If you don't think the highly contained agents we have today will eventually be controlling their own dedicated hardware, you are mistaken.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
OpenAI's First AI Device Will Be a Portable Smart Speaker
- Thread starter MacRumors
- Start date
- Sort by reaction score
You don't have the first clue what this product is. And as I said, this is the beginning. It is very early. Hardware controlled and directed by AI is coming.
Happy_John
macrumors 68000
“Eventually.” Eventually the sun will burn out. Eventually we’ll all be dead.
This thread isn’t about eventually. It’s about OpenAI’s supposed first consumer hardware product.
And OpenAI sells access to big LLMs as a service, either as a flat subscription, per (million of) token.
This device is a means to get people to regularly pay OpenAI for compute/inference on those big models. It’s renting, not buying.
Smaller “on device” models is not their business strategy.
Siri was stupid long before AI was the center of attention. Privacy… only time will tell.
Where Apple shines is their world class ability to control the narrative and convince the faithful.
To be honest, that’s a good theory.
Although I can’t imagine the general population have much to offer in terms of model training data.
Advertising data though, there you could be honest something.
Thing is, though I can’t imagine this product being too successful, but we’ll see what happens.
I agree, the HomePods barely qualify as a smart speaker.
Good speakers, but not the most stable software.
I genuinely think there is a market for HomePods if they can just make the software stable and provide Siri AI so that it can actually be useful as a smart speaker.
Who said anything about AI? I was comparing to Alexa and Google Assistant.
How much time? Apple has been around for 50 years.
This is just something embittered people who don't agree with the mainstream say to make themselves feel superior.
No, you simply jumped to conclusions and, being MacRumors, those conclusions are just more negativity. You have no idea what their device will do, nor does the author of the article to which you applied your astute reading comprehension skills.
Yeah, like what they are trying to make Siri 2.0 to be.
But this smart speaker thing from OpenAI is just a shell that takes up your Wi-Fi bandwidth and drains your wallet by burning tokens. And in a world of increasing software-as-a-service products and subscriptions, hardware is becoming merely a beautiful wrapper.
And we are still decades away from what you describe as sentient AI assistant in a robot, or any other form. (By any chance, did you watch Terminator and/or Matrix trilogy?)
I'm probably not the first with this observation, but consider me a datapoint.
I think I would trust a company run by Mark Zuckerberg with my private information before I would trust a company run by Sam Altman. And I would run into traffic with my eyes closed before I would trust Mark Zuckerberg with my private information.
I think I would trust a company run by Mark Zuckerberg with my private information before I would trust a company run by Sam Altman. And I would run into traffic with my eyes closed before I would trust Mark Zuckerberg with my private information.
Numbah One
macrumors regular
I’m comparing rumors of one company with a vision and actual AI technology with a company that hasn’t been able to improve its home speaker due to lack of AI. When the new HomePod comes out, you might as well call it Google Nest Pro.You’re comparing 10 year the of technology to technology not yet released. That’s not a useful comparison.
The only useful comparison you will be able to make is when both OpenAI release their device and Apple release a new HomePod (or similar device) with Siri AI. Neither have been released, so you can’t say which is better.
Both will have very similar functions though, so it’s very doubtful that the release of both will squash the accusations that OpenAI stole Apple’s internal development and manufacturing processes.
There’s nothing about this proposed device that us original - all the “features” mentioned in the article have been attempted in other devices, and have been leaked as features in Apple’s (and other companies’) upcoming devices.
Take away the “OpenAI” branding, there’s not much here to suggest that thus will be gamechanger. Sticking te letters “AI” to a product doesn’t automatically magically make that product more exciting or innovative. Internally, it’s most likely a smartwatch or smartphone-based mainboard, some RAM, some storage, speakers, batteries, sensors + mic, and some servos to make the thing move, which is a purely cosmetic feature.
That type of hardware already exists in a cute form factor l:
There’s a strong smell of “air fryer” about this - people had ovens in their kitchens, but went out about “air fryers” anyway, ie, a second smaller oven.
So far, there is no indication that this OpenAI device will do anything that can’t be done on your smartphone. If you want a second device that does the same thing as your first device, cool, go for it, but there’s no added value o doing so except “cuteness” or aesthetics.
Speaker seems to me a sub-optimal choice.
I think they'd do better by embedding the AI in a companion "animal", like this sort of guy:
(These sorts of robot cats are, IMHO, charming, and capable of a few movements, but not currently of any sort of speech.)
I suspect people would enjoy talking to an animal more than to a speaker, and there's a ready market available in the form of lonely old people, people with mobility issues, that sort of population.
I think they'd do better by embedding the AI in a companion "animal", like this sort of guy:
(These sorts of robot cats are, IMHO, charming, and capable of a few movements, but not currently of any sort of speech.)
I suspect people would enjoy talking to an animal more than to a speaker, and there's a ready market available in the form of lonely old people, people with mobility issues, that sort of population.
I don’t know if it’s wise to introduce the elderly to tech with a propensity to provide suicide directions. Even if it’s smuggled in a cat-shaped object.
Mister Fluffy providing drano dosage recommendations seems like a lawsuit waiting to happen.
Decades? No I don't think. If you've ever used voice agents, you'd know that's not remotely true. You could put a voice agent inside a basic animatronic right now and you'd have a really convincing albeit limited android. The robotics part is the biggest hurdle, and that has been making massive strides itself.
quietstormSD
macrumors 65816
I feel like this would struggle with the 2 current juggernauts of Google and Amazon in the smart home speaker market. Apple sucks in this market too and can't compete with Google and Amazon.
I celebrate that we don't all agree, especially when a post is mis-read or misinterpreted.
turbineseaplane
macrumors Penryn
axlpicanha
macrumors newbie
It seems silly to latch onto complaints very specific to a particular time and place in a field that changes every month...
Kinda like complaining that the iPhone will never take off because, sure, EDGE works, but do you know how slow its data rate is?
Piggie
macrumors G3
I’d take it as a sign that the industry is less concerned with appropriate safeguards than pushing products. That seems to be consistent across all AI platforms.
I think there’s been enough high profile safety failures that it’s reasonable to default to skepticism. “Move fast and break stuff” sounds great in a PowerPoint, but sometimes break stuff means dead humans.
jdawgnoonan
macrumors 65816
I am sure that they will be controlling their own hardware, but a speaker and microphone system is still only a conduit.
Someone asked me to look over an interaction they had with an LLM to see how they communicate with it. The way they were typing to it… it was so creepy and I couldn’t imagine how someone fairly intelligent could be attracted to that. Then, I’m reminded how intelligent people ACTUALLY believe that the barista at their coffee place REALLY wants to know what their plans are for the day. And not that they were required to say that by the corporation that signs their check. Then, it makes a little more sense.
The things I’m interested in haven’t been included in models that I’ve used, so I get the predictable hallucination, I type a correction, followed by “You’re right,” blah blah blah