to neuter the "home router" to stop it displacing expensive products in the business sector. The WRT54G was the best product for the market, but not the best product to exist for vendor profitablity.
Remember when the IoT was very new? There were similar grumblings of "Why would I want talk to my refridgerator?" And now more and more things are just IoT connected for no reason.
I suspect AI will follow as similar path into the consumer mainstream.
So the trades have been unknowingly fucking with AI for decades, because of the time honored tradition of fucking with apprentices.
A lot of forums are filled with absolutely unhinged advice, and sprinkled in there is some good advice. If you know what you're doing, you can spot the bullshit.
But if you don't know anything about it, the advice seems perfectly reasonable. There's a skill in giving unhinged advice. Literally you can't get your master cert without convincing at least one apprentice to ask where the board stretcher is.
Do I actually have a dedicated vise for Vaseline when I run a tap cycle or is that old timer bullshit? HOW WOULD YOU POSSIBLY KNOW??
I still don’t understand how the buzzword of AI 10x’d all these valuations, when it’s always either:
a) exactly what they’ve been doing before, now with a fancy new name
b) deliberately shoehorning AI in, in ways with no practical benefit
Isn’t that the entire point behind what most business people do? The whole goal is to upsell some schmuck by speaking too fast, and mentioning a lot of words that don’t really mean anything. Except the difference now is that the business person in this case is the leadership behind most of the tech industry
That's kind of abstract. Like, nobody pays purely for hardware. They pay for the ability to run software.
The real question is, would you pay $N to run software package X?
Like, go back to 2000. If I say "would you pay $N for a parallel matrix math processing card", most people are going to say "no". If I say "would you pay $N to play Quake 2 at resolution X and fps Y and with nice smooth textures," then it's another story.
I paid $1k for a fast GPU so that I could run Stable Diffusion quickly. If you asked me "would you pay $1k for an AI-processing card" and I had no idea what software would use it, I'd probably say "no" too.
I'm still not sold that dynamic text generation is going to be the major near-term application for LLMs, much less in games. Like, don't get me wrong, it's impressive what they've done. But I've also found it to be the least-practically-useful of the LLM model categories. Like, you can make real, honest-to-God solid usable graphics with Stable Diffusion. You can do pretty impressive speech generation in TortoiseTTS. I imagine that someone will make a locally-runnable music LLM model and software at some point if they haven't yet; I'm pretty impressed with what the online services do there. I think that there are a lot of neat applications for image recognition; the other day I wanted to identify a tree and seedpod. Someone hasn't built software to do that yet (that I'm aware of), but I'm sure that they will; the ability to map images back to text is pretty impressive. I'm also amazed by the AI image upscaling that Stable Diffusion can do, and I suspect that there's still room for a lot of improvement there, as that's not the main goal of Stable Diffusion. And once someone has done a good job of building a bunch of annotated 3d models, I think that there's a whole new world of 3d.
I will bet that before we see that becoming the norm in games, we'll see LLMs regularly used for either pre-generated speech synth or in-game speech synthesis, so that the characters say text (which might be procedurally-generated, aren't just static pre-recorded samples, but aren't necessarily generated from an LLM). Like, it's not practical to have a human voice actor cover all possible phrases with static recorded speech that one might want an in-game character to speak.
This. Apple is doing it the right way, avoiding the term AI and instead focusing on what benefits it brings in iOS18. Other companies need to figure out what problem people need to solve and what AI would do to solve it. Instead they’re trying to cram it into everything and people are largely nonplussed about it.
I have no clue why any anybody thought I would pay more for hardware if it goes with some stupid trend that will be blow up in our faces soon or later.
I don't get they AI hype, I see a lot of companies very excited, but I don't believe it can deliver even 30% of what people seem to think.
So no, definitely not paying extra. If I can, I will buy stuff without AI bullshit. And if I cannot, I will simply not upgrade for a couple of years since my current hardware is fine.
In a couple of years either the bubble is going to burst, or they really have put in the work to make AI do the things they claim it will.
I agree that we shouldn't jump immediately to AI-enhancing it all. However, this survey is riddled with problems, from selection bias to external validity. Heck, even internal validity is a problem here! How does the survey account for social desirability bias, sunk cost fallacy, and anchoring bias? I'm so sorry if this sounds brutal or unfair, but I just hope to see less validity threats. I think I'd be less frustrated if the title could be something like "TechPowerUp survey shows 84% of 22,000 respondents don't want AI-enhanced hardware".
I honestly have no Idea what AI does to a processor, and would therefore not pay extra for the badge.
If it provided a significant speed improvement or something, then yeah, sure. Nobody has really communicated to me what the benefit is. It all seems like hand waving.
what they mean is that they are putting in dedicated processors or other hardware just to run an LLM . it doesnt speed up anything other than the faux-AI tool they are implementing.
LLMs require a ton of math that is better suited to video processors than the general purpose cpu on most machines.
I honestly have no Idea what AI does to a processor
Parallel processing capability. CPUs historically worked with mostly-non-massively-parallelizable tasks; maybe you'd use a GPU if you wanted that.
I mean, that's not necessarily "AI" as such, but LLMs are a neat application that uses them.
On-CPU video acceleration does parallel processing too.
Software's going to have to parallelize if it wants to get much by way of performance improvements, anyway. We haven't been seeing rapid exponential growth in serial computation speed since the early 2000s. But we can get more parallel compute capacity.
I use a llm fine tuned on medical stuff for minor medical questions or to prep for medical appointments (getting on the same page as the doc can save some serious time, say the wrong thing and they'll get hung up on it for a year lol).
I really want to combine it something like fasten health so I can go over my medical data on my own machines faster. Pipe dream rn because getting that data from the docs is a pain in the dick, but still would be cool to me.
Depends on what kind of AI enhancement. If it's just more things nobody needs and solves no problem, it's a no brainer.
But for computer graphics for example, DLSS is a feature people do appreciate, because it makes sense to apply AI there.
Who doesn't want faster and perhaps better graphics by using AI rather than brute forcing it, which also saves on electricity costs.
But that isn't the kind of things most people on a survey would even think of since the benefit is readily apparent and doesn't even need to be explicitly sold as "AI". They're most likely thinking of the kind of products where the manufacturer put an "AI powered" sticker on it because their stakeholders told them it would increase their sales, or it allowed them to overstate the value of a product.
Of course people are going to reject white collar scams if they think that's what "AI enhanced" means.
If legitimate use cases with clear advantages are produced, it will speak for itself and I don't think people would be opposed.
But obviously, there are a lot more companies that want to ride the AI wave than there are legitimate uses cases, so there will be quite some snake oil being sold.
well, i think a lot of these cpus come with a dedicated npu, idk if it would be more efficient than the tensor cores on an nvidia gpu for example though
edit: whatever npu they put in does have the advantage of being able to access your full cpu ram though, so I could see it might be kinda useful for things other than custom zoom background effects
But isn't ram slower then a GPU's vram? Last year people were complaining that suddenly local models were very slow on the same GPU, and it was found out it's because a new nvidia driver automatically turned on a setting of letting the GPU dump everything on the ram if it filled up, which made people trying to run bigger models very annoyed since a crash would be preferable to try again with lower settings than the increased generation time a regular RAM added.
AI in Movies: "The only Logical solution, is the complete control/eradication of humanity."
AI in Real Life: "Dave, I see you only have beer, soda, and cheese in your fridge. I am concerned for your health. I can write you a reminder to purchase better food choices."
Dave: "THESE AI ARE EVIL, I WILL NEVER SUBMIT TO YOUR POWER!"
There's a lot of use cases in manufacturing where you can do automated inspection of parts as they go by on a conveyor, or have a robot arm pick and place parts/boxes/pallets etc.
Those types of systems have been around for decades, but they can always be improved.
It's not really done better in the cloud if you can push the compute out to the device. When you can leverage edge hardware you save bandwidth fees and a ton of cloud costs. It's faster in the cloud because you can leverage a cluster with economies of scale, but any AI company would prefer the end-user to pay for that compute instead, if they can service requests adequately.
Most people won't pay for it because a lot of AI stuff is done cloud side. Even stuff that could be done locally is done in the cloud a lot. If that wasn't possible, probably more people would wand the hardware. It makes more sense for corporations to invest in hardware.
It just doesn't really do anything useful from a layman point of view, besides being a TurboCyberQuantum buzzword.
I've apparently got AI hardware in my tablet, but as far as I'm aware, I've never/mostly never actually used it, nor had much of a use for it. Off the top of my head, I can't think of much that would make use of that kind of hardware, aside from some relatively technical software that is almost as happy running on a generic CPU. Opting for AI capabilities would be paying extra for something I'm not likely to ever make use of.
And the actual stuff that might make use of AI is pretty much abstracted out so far as to be invisible. Maybe the autocorrecting feature on my tablet keyboard is in fact powered by the AI hardware, but from the user perspective, nothing has really changed from the old pre-AI keyboard, other than some additions that could just be a matter of getting newer, more modern hardware/software updates, instead of any specific AI magic.
They'll pay for it. When the tech companies decide, it's a thing to make money off & advertise it, all the good ants will buy, buy, buy and the rest of the time they will work, work, work for it.