One thing about AR glasses that I don't get so far: Why do ALL of the AR glasses use only 1080p displays per eye whereas all the other VR solutions that also have a display right in front of your eye have even bigger resolutions?
I really do want AR glasses that can act as an alternative to a physical display. But they're always 1080p for some reason. The companies I've seen making them appear to me to keep fixing important issues and adding really useful features - but seem to avoid touching anything that would improve the resolution.
Edit : There's a part of me that believes the situation here is a lot like how Linus Tech Tips described the Dashcam situation.
The display panels used for AR(and some pre-Oculus VR) displays are completely different thing from normal LCDs. Lots of them used to be a reflective panel made of silicon wafers with RGB front lights flashing in sequence. Driving circuits and interfacing protocols are all non-standard, under NDA, and quite complicated. The panels also needs to be coupled to a complicated optical lens - they're just magnifying glasses, only fancier. The complete thing that has a digital input and an optical output is often referred to as an Optical Engine.
What that means is, these glasses are made by someone paying $$$ once to Asian engineering companies to have it figured out, and everyone reuse that exact setup for years on.
And companies like Sony or Canon, they develop stuffs by scraping engineers off a wall instead of first throwing them at the wall, so every Sony or Canon cameras and projectors come with novel viewfinders and crazy patented lenses. Or panel suppliers like Kopin, Epson, or Sharp can arrange contacts to engineering consultants. I guess. Same likely goes for all the Chinese companies too, though I'm not familiar enough on that front to be able to offer googlable keywords.
Startups and even mid-sized consumer electronics companies don't have that kind of time or financial backings or margins to do the same. They barely manage to pay for assembly and ship it. And so the spec of the final product is whatever spec of parts that they could buy off the shelf.
What about Apple? Well, they pay to have 2 extra display factories built for redundancy by policy - I'm sure that most military organization don't do that. And even then they use a rather simplistic, completely concentric and rotationally symmetric optical design.
(I kind of have a crazy idea to bypass some of that, but unfortunately I'm crazy and so is the idea)
It's depressing but not surprising when you remember that that only reason any of this exists is because Lucky Palmer pointed out that you can kickstart modern VR R&D with off-the-shelf smartphone displays (IIRC).
In my experience you need at least 8k per eye to get sharp text. But if you just want to watch a movie 4k will look very crisp and 1080p@30hz also works for movies.
So what you get is basically a larger mobile screen for watching videos, something that competes with video projectors, but you can use it on the bus. It's an easier sell compared to a device that will give you a good reading experience and competes with paper books.
Those of us that want to use them as terminal emulators are still too few.
Not disagreeing with the resolution but I assume for most last-gen use cases it didn't make much sense because you also need to be able to drive that somehow and you can be happy that you get a 2 Lane display port interface up until the more recent phones.
So if you have high res displays that no one can realistically use for the intended mobile cinema use case then you are wasting money and energy
The screens are two entirely different products. The current Sony micro-OLED screens being used by product manufacturers for AR glasses require out-of-plane reprojection versus VR headset screens which require focal lenses.
The top-of-the-line solutions by Sony used in current products only go up to 1080p120.
I don't know what I would do with higher resolution personally. This is already enough that xterm's "tiny" font is extremely hard to read. Personally I'd prefer lower power draw to more resolution.
One of the reason is IMHO that the optical resolution with those lense designs does not get better than 1080p. It seems even that resolution is blurry at the edges. So anything beyond probably would be a waste without totally different optics.
* afaik I'm capped at 4 FPS (just how often you can update tabs)
* more of the screen is likely to update on each frame, which requires way more "turning canvases into data urls to update favicons," which is expensive with my current implementation (although I wasn't aware of offscreencanvas, which might help here)
I think you can reduce state. Rather than tracking maxElementReached per-element, maintain a single maxElementReached for the first n elements. March the first n elements forward in lockstep, and grow n by 1 whenever you exhaust all available combinations for that set
1. Combine the first element with every next element until exhausted.
2. Catch up the second element to where the first element got to.
3. Combine the first two elements with every next element, until exhausted.
4. Catch up the third element.
5. Combine the first three elements with every next element
6. etc.
In pseudocode...
n = 1
maxElementReached = -1
while(n < totalElements()) {
while(maxElementReached + 1 < totalElements()) {
maxElementReached = maxElementReached + 1
Combine each of the first n elements with element[maxElementReached]
}
// we've exhausted all possible combinations for the first n elements.
// increase n by 1 and catch up the new element to keep going
Combine element[n] with each element from n to maxElementReached
n = n + 1
}
minified version courtesy of GPT-4
(disclaimer I have no clue how this works)
let m={},t=0,f=0,s=0;setInterval(function(){document.getElementsByClassName('mobile-item')[f].getElementsByClassName('item')[0].click();document.getElementsByClassName('mobile-item')[s].getElementsByClassName('item')[0].click();t=document.getElementsByClassName('mobile-item').length;s=(s+1)%t;if(s==0){m[f]=t;if(Object.keys(m).some(i=>m[i]<t)){let p=Object.keys(m).find(i=>m[i]<t);f=p;s=m[p];}else{f=(f+1)%t;s=f;}}document.title=f+'+'+s+'|'+t;},500);
theoretically their concern is around AI safety - whatever it is in practice doing something like that would instantly signal to everyone that they are the bad guys and confirm everyone's belief that this was just a power grab
Edit: since it's being brought up in thread they claimed they closed sourced it because of safety. It was a big controversial thing and they stood by it so it's not exactly easy to backtrack
Not sure how that would make them the bad guys. Doesn't their original mission say it's meant to benefit everybody? Open sourcing it fits that a lot better than handing it all to Microsoft.
All of their messaging, Ilya's especially, has always been that the forefront of AI development needs to be done by a company in order to benefit humanity. He's been very vocal about how important the gap between open source and OpenAI's abilities is, so that OpenAI can continue to align the AI with 'love for humanity'.
It benefits humanity. Where humanity is very selective part of OpenAI investors.
But yea, declare we are non-profit and after closing sourcing for "safety" reasons is smart. Wondering how can it be even legal. Ah, these "non-profits".
I can read the words, but I have no idea what you mean by them. Do you mean that he says that in order to benefit humanity, AI research needs to be done by private (and therefore monopolising) company? That seems like a really weird thing to say. Except maybe for people who believe all private profit-driven capitalism is inherently good for everybody (which is probably a common view in SV).
the view -- as presented to me by friends in the space but not at OpenAI itself -- is something like "AGI is dangerous, but inevitable. we, the passionate idealists, can organize to make sure it develops with minimal risk."
at first that meant the opposite of monopolization: flood the world with limited AIs (GPT 1/2) so that society has time to adapt (and so that no one entity develops asymmetric capabilities they can wield against other humans). with GPT-3 the implementation of that mission began shifting toward worry about AI itself, or about how unrestricted access to it would allow smaller bad actors (terrorists, or even just some teenager going through a depressive episode) to be an existential threat to humanity. if that's your view, then open models are incompatible.
whether you buy that view or not, it kinda seems like the people in that camp just got outmanuevered. as a passionate idealist in other areas of tech, the way this is happening is not good. OpenAI had a mission statement. M$ manuevered to co-opt that mission, the CEO may or may not have understood as much while steering the company, and now a mass of employees is wanting to leave when the board steps in to re-align the company with its stated mission. whether or not you agree with the mission: how can i ever join an organization with a for-the-public-good type of mission i do agree with, without worrying that it will be co-opted by the familiar power structures?
the closest (still distant) parallel i can find: Raspberry Pi Foundation took funding from ARM: is the clock ticking to when RPi loses its mission in a similar manner? or does something else prevent that (maybe it's possible to have a mission-driven tech organization so long as the space is uncompetitive?)
Exactly. It seems to me that a company is exactly the wrong vehicle for this. Because a company will be drawn to profit and look for a way to make money of it, rather than developing and managing it according to this ideology. Companies are rarely ideological, and usually simply amoral profit-seekers.
But they probably allowed this to get derailed far too long ago to do anything about it now.
Sounds like their only options are:
a) Structure in a way Microsoft likes and give them the tech
b) Give Microsoft the tech in a different way
c) Disband the company, throw away the tech, and let Microsoft hire everybody who created the tech so they can recreate it.
No, that's backwards. Remember that these guys are all convinced that AI is too dangerous to be made public at all. The whole beef that led to them blowing up the company was feeling like OpenAI was productizing and making it available too fast. If that's your concern then you neither open source your work nor make it available via an API, you just sit on it and release papers.
Not coincidentally, exactly what Google Brain, DeepMind, FAIR etc were doing up until OpenAI decided to ignore that trust-like agreement and let people use it.
They claimed they closed sourced it because of safety. If they go back on that they'd have to explain why the board went along with a lie of that scale, and they'd have to justify why all the concerns they claimed about the tech falling in the wrong hands were actually fake and why it was ok that the board signed off on that for so long
What would that give them? GPT is their only real asset, and companies like Meta try to commoditize that asset.
GPT is cool and whatnot, but for a big tech company it's just a matter of dollars and some time to replicate it. Real value is in push things forward towards what comes next after GPT. GPT3/4 itself is not a multibillion dollar business.
If I can get this to sort my huge mess of photos based on geolocation and time into folders for me based on my own folder and naming structure... This just might be useful.
Edit : even older folks who have no idea how tech works might be able to finally sort their photos as well.
That sounds very helpful, and if AI can do it, then that's great. But also, it seems like rule based file restructuring should already be easily possible without AI. Shame something like that doesn't exist.
In Apple-land, it's done that for ages: group via face, put name to face, group by geolocation etc. I think it's started recognizing my dogs too. Happens all device side.
Google photos does amazing work with this. It's face people detection is so good it recognises baby photos all the way to adulthood as the same person.
yeah. I have spent two weeks hashing thousands of pictures scattered around the informal backups and now I still have to deal with different size versions of the same image.
I wish the computer could rename them consistently in my stead and removed the lesser quality versions. I would prefer if it was a tool developed by anyone else but Microsoft.
Short circuit the problem by not taking so many photos. Ask yourself honestly, how often do you ever look at them? If you do, you've probably also got them organized. If you don't, then you're just hoarding.
Out of curiosity... Does anyone use the rss reader in MS Outlook? I know MS is supposed to pack a lot of features to catch as much audience as possible, but are there enough people still using it to this day?
I used thunderbird like that a lot but everytime I needed to read or send an email for something productive I'd see that there's a new post somewhere and I'd check it.
It was actually a pretty decent thing if you set up a dedicated account (I used to maintain rss2mail and https://github.com/rcarmo/rss2imap).
I still have archived posts in MIME format, complete with images, etc. Amazing for archival purposes.
The tricky thing is actually dealing with title-only feeds - these days apps like Reeder will fetch a readable version of posts automatically, and I never got around to implementing that in earnest.
I used it for a few years to access rss feeds from non-public internet routable jira servers. It was pretty nice to monitor new issues across lots of projects.
One thing about AR glasses that I don't get so far: Why do ALL of the AR glasses use only 1080p displays per eye whereas all the other VR solutions that also have a display right in front of your eye have even bigger resolutions?
I really do want AR glasses that can act as an alternative to a physical display. But they're always 1080p for some reason. The companies I've seen making them appear to me to keep fixing important issues and adding really useful features - but seem to avoid touching anything that would improve the resolution.
Edit : There's a part of me that believes the situation here is a lot like how Linus Tech Tips described the Dashcam situation.