Inference should ideally be done on an instance with a GPU, and it should have enough GPU VRAM to hold the entire model. This one is only a 7b model, so it should run pretty easily in a modest amount of vram. You can likely run it locally in CPU-only mode, too, though it's likely to be rather slow.
You could try running these in Google Collab (though tweaks would have to be made to load the files), or you might try something like runpod.io, as well, which gets you GPU instances for a lot less than you'd pay with AWS - for example, a Tesla V100 (16GB) community cloud instance runs you $0.24/hr vs the g5g.xlarge which runs you $0.42/hr.
It feels like Apple has been very quiet about their approach in some of the future spaces (self-driving, VR/AR, AI) but then they suddenly burst out with something like Vision Pro.
It seems like they are biding their time and avoiding the hype knowing that they don't need to get their first. If they can produce a superior product their brand will let them dominate those markets whenever they decide to release.
Speaking as a VR enthusiast who has virtually no interest in it: The Vision Pro is an entirely new product relative to any other VR headset I know of. It offers a comprehensive set of computing features (discounting an App Store which has only really just started) that puts it roughly on par with something like iPad OS. I would say it dominates currently for VR headsets if that is what you're after, because it's basically your only option. Desktop computing in VR is possible with things like a Vive or an Index but the experience is entirely different, worse on every front, and is basically done as a workaround in case you need to operate your PC while in VR without taking off the headset. And of course, all of those headsets are (quite intelligent, but remain merely) displays: they require a fairly powerful computer to attach to. There are headsets that operate independently too, like Meta's offerings, but those are much, much, much more restrictive in terms of capability than the Vision Pro.
My interest in VR is basically solely for gaming, so the Vision Pro for me is a complete miss and I have no plans to purchase one. That being said, in the realm of productivity VR, I'd say they are dominating in that they're the sole entrant to the market that I'm aware of and their offering is extremely compelling.
To paraphrase YC/Sama etc. If there is a small grouop of initial users who are obsessively using it and can't get enough of it. Then VP will work and be here to stay. If not, then it is not their time.
So, the question is this. If you know anyone who actually got VP. Are they using it obsessively? I got occulus rift a while ago and so did many of my friends. We bought it but never really used it. Like probably a few times in a year to show friends.
I don't know enough to really say anything with confidence, apart from that if Apple's track record is anything to go by, if any company can get VR-for-work right, I suspect it's them. Not my 10 but I suspect it will be a lot of people's.
Primarily using it for software development. I'm a web dev. I'm on day three of all-day usage (~8hrs, with breaks) and my experience has improved with each day, both in terms of comfort and as I get used to visionOS. I have 10 windows open in front of me right now, including my calendars, Slack, Messages, Spotify (in Safari), ChatGPT, a reference browser for documentation, and a few other odds and ends. My main front and center window is my 4k MacBook Pro 16" screen in 1080p (larger than that and the screen is too large for me or the text too small, but I prefer slightly larger text usually).
I've also used it for reading, watching movies, and reminiscing over old photos. Something I don't think that has really been captured well in reviews is how incredible it is to look at life-size recreations of photos you took on your phone, even plain old 2D ones. There's something really special about seeing a photo at life-size scale wherever you want, even if you technically can achieve this with a massive projector. And that's to say nothing of the bittersweet nostalgic nature of spatial video / photos, which has, if anything, been profoundly understated.
It is not my first headset. Previous to the AVP, I owned Quests 1, 2, and 3, and I experienced the Rift before that. I tried, hard, to work all day in each of the Quests as that was my primary use case for them, but flamed out after a couple of hours every time.
It does, but you can connect a Bluetooth keyboard, which is what I do.
Or, if you are using virtual desktop, you can use your mouse and keyboard to control VP natively as well. The only drawback is that this may not work if you are on VPN (even if the desktop comes through OK). I'm hoping this is a software issue that will get straightened out soon.
What do you think of the AVP’s potential as a DAM tool, at least for hobbyists? Obviously a Lightroom subscription is way, way cheaper, but it seems like it would be awesome to have a native solution that ties in with iCloud sync and lets you blow up your photos to huge size for side-by-side comparisons.
Yeah, Not sure you can call “dominating” at having a unique feature that nobody uses. Time will tell but I think Vision Pro will have similar retention issues to other headsets. Fundamental problems still there (weight, vergence accommodation conflict, isolation…)
I'm not trying to come across as vehemently pro Vision Pro here, I'm genuinely not sure if there's a market for this, but like... I really feel the need to say, this is not like any other headset, just, at all.
On the weight front, it's significantly lighter than a Valve Index, by roughly 20%, and that's just the unit I'm pretty sure, the Index has an umbilical cord that is not-insubstantial in mass by itself that dangles off the back of your head, so already the Vision Pro is quite far ahead on comfort.
VAC is of course a factor in any headset and this really comes down to the per-user situation: I personally got my "VR legs" in about thirty minutes way back when I got started, and am able to use full motion movement pretty well now with only minimal vertigo when things go up and down too fast for my brain's liking. However, the VP is going to have a distinct advantage here with it's big complicated camera system that allows it to operate both in VR and in AR, with most applications seeming to live in the latter realm. I think that's going to lend itself to big comfort gains for people who aren't as blessed as me in that regard.
Isolation also is much improved with the VP, what with it employing again, the camera system, and also the fact that the Vision Pro (much like my Valve Index) doesn't cover or obstruct the ears. And you get those weird CG eyes you can plaster on the front of the thing, haha.
Like again, I don't know if this will work out long term, I'm interested to see if it will really. But I see people keep comparing it to other headsets and as a fervent enjoyer of many of these headsets, I really cannot overemphasize how bad that comparison is.
I personally don’t think is significantly better (own all the headsets since DK1) in terms of comfort. In some aspects worse. Yesterday had to change clothes to use the AVP because my pajamas don’t have pockets for the battery pack. Each headset has a different set of annoyances. not sure AVP is net positive in comparison.
Wearing a headset / computer on your face is a cost (vs wearing nothing) that the value prop gotta offset. We’ll see if it is a net positive, worth the tradeoff for a significant amount of people. Previous evidence indicates it isn’t.
It’s also got nearly the entire library of iPad apps available to it to tide users over until native alternatives appear, plus for many apps providing a native version is little more than ticking a box in Xcode on existing iOS projects.
This might sound trivial but it determines how long the user is able to use the device exclusively without relying on their phone, computer, etc which is a major factor in it being used at all.
There's only a slim chance of history repeating here, but the olds here may recall that in the early days of discrete graphics cards, particularly before 3D took over, there were a lot of people buying graphics cards for Serious Business Reasons but really in order to play games when off the clock.
A graphics card could make your spreadsheets go faster, but that's not really why they bought them.
It's plausible, however improbable, that a VR headset targeted at Serious Business Reasons could open doors that entertainment has not. And in particular, slightly unwieldy price points that can be rationalized a bit easier.
You can run your laptop and other apps with multiple screens as big and layed out where how you want while still having a usable (seeable) keyboard? Is the latency such that you don't get nauseous? Seems useful for productivity. Do other goggles have these feature, yes, but there's a difference between being a toy versus actually being as good/better than how I work now.
Someone just noticed Apple’s decades old SOP. Watch the aggregate fumble about, fail over and over, then take the few ideas that survive rapid, aimless iteration attempts and add the last mile of polish.
A decade of crappy Palm, WinCE, and easily forgotten Java based mobile devices came and went before iPhone.
They know what software people refuse to accept; it’s the hardware experience that matters and they’ll wait until the hardware is there.
Modern hardware is responsible for AI, more reliable networks, and the rest of modern compute. Has little to do with the bloated software stacks we git pull into the data center
Edit: Also consider that Intel and Apple don’t just make a consumer device. They design an advanced manufacturing pipeline. The difference between print(“hello world”) and for index in array: print(contents of index)
That’s the basis of their value and why they are propped up by government and why software focused startups will always just be pump and dump schemes.
Apple is the master of the second mover advantage. Let the existing companies burn money in R&D, but closely monitor how customers perceive their products. What pain points exist that the current companies haven't addressed? What are the pitfalls?
The first mover advantage gives you a chance to get most of the market share, but with Apple's walled garden, they don't need the first mover advantage to do that.
Even when they didn’t have that garden. Look at the iPod.
Other MP3 players existed, but flash media was too small. Other hard drive players existed, they were too big. And both had bad interfaces (by Apple standards).
Apple didn’t show up until they could meet all their goals.
They could have made a phone sooner, it wouldn’t have been the iPhone. They could have made a headset sooner, it wouldn’t be the Vision Pro.
They set their standards for what’s necessary and wait to be able to deliver that.
>> Because of the iPod Nano unveiling on the same day, relations between Motorola and Apple were strained and Motorola CEO Ed Zander later accused Apple of purposely undercutting the Rokr.
Apple probably saw it as an opportunity for free iTunes and Nano advertising.
I don’t know about advertising. But I would be willing to bet money that Steve Jobs chose that day for the Nano to do exactly what Zander thought: screw them over.
They (whoever ‘they’ is) caused the phone to be terrible even though it had the iTunes name on it. So Jobs got revenge, as he always did if he could.
I don’t remember the story of how Apple got involved in that, but it was incredibly clear when it was released that Steve Jobs hated the thing. Even when he demoed it on stage.
Perhaps they agreed to a partnership and didn’t realize that Motorola and the carriers would ruin it? Or perhaps it was some sort of “we’ll do this to learn some of the business” kind of a thing?
I’d love to know more of the story behind how it started and it’s development.
“They could have made a headset sooner, it wouldn’t be the Vision Pro“
Hmm I guess. Vision Pro seems rushed from vantage point. Odd battery pack … only 2 hour charge … limited field of vision. Also the screens that show a picture of your eyes are off putting and strange.
Think Apple may have rushed here too quickly. We need more battery life.. think we may be looking more for an Augmented Reality mode which works on top of what you see with your actual eyes rather than your eyes just looking at high def screens inside of a headset that is relaying a video captured from cameras of what is in front of you.
I agree that 3 would be nice, since it would get you through pretty much any movie, but I feel like when I'm watching a movie I don't mind plugging in. After all, I'm not likely moving around a whole bunch. I have USB-C plugs by my bed and near my couches, so it doesn't seem like this is a huge inconvenience (though it would be great if there were a "you're plugged in" warning if you start getting up while you're plugged to a wall).
But if the tradeoff is a heavier battery (or a more expensive device), I'm fine with 2 hours. No hot-swapping, that seems like a bigger miss! People don't want to shut down just to change battery packs.
That’s the tradeoffs they wanted to hit whatever guidelines they had. They clearly didn’t want to ship a 2 pound battery.
What I meant is they didn’t need chips as powerful. They didn’t need screens as high resolution or bright. They didn’t need so many cameras.
They could have released something closer to a Quest model in specs years earlier.
They waited for this.
There are always tradeoffs. First gen Apple stuff is often weak in some way (CPU, memory, battery, software) because they like to be at the edge of what’s possible. Then they improve from there.
Things are a bit different now, compared to the iPod. If Microsoft had released a competitor a year or two in advance of the iPod, Apple would have had a much harder time getting traction for their iPod because they didn't have a large walled garden back then. They only had a very small subset of the computer market, and iPods were only compatible with Macs (at first).
Apple can afford to be a much later mover, even against very well-resourced competitors, now that they have a huge walled garden. If they didn't have the walled garden, it would be very difficult for them to catch up.
I think people give way way too much credit to the walled garden for their success.
The iPod wasn’t a success because of the walled garden. They didn’t have one (once they released on Windows).
If MS had released a player that was good enough they could have stolen the market, yes. But that’s the point. No one did because everyone released something akin to a flash-based CD player. They didn’t rethink the interface much. The interface is why the iPod grabbed the hold it did.
Yeah it became cool over time.
But I contend an iPod with the interface of a Creative Nomad Jukebox wouldn’t have succeeded the way it did.
> The iPod wasn’t a success because of the walled garden. They didn’t have one (once they released on Windows).
My reading of the upthread comments is that people are saying a walled garden isn't necessary, and using the iPod as an example of Apple's success before it had a walled garden. No one is saying that they had a walled garden back then. They did have a built-in audience (Mac users), but that is different from a walled garden. On iOS, they can prevent app devs from accessing their hardware in the same way they can, effectively neutering their efficiency.
Except with gen AI the interface is text and there are multiple companies that have equally acceptable products. Either apple solves hallucination or makes some other AI breakthrough (i doubt it since the best researchers are probably at the hottest AI companies), or else it's just losing out as other companies capture a new market.
I think the second mover will have a huge disadvantage here as companies integrate the first mover into their products.
Apple will be able to do on-device AI, but I can't see that being a game changer unless paired with AR or some other new tech.
Some answers are subjective and rely on justification/argument for support. Provided the argument is sound that's as good as we'll ever get. But it's farcical to claim AI could never always be objectively right.
Apple's real SOP is to make white males believe in the inevitability of Apple cultural dominance so they feel inclined to buy anything with the logo on it as a signal
i.e. you convinced yourself Vision Pro was great before ever using it
I don't 'do' Apple, neither do the vast majority of my friends.
But, having used a borrowed mac product in the past, I do recognize that using them is just... pleasant. Not enough to justify the higher price point, but the superior user experience isn't just marketing.
>Apple customers demographics from 2019 show that about 51% of all iOS users were female, while 43% were male. Moreover, the percentage of female users in the US was 58%. While it’s a small majority, it is the only major smartphone vendor with a majority of female users.
Strage to point out white males in your statement.
It's not the hardware, it's the combination of good hardware and software. That's paraphrasing Steve Jobs. He also stated that one of the primary advantages Apple had was software with better usability, generally higher quality software.
> Modern hardware is responsible for AI
No, again, it's the hardware and software. Tensorflow is what formally sparked the AI boom that we're still rolling on, and of course it needed the underlying hardware.
The hardware does nothing without the software. Python and Linux are good software, to name two.
I still think the iPhone was "safer" than all those others (every single one of those listed had its ardent defenders, mind you) because so many people had iPods and it was "an iPod that can make phone calls" - so there was less risk buying it, especially after the price drop to $399 for the 8 GB (you could pick up 4GB quite cheap, then).
The iPod classic was $349 (for many more songs, admittedly), but the prices were comparable.
> If they can produce a superior product their brand will let them dominate those markets whenever they decide to release.
I anticipate that Apple's Pro devices (iPhone, MacBooks) will have beefier hardware designed specifically for this use case. My guess is they will sell a much higher proportion of Pro iPhones as a result. They'll probably also sell a lot more new devices overall, contra the trend toward slower upgrade cycles.
At the same time, I think third-party companies will find a way to do pretty decent on-device (and very good cloud-based) AI, so it will be a tradeoff in terms of speed and privacy. If you want the best speed and no data being shared with a cloud-based AI provider, the Pro devices will be aimed at you.
> It seems like they are biding their time and avoiding the hype knowing that they don't need to get their first
This idea that Apple gets to things late to "perfect them" is an artifact of the iPod/iPhone era. Nothing Apple put out since then has met this bar.
Apple does things low-key these days because it has to. Remember the Apple Maps fiasco? A few memes was all it took to send everyone scurrying back to GMaps for a decade.
iPad? Apple watch? Slim laptops (involving aircraft metallurgy tricks)?
Apple's been accused of copycatting practically since the beginning. But if you only copycat, you don't typically last this long. They can always offshore somewhere cheaper if you stand still long enough.
I’m not at all convinced about the inevitable success of the Vision Pro (which I maintain is in its current form, a dev kit being marketed as a mass-consumer device), but this just isn’t true.
To use a few examples that have happened after Apple Maps:
* The Apple Watch — Apple was at least 2 years behind other smartwatches when it launched in 2015 and now it basically owns the category. The first version wasn’t quite there but they nailed it with the series 2 and the pivot from fashion to fitness.
* AirPods - tons of companies had been trying truly wireless headphones for years. I know, I had tested basically every option commercially available. Although imperfect at launch (and the software got better over time for once here), AirPods from the moment I first used them in September of 2016 was category defining and the absolute best truly wireless earbuds you could get on the market, at any price point — even tho the quality was mid. Even now, you can do better than standard AirPods for the price/performance, but it’s hard to top AirPods Pro 2s unless you want to spend significantly more money and sacrifice ecosystem niceties (assuming you’re an iOS user. And the Venn diagram between people who are willing to spend over $250 on truly wireless earbuds and iPhone users is basically a circle).
What I do think is instructive about the Apple Maps example is that when it does come to services, and I put generative AI in the services camp much more than hardware, is that Apple is not best-in class at services at all. As a package, Apple services can be beneficial, especially if you’re part of the ecosystem - but even tho I pay Apple $38 a month for Apple One Premier, I’m still paying separately for Spotify, Dropbox, OneDrive (part of Office 365 but still), and Netflix (and a bunch of other video services) and I don’t think I could reasonably drop any of those services in-place of just using iCloud/Apple TV/Apple Music, etc. My parents can (except for Netflix/prime video), and Apple services are definitely “good enough” for a lot of people in the ecosystem, but none are best-in-class.
Unclear if they’d be able to be “good enough” to displace other generative AI services here or not, but to your point, Apple Maps still hasn’t gotten over its bad launch. And Siri on the phone is still dog shit. Siri is so bad that I won’t even use it on platforms where it is objectively good, like Apple TV, unless I specifically remember to.
Fitbit and various other "exercise-aimed" smartwatches were already out way before the Apple Watch, and there were other smartwatches (such as Pebble) - https://en.wikipedia.org/wiki/Smartwatch has a whole listing.
Apple came in quite late and slurped up much of the market. If their watch was Android-compatible, it'd be even more.
They may not have been the first in the category, but those products were arguably a generation behind what apple introduced. Fitbit in particular was like comparing a flip phone to an iPhone.
That's usually what they do (from a small number of examples, to be sure) - come in polished enough that it feels (or is) a generation ahead to those who haven't been watching.
I don't know if they've been able to pull this off with AVP, as most people are just ignoring it.
I'm not sure why you'd draw that conclusion. I overheard a random group of people talking it at the dog park yesterday morning. Then I had breakfast with a friend today who isn't in tech at all, who does corporate conflict resolution, and he asked me if I'd tried a demo. Two other friends, also in no way connected to tech asked me about it and I did a demo to them over FaceTime.
I think if you personally are ignoring it, you probably don't notice how much other people aren't.
That said, I can't think of any of my friends who I would recommend to buy one at this stage, but I see Apple having no trouble selling as many as they can make.
I’m sure they’ll sell, and people know about it, but by “ignoring it” I was referring to the whole virtual reality goggles market, which while it’s ahead of the VirtualBoy just hasn’t really made anything near even “Apple Watch” waves, let alone smartphone waves.
I mean, if we're talking right now -- it absolutely has not. IDC predicts ~12 million AR/VR headset shipments in 2024, which is up 46% from 2023 (which was down 8.3% from 2022) but the Apple Watch, even though it had massive sales declines in 2023, was estimated to have sold 36-38 million units in 2023 [2]. Its wearables category did $39 billion in revenue in fiscal year 2023, which analysts estimate about half is attributable to Apple Watch.
So right now, the entire AR/VR category is expected to ship 1/3 of the number of units of just the Apple Watch. Even if Apple sells 180,000 AVPs all with an MSRP of $4000 (so I'm being extra generous here), that's only $720m in revenue, or less than 3.7% of Apple Watch sales.
And that doesn't even factor in all the Apple Watch units that are in use, which realistically, is probably 125 - 150 million units.
Does that mean AVP or headsets like it won't ever achieve Apple Watch-like sales or adoption? Not at all. But it certainly isn't now and it's probably doing the category a disservice to expect it to reach those heights anytime soon, let alone pretend that just because awareness is super high that sales are guaranteed to follow.
>but I see Apple having no trouble selling as many as they can make.
But right now, supply-chain analysts suggest that they can only make 160,000 or something this year. Apple could sell literal boxes of dog shit to 160,000 people, so I don't think we really know what the size of the market is.(And incidentally, although shipping is now two or three weeks out (an improvement from two weeks ago), I can pick up an AVP from a number of stores in my area today and at another store on Saturday. I don't think we can make any assessments as to the market size/demand right now, but I will say that in-store availability one-week post release was not something we saw for the first iPad or Apple Watch. The iPad 2 was consistently out of stock its first few months, even -- and not because they were supply constrained but because Apple sold every device they got in-store.)
Is interest/intrigue/vauge knowledge super high -- yes -- it's a new Apple product, that's a given. But prodct awareness != product success.
Fitbit was a fitness tracker that later adopted more smartwatch features -- but things like Pebble (early 2013), Android Wear (now Wear OS) (spring 2014), Samsung's Galaxy Gear (fall 2013) all already existed. Now were any of them as good as Apple Watch -- aside from Pebble (which could never have the same level of integration as Apple could offer for iOS, but still did a very good job), I would say no -- but I would also argue Apple Watch was fairly medicore outside of its core notification features at first too.
WatchOS until like fall 2016 really wasn't that great. The third-party apps were all garbage and we didn't even get "native" apps until six months after the device launched and that original Apple Watch (the Series 0) was woefully underpowered for what it needed to be. I bought one and don't regret spending $800 on the stainless steel variant with the modern buckle combo, but it also got like two years of software support, and even the support was limited compared to even the low-cost Series 1 released a year later.
I agree that Apple software has a big hill to climb when it doesn't have the iPhone halo to buttress its bugs.
The Apple Watch has a big advantage: it's one of the cheapest things they sell. It is the iPod Shuffle of its time. But as it doesn't work without an iPhone, I cannot agree that it owns the market.
I don't think Apple Watch owns the smartwatch category, I think it exists in an entirely separate category: a wearable companion to iPhone.
AirPods...I have about 6 different "true wireless headphones". AirPods IMO worse than cheap version of Pixel Buds that google give away for free just for existing. Max and Pro are alright, but the original and 2nd gen AirPods are my least favorite headphones.
Their nice seamless device migration is nice, but only possible because they control all sides of it. Also, doesn't work well if connected to Apple TV for some reason - doesn't automatically switch when you pickup the phone.
I agree with you as someone who had Airpods and still use an Apple Watch.
For Airpods I fell for the hype in a moment of need and weakness: I needed new earbuds for a travel because they couldn't swap my old ones that needed a warranty return. I regreted it pretty fast after the honeymoon period. Sure, they wouldn't drop too much and the 2 buds were perfectly in sync (but that's a low bar to clear for earbuds...) but most of the rest sucked big time. In fact, their poor battery life made me regret the wire pretty fast and their mild sound quality doesn't really help being more forgiving of its shortcomings. Nowadays they make no sense for their price, and the "Pro" version is alright, but to get half decent battery life they had to increased case size wich was basically the only useful difference and in the end, they perform like buds half their price for sure.
I think the iCloud sync bullshit to be unreliable and a stupid "feature" at that because it guarantees that it will only be useful with other Apple products when I definitely do not want to limit my tech to only one brand especially when said brand only sells at luxury pricing... What you really want is Bluetooth multipoint, a feature that many earbuds at this price have anyway.
I bought the Watch to track swimming because it seemed like the thing that would integrate easier with iOS and it seemed not too bad quality tracking. But in the end after years of using an Apple Watch for sports tracking (including thousands kms of running) I don't think the Watch is really good and most will get it to show off or because it's actually the only that can properly integrate with iOS thanks to limitations put in place by Apple themselves. Considering the price, the only really useful features (health/sports tracking) is extremely lacking, especially if you do not spend time looking for 3rd party solutions. The only reason it got such success is because of Apple brand recognition and other secondary factors like social status.
In the end it's not a completely terrible product but if evaluated on its own at its price point its really not great...
Suffise to say I'm rather disenchanted with today's Apple gear and I feel they really need to make their products MUCH better for the price or lower pricing across the board. In the long term something will have to give because people won't be captivated by the brand forever and désillusion has already started.
> What you really want is Bluetooth multipoint, a feature that many earbuds at this price have anyway.
They do, but not as good as within apple ecosystem. My only gripe is how it works together with ATV.
Fitness tracking, I think it's a subset of what I said before: it isn't a fitness tracker, but an iPhone wearable that can be used to track some fitness things. Yeah, Ultra has some niche features, but I got just because it has bigger screen.
I presume they have an SE model in mind for their second iteration. Maybe there's a mothballed project we don't know about, or maybe someone at Apple is tired of rebranding product lines and insisted on making space for both.
Or maybe someone thinks they're managing expectations on price by putting Pro in the name, since it's on par with their other Pro products.
Apple has always been aware that "first mover advantage" is largely bullshit. "The pioneers get all the arrows." The sweet spot is to be second or third to market.
The only real "trickle down" that actually exists, heh. Invent cool stuff, protect yourself from having it stolen, eventually give into the ridiculous sums the giants will throw at you and pivot with that to your next endeavor.
I was not expecting for Apple to be releasing any open source generative ai product. So the big players in the download it yourself gang are Meta & Apple? I don't think anyone would have taken that bet when OpenAI blew up.
Apple makes a ton of sense. They sell hardware (and services for that hardware). It would make sense that they want to commoditize foundation models while at the same time creating yet another reason to buy their most powerful hardware.
Meta is the real wildcard here. They're betting on "commoditize your compliment" and hoping by making GenAI training nearly free (by doing it for you) that none of their competitors can use GenAI models to gain an advantage over them.
Why is Meta the wildcard? Meta has the largest track record of open sourcing everything. From hardware (OpenCompute suite of releases ) to pretty much all software ( HHVM, Hack, Buck, etc etc ) that would benefit wider use.
Even Meta's ML research has a history of open sourcing everything, hello PyTorch. So Llama is just a continuation
Only because all their open source didn't cost them billions to produce. Making a foundation model is at least a billion dollars in just hardware and power.
But you're right they do have a good history of open source, and they should get credit for that.
Historically, Apple doesn't sell hardware, they sell a combination of software and hardware that are supposedly more tightly integrated than is possible when those things are developed separately. It's interesting and surprising that they would develop and release an open source model.
Interesting read, but then suddenly there's this passive aggressive paragraph:
Apple stock has taken a beating as of late, in part because analysts have loudly proclaimed that the company is behind Meta, Google, and Microsoft in generative AI implementation. It's not clear why this wasn't a problem when it wasn't first to a mobile phone, a tablet, a smartwatch, or a VR headset, but is with generative AI.
I also found that an odd statement. If you need a “why” analysts may not feel confident about Apple’s ability to deliver when it comes to AI, look no further than Siri and Apple Maps.
I'll grant you Siri, of course, but why Apple Maps? From my use, it should be held as an example of how Apple can convert a bad launch into a good product.
The significant part of this, for me, is not the image model. It's that Apple rarely publishes in the vision field. Back in Neurips (2016 I want to say?) Apple had a workshop where they said "we're going to engage more with the academic community, we're going to publish more," but since then only a small number of CV projects have seen the academic light of day. I remember one on detection of flashing lights, a whitepaper/tech report on Hey Siri wakeword detection, some interesting synthetic image generation for iris recognition, now this.
If I had to guess, I think Apple Research may have some institutional obstacles to getting research work out there. Perhaps this could indicate those walls are continuing to dissolve bit by bit?
They have at least two recent papers (MobileOne and FastViT) with code and weights for low-latency (as measured on iPhone) vision backbones. Also AIM (Autoregressive Image Models) released last month.
The code repo this article is based on - https://github.com/apple/ml-mgie - is based around InstructPix2Pix (which is based on Stable Diffusion) and LLaVA (multi-modal LLM that supports vision input).
It's research code, not a product. I don't think Apple would ever productionize anything with data lineage from Stable Diffusion.
This is going to be really cool. Being able to have pictures described, with AI, and then edit them using text, like underline this part of the screenshot, or something like that. Especially with Llava 1.6, and Apple's other model where it understands spatial parts of images, this should be pretty possible for me, as a blind person, to do. So yeah I think I'll wait to get this September's iPhone.
I’m curious why so many AI papers consist mostly of people with Chinese-like names. I’m surprised there aren’t more witch hunts from the far right & skepticism that they can embed CCP propaganda into the models.
111 comments
[ 4.2 ms ] story [ 59.4 ms ] threadhttps://instances.vantage.sh/?min_gpus=1 (wait for it to finish loading to filter)
You could try running these in Google Collab (though tweaks would have to be made to load the files), or you might try something like runpod.io, as well, which gets you GPU instances for a lot less than you'd pay with AWS - for example, a Tesla V100 (16GB) community cloud instance runs you $0.24/hr vs the g5g.xlarge which runs you $0.42/hr.
It seems like they are biding their time and avoiding the hype knowing that they don't need to get their first. If they can produce a superior product their brand will let them dominate those markets whenever they decide to release.
My interest in VR is basically solely for gaming, so the Vision Pro for me is a complete miss and I have no plans to purchase one. That being said, in the realm of productivity VR, I'd say they are dominating in that they're the sole entrant to the market that I'm aware of and their offering is extremely compelling.
So, the question is this. If you know anyone who actually got VP. Are they using it obsessively? I got occulus rift a while ago and so did many of my friends. We bought it but never really used it. Like probably a few times in a year to show friends.
I've also used it for reading, watching movies, and reminiscing over old photos. Something I don't think that has really been captured well in reviews is how incredible it is to look at life-size recreations of photos you took on your phone, even plain old 2D ones. There's something really special about seeing a photo at life-size scale wherever you want, even if you technically can achieve this with a massive projector. And that's to say nothing of the bittersweet nostalgic nature of spatial video / photos, which has, if anything, been profoundly understated.
It is not my first headset. Previous to the AVP, I owned Quests 1, 2, and 3, and I experienced the Rift before that. I tried, hard, to work all day in each of the Quests as that was my primary use case for them, but flamed out after a couple of hours every time.
Or, if you are using virtual desktop, you can use your mouse and keyboard to control VP natively as well. The only drawback is that this may not work if you are on VPN (even if the desktop comes through OK). I'm hoping this is a software issue that will get straightened out soon.
On the weight front, it's significantly lighter than a Valve Index, by roughly 20%, and that's just the unit I'm pretty sure, the Index has an umbilical cord that is not-insubstantial in mass by itself that dangles off the back of your head, so already the Vision Pro is quite far ahead on comfort.
VAC is of course a factor in any headset and this really comes down to the per-user situation: I personally got my "VR legs" in about thirty minutes way back when I got started, and am able to use full motion movement pretty well now with only minimal vertigo when things go up and down too fast for my brain's liking. However, the VP is going to have a distinct advantage here with it's big complicated camera system that allows it to operate both in VR and in AR, with most applications seeming to live in the latter realm. I think that's going to lend itself to big comfort gains for people who aren't as blessed as me in that regard.
Isolation also is much improved with the VP, what with it employing again, the camera system, and also the fact that the Vision Pro (much like my Valve Index) doesn't cover or obstruct the ears. And you get those weird CG eyes you can plaster on the front of the thing, haha.
Like again, I don't know if this will work out long term, I'm interested to see if it will really. But I see people keep comparing it to other headsets and as a fervent enjoyer of many of these headsets, I really cannot overemphasize how bad that comparison is.
Wearing a headset / computer on your face is a cost (vs wearing nothing) that the value prop gotta offset. We’ll see if it is a net positive, worth the tradeoff for a significant amount of people. Previous evidence indicates it isn’t.
This might sound trivial but it determines how long the user is able to use the device exclusively without relying on their phone, computer, etc which is a major factor in it being used at all.
A graphics card could make your spreadsheets go faster, but that's not really why they bought them.
It's plausible, however improbable, that a VR headset targeted at Serious Business Reasons could open doors that entertainment has not. And in particular, slightly unwieldy price points that can be rationalized a bit easier.
A decade of crappy Palm, WinCE, and easily forgotten Java based mobile devices came and went before iPhone.
They know what software people refuse to accept; it’s the hardware experience that matters and they’ll wait until the hardware is there.
Modern hardware is responsible for AI, more reliable networks, and the rest of modern compute. Has little to do with the bloated software stacks we git pull into the data center
Edit: Also consider that Intel and Apple don’t just make a consumer device. They design an advanced manufacturing pipeline. The difference between print(“hello world”) and for index in array: print(contents of index)
That’s the basis of their value and why they are propped up by government and why software focused startups will always just be pump and dump schemes.
The first mover advantage gives you a chance to get most of the market share, but with Apple's walled garden, they don't need the first mover advantage to do that.
Other MP3 players existed, but flash media was too small. Other hard drive players existed, they were too big. And both had bad interfaces (by Apple standards).
Apple didn’t show up until they could meet all their goals.
They could have made a phone sooner, it wouldn’t have been the iPhone. They could have made a headset sooner, it wouldn’t be the Vision Pro.
They set their standards for what’s necessary and wait to be able to deliver that.
The first mover rarely wins in the long run.
It's a bit like Debian*, but closed source.
*: It's ready when it's ready.
https://en.wikipedia.org/wiki/Motorola_ROKR
>> Because of the iPod Nano unveiling on the same day, relations between Motorola and Apple were strained and Motorola CEO Ed Zander later accused Apple of purposely undercutting the Rokr.
Apple probably saw it as an opportunity for free iTunes and Nano advertising.
They (whoever ‘they’ is) caused the phone to be terrible even though it had the iTunes name on it. So Jobs got revenge, as he always did if he could.
Perhaps they agreed to a partnership and didn’t realize that Motorola and the carriers would ruin it? Or perhaps it was some sort of “we’ll do this to learn some of the business” kind of a thing?
I’d love to know more of the story behind how it started and it’s development.
Hmm I guess. Vision Pro seems rushed from vantage point. Odd battery pack … only 2 hour charge … limited field of vision. Also the screens that show a picture of your eyes are off putting and strange.
Think Apple may have rushed here too quickly. We need more battery life.. think we may be looking more for an Augmented Reality mode which works on top of what you see with your actual eyes rather than your eyes just looking at high def screens inside of a headset that is relaying a video captured from cameras of what is in front of you.
Looks more like a Prototype IMO
Six hours of connected battery [1] (compared to days for the competition), no copy paste [2], no third-party apps.
[1] https://everymac.com/systems/apple/iphone/iphone-faq/iphone-...
[2] https://9to5mac.com/2022/06/20/apple-engineer-first-iphone-c...
I agree that 3 would be nice, since it would get you through pretty much any movie, but I feel like when I'm watching a movie I don't mind plugging in. After all, I'm not likely moving around a whole bunch. I have USB-C plugs by my bed and near my couches, so it doesn't seem like this is a huge inconvenience (though it would be great if there were a "you're plugged in" warning if you start getting up while you're plugged to a wall).
But if the tradeoff is a heavier battery (or a more expensive device), I'm fine with 2 hours. No hot-swapping, that seems like a bigger miss! People don't want to shut down just to change battery packs.
What I meant is they didn’t need chips as powerful. They didn’t need screens as high resolution or bright. They didn’t need so many cameras.
They could have released something closer to a Quest model in specs years earlier.
They waited for this.
There are always tradeoffs. First gen Apple stuff is often weak in some way (CPU, memory, battery, software) because they like to be at the edge of what’s possible. Then they improve from there.
Apple can afford to be a much later mover, even against very well-resourced competitors, now that they have a huge walled garden. If they didn't have the walled garden, it would be very difficult for them to catch up.
The iPod wasn’t a success because of the walled garden. They didn’t have one (once they released on Windows).
If MS had released a player that was good enough they could have stolen the market, yes. But that’s the point. No one did because everyone released something akin to a flash-based CD player. They didn’t rethink the interface much. The interface is why the iPod grabbed the hold it did.
Yeah it became cool over time.
But I contend an iPod with the interface of a Creative Nomad Jukebox wouldn’t have succeeded the way it did.
My reading of the upthread comments is that people are saying a walled garden isn't necessary, and using the iPod as an example of Apple's success before it had a walled garden. No one is saying that they had a walled garden back then. They did have a built-in audience (Mac users), but that is different from a walled garden. On iOS, they can prevent app devs from accessing their hardware in the same way they can, effectively neutering their efficiency.
Not enough patents, tbh. It's so hard to innovate and scrap things together, that this gets left as an afterthought.
Then the giants move in and steal your lunch, leaving you with nothing.
I think the second mover will have a huge disadvantage here as companies integrate the first mover into their products.
Apple will be able to do on-device AI, but I can't see that being a game changer unless paired with AR or some other new tech.
i.e. you convinced yourself Vision Pro was great before ever using it
But, having used a borrowed mac product in the past, I do recognize that using them is just... pleasant. Not enough to justify the higher price point, but the superior user experience isn't just marketing.
Strage to point out white males in your statement.
[0] https://www.apple.com/apple-vision-pro/
https://web.archive.org/web/20240207021726/https://www.apple...
> Modern hardware is responsible for AI
No, again, it's the hardware and software. Tensorflow is what formally sparked the AI boom that we're still rolling on, and of course it needed the underlying hardware.
The hardware does nothing without the software. Python and Linux are good software, to name two.
Including the Apple Newton, arguably the first handheld device.
The iPod classic was $349 (for many more songs, admittedly), but the prices were comparable.
I anticipate that Apple's Pro devices (iPhone, MacBooks) will have beefier hardware designed specifically for this use case. My guess is they will sell a much higher proportion of Pro iPhones as a result. They'll probably also sell a lot more new devices overall, contra the trend toward slower upgrade cycles.
At the same time, I think third-party companies will find a way to do pretty decent on-device (and very good cloud-based) AI, so it will be a tradeoff in terms of speed and privacy. If you want the best speed and no data being shared with a cloud-based AI provider, the Pro devices will be aimed at you.
This idea that Apple gets to things late to "perfect them" is an artifact of the iPod/iPhone era. Nothing Apple put out since then has met this bar.
Apple does things low-key these days because it has to. Remember the Apple Maps fiasco? A few memes was all it took to send everyone scurrying back to GMaps for a decade.
Apple must have got to him...
Apple's been accused of copycatting practically since the beginning. But if you only copycat, you don't typically last this long. They can always offshore somewhere cheaper if you stand still long enough.
Even chatgpt says they are the best company of all time at marketing.
Also, AFAIK, it won’t say anything specific until you ask it a question, and the away you phrase that can hugely affect its reply.
Even if it didn’t, its response still assumes (more or less) that the Internet is always right, and that is questionable.
To use a few examples that have happened after Apple Maps:
* The Apple Watch — Apple was at least 2 years behind other smartwatches when it launched in 2015 and now it basically owns the category. The first version wasn’t quite there but they nailed it with the series 2 and the pivot from fashion to fitness.
* AirPods - tons of companies had been trying truly wireless headphones for years. I know, I had tested basically every option commercially available. Although imperfect at launch (and the software got better over time for once here), AirPods from the moment I first used them in September of 2016 was category defining and the absolute best truly wireless earbuds you could get on the market, at any price point — even tho the quality was mid. Even now, you can do better than standard AirPods for the price/performance, but it’s hard to top AirPods Pro 2s unless you want to spend significantly more money and sacrifice ecosystem niceties (assuming you’re an iOS user. And the Venn diagram between people who are willing to spend over $250 on truly wireless earbuds and iPhone users is basically a circle).
What I do think is instructive about the Apple Maps example is that when it does come to services, and I put generative AI in the services camp much more than hardware, is that Apple is not best-in class at services at all. As a package, Apple services can be beneficial, especially if you’re part of the ecosystem - but even tho I pay Apple $38 a month for Apple One Premier, I’m still paying separately for Spotify, Dropbox, OneDrive (part of Office 365 but still), and Netflix (and a bunch of other video services) and I don’t think I could reasonably drop any of those services in-place of just using iCloud/Apple TV/Apple Music, etc. My parents can (except for Netflix/prime video), and Apple services are definitely “good enough” for a lot of people in the ecosystem, but none are best-in-class.
Unclear if they’d be able to be “good enough” to displace other generative AI services here or not, but to your point, Apple Maps still hasn’t gotten over its bad launch. And Siri on the phone is still dog shit. Siri is so bad that I won’t even use it on platforms where it is objectively good, like Apple TV, unless I specifically remember to.
Apple came in quite late and slurped up much of the market. If their watch was Android-compatible, it'd be even more.
I don't know if they've been able to pull this off with AVP, as most people are just ignoring it.
I think if you personally are ignoring it, you probably don't notice how much other people aren't.
That said, I can't think of any of my friends who I would recommend to buy one at this stage, but I see Apple having no trouble selling as many as they can make.
So right now, the entire AR/VR category is expected to ship 1/3 of the number of units of just the Apple Watch. Even if Apple sells 180,000 AVPs all with an MSRP of $4000 (so I'm being extra generous here), that's only $720m in revenue, or less than 3.7% of Apple Watch sales.
And that doesn't even factor in all the Apple Watch units that are in use, which realistically, is probably 125 - 150 million units.
Does that mean AVP or headsets like it won't ever achieve Apple Watch-like sales or adoption? Not at all. But it certainly isn't now and it's probably doing the category a disservice to expect it to reach those heights anytime soon, let alone pretend that just because awareness is super high that sales are guaranteed to follow.
[1]: https://www.idc.com/promo/arvr [2]: https://medium.com/@mingchikuo/apple-watch-estimates-and-pre...
But right now, supply-chain analysts suggest that they can only make 160,000 or something this year. Apple could sell literal boxes of dog shit to 160,000 people, so I don't think we really know what the size of the market is.(And incidentally, although shipping is now two or three weeks out (an improvement from two weeks ago), I can pick up an AVP from a number of stores in my area today and at another store on Saturday. I don't think we can make any assessments as to the market size/demand right now, but I will say that in-store availability one-week post release was not something we saw for the first iPad or Apple Watch. The iPad 2 was consistently out of stock its first few months, even -- and not because they were supply constrained but because Apple sold every device they got in-store.)
Is interest/intrigue/vauge knowledge super high -- yes -- it's a new Apple product, that's a given. But prodct awareness != product success.
WatchOS until like fall 2016 really wasn't that great. The third-party apps were all garbage and we didn't even get "native" apps until six months after the device launched and that original Apple Watch (the Series 0) was woefully underpowered for what it needed to be. I bought one and don't regret spending $800 on the stainless steel variant with the modern buckle combo, but it also got like two years of software support, and even the support was limited compared to even the low-cost Series 1 released a year later.
The Apple Watch has a big advantage: it's one of the cheapest things they sell. It is the iPod Shuffle of its time. But as it doesn't work without an iPhone, I cannot agree that it owns the market.
https://www.statista.com/statistics/910862/worldwide-smartwa...
AirPods...I have about 6 different "true wireless headphones". AirPods IMO worse than cheap version of Pixel Buds that google give away for free just for existing. Max and Pro are alright, but the original and 2nd gen AirPods are my least favorite headphones.
Their nice seamless device migration is nice, but only possible because they control all sides of it. Also, doesn't work well if connected to Apple TV for some reason - doesn't automatically switch when you pickup the phone.
For Airpods I fell for the hype in a moment of need and weakness: I needed new earbuds for a travel because they couldn't swap my old ones that needed a warranty return. I regreted it pretty fast after the honeymoon period. Sure, they wouldn't drop too much and the 2 buds were perfectly in sync (but that's a low bar to clear for earbuds...) but most of the rest sucked big time. In fact, their poor battery life made me regret the wire pretty fast and their mild sound quality doesn't really help being more forgiving of its shortcomings. Nowadays they make no sense for their price, and the "Pro" version is alright, but to get half decent battery life they had to increased case size wich was basically the only useful difference and in the end, they perform like buds half their price for sure. I think the iCloud sync bullshit to be unreliable and a stupid "feature" at that because it guarantees that it will only be useful with other Apple products when I definitely do not want to limit my tech to only one brand especially when said brand only sells at luxury pricing... What you really want is Bluetooth multipoint, a feature that many earbuds at this price have anyway.
I bought the Watch to track swimming because it seemed like the thing that would integrate easier with iOS and it seemed not too bad quality tracking. But in the end after years of using an Apple Watch for sports tracking (including thousands kms of running) I don't think the Watch is really good and most will get it to show off or because it's actually the only that can properly integrate with iOS thanks to limitations put in place by Apple themselves. Considering the price, the only really useful features (health/sports tracking) is extremely lacking, especially if you do not spend time looking for 3rd party solutions. The only reason it got such success is because of Apple brand recognition and other secondary factors like social status. In the end it's not a completely terrible product but if evaluated on its own at its price point its really not great...
Suffise to say I'm rather disenchanted with today's Apple gear and I feel they really need to make their products MUCH better for the price or lower pricing across the board. In the long term something will have to give because people won't be captivated by the brand forever and désillusion has already started.
They do, but not as good as within apple ecosystem. My only gripe is how it works together with ATV.
Fitness tracking, I think it's a subset of what I said before: it isn't a fitness tracker, but an iPhone wearable that can be used to track some fitness things. Yeah, Ultra has some niche features, but I got just because it has bigger screen.
I also wonder why they broke with their naming scheme ("Pro" is not the Pro of any smaller model)
Or maybe someone thinks they're managing expectations on price by putting Pro in the name, since it's on par with their other Pro products.
https://apple.fandom.com/wiki/AppleVision_Display
Spend a few thousand dollars to get provisionals. Then if your idea works and grows, file the full patents.
Don't let the giants steal your blood, sweat, and tears finding the gradients.
Meta is the real wildcard here. They're betting on "commoditize your compliment" and hoping by making GenAI training nearly free (by doing it for you) that none of their competitors can use GenAI models to gain an advantage over them.
Even Meta's ML research has a history of open sourcing everything, hello PyTorch. So Llama is just a continuation
But you're right they do have a good history of open source, and they should get credit for that.
They don't want to be left behind with a "everyone with an Nvidia GPU can play" world. Open-source is the best counter to that right now.
https://huggingface.co/spaces/tsujuifu/ml-mgie
(Long queue, I could not test it)
Apple stock has taken a beating as of late, in part because analysts have loudly proclaimed that the company is behind Meta, Google, and Microsoft in generative AI implementation. It's not clear why this wasn't a problem when it wasn't first to a mobile phone, a tablet, a smartwatch, or a VR headset, but is with generative AI.
somebody is hurting
If you are only looking at objective metrics, Apple is always going to perform at B and C levels.
If you are looking at fanaticism, they are A++.
If I had to guess, I think Apple Research may have some institutional obstacles to getting research work out there. Perhaps this could indicate those walls are continuing to dissolve bit by bit?
It's research code, not a product. I don't think Apple would ever productionize anything with data lineage from Stable Diffusion.