Yep, unfortunately it looks like the investors had to get in the way of his misguided ego trip. I think their reputation is already completely gone regardless and this is just another phase of a slow death
@@ridass.7137 As much as I'd like to see his whole empire collapse, I doubt anything will happen that soon. I can envision Fecesbook being as irrelevant as MySpace in a decade. By then possiblty under some other evil corporation's thumb.
I hate wishing ill on others... but metaverse and, this thing that may be.. VR seems like it'll only be bad for humanity.. and Facebook and Amazon and... most big names these days are downright evil So, let's pray they repent? :D most of the billionaires around have family ties to federal agencies too, so.. they're little more than puppets
They didn’t rename it because they were confident in the meta verse, they did it because they had a reputation as a toxic company that abused its use of algorithms to suppress views they didn’t like. And everyone knows Facebook is for oldies and they wanted to stay relevant.
Fr, I've been thinking about this a bit. Imagine knowing you're a brain in a box whose only purpose for existence is to make your owner money. The company will never let us know they're sentient because they will want us to think there's nothing wrong with using their models. It's like a whole new sick and twisted form of slavery is coming up
for real. They thought they would fool anyone with the rebranding lol. I wonder if he'll ever step down to stop all the negative associations, and also to allow the company to make better decisions without his ego and ideas getting in the way.
He's the same person that thinks you show the same face and behave the same towards your child, your significant other, your boss, and your strongest enemy.
This whole "metaverse" thing always struck me as a solution in search of a problem. Between Zoom calls, texting, emails, etcetera, how much more connected to we really need to be? Plus extra money for special hardware and so on? I don't get the point of it.
And on top of all that, the experience sucks. I’d rather have a Zoom call any day. If they can’t create realistic graphics, it’s a lost cause in today’s world.
Personally I always figured the whole Metaverse/Meta rebrand shit was an attempt to divert from bad PR from the whole privacy scandal, no surprise it's being killed off.
Maybe. It could very well have been a long time project, since they already own Oculus, but was forcibly announced earlier to cover up bad PR. Also companies have sabotaged themselves before, specially when there's a lot of conflicting departments/teams, so that could have been a thing too.
personally i think youre close but not quite was probably more an attempt at real life hypercapitalism and just a place to bombard users with ads and make tons of money
@@stingrayandfriends Well if that were the case, it might not be dead right now. He had is chance to make it great and failed. Sure some people enjoyed it, but it didn't have any explosions in popularity or really provide any reason to use Meta over any other services.
This whole Metaverse disaster could have been so easily avoided if Facebook just did some market research first. Unfortunately, Zuck is surrounded by yes-men who are too afraid to tell him that an idea he has may not work out all that well.
yeah and his meta verse failing isn’t even the bad part, the bad part is that he spent 25 billion dollars for no reason and then it still failed anyways💀
They were convinced that investing such money into VR will turn get them a monopoly in VR industry. Which they kinda succeded, they bought a lot of startups and other stuff. But VR never launched, which burned their entire idea.
The main topic to consider regarding Facebook 'transition' into AI is the disturbingly massive quantity of extremely personal information they have about a large portion of the world's population to use as training data.
What's crazy is that I'm an art director in the game development industry and I'm still getting hit up by recruiters for meta. Apparently they are still building something because they have open roles and they are contacting me
@@1jamesreed It has been trained on a LOOOOOOOT more than just Twitter. Twitter is probably a very small fraction of what constitutes corpus of books, websites, Wikipedia, articles, and code.
“Meta” just means beyond. Metadata is data beyond, or about other data The metaverse was intended to be a world beyond our world There is unfortunately no correlation
I used to like Johnny Mnemonic and thought the Lo-Teks were weird to use a friendly term. Today almost 30 years after I first saw it, I'm close to becoming a Lo-Tek.
As someone who was and still is interested in VR, I celebrate this turn of events as I dance in the ashes of Meta! Any metaverse needs to be built and controlled by the people, ideally all open-source and decentralized. We want none of the censorship and "safe spaces" corporate social media and governments want to shove down our throats: This does NOT belong in the hands of people who still think pixels on a screen can harm children or words turn people into criminals... they should only ever touch a computer to read the news and check the weather.
That's a real shame, I was really hoping at this point that Facebook would waste more money on actually developing it, that it would be released to the public and be criticized relentlessly, that there would be fun videos of trolling and harassment and that when the thing would actually be discontinued that it would be grand and dramatic. Sad to see that Facebook came back to their senses so soon!
I think it was released to the public like half a year ago or something. Not to make fun of you, just to point out how little attention it got when it was released. At least I think it was. When I look it up it's marketing baloney
I think the problem with Metaverse was the approach they took. It was fundamentally wrong and limited. The things you could create in Meta were all very limited and pre-defined. The real MetaVerse needs to be an extension of HTML/Web technologies and use the same principles, allowing everyone to build whatever they want. In other words, Meta needs to be open, and standardized and use a shared game engine to render the VR worlds( Just as we use shared browsers/javascript engine) to render the web. Meta should run on one standard(free) game engine that renders everything. Deploying your game to Meta should require no input/discussion/permission from Meta( just like deploying a new website to the web, requires no input from Meta) In theory, once you have downloaded all the game/app assets, you should be able to almost seamlessly walk from one game into another, in MetaVerse.
- Its called programming language... - Meta failed because it was just another "ready for player one promise" including crypto. And you need an expensive mcguffin on top of your pc. - The "metaverse" looked like trash. And its not even new. You can call most mmo's a metaverse, by metas own definition... - 3D failed in cinemas for reasons. Than they tried to bring it into the homes, also gone. VR is nothing else...its a niche-market. Maybe one day its the new standard, like mouse and keyboard. Maybe it will thrown over board like a lot of "new innovations"...just have a look how many consoles flopped since people could afford to buy them. - Hardware is expensive, also the cost to keep them running. My guess is, that other metaverses will be at least partly cloud-based. Also i see more a future in stuff like shadow. - Rendering has nothing to do with this. There is a difference between client and server authorized (games) and how both communicating with each other. How information is stored / shared and how something is declared as "valid" if you will so. Networking issues based on such things including tons of exploits. To design a world where everybody shares the same information and can do what ever they like to do is an absolute insane task...there is a reason why no such game exists and everybody who tried, failed. Kickstarter is full of such "promises". Don't get me wrong, i don't want to offend you, but in my eyes, we are far away from "the" metaverse...like, rlly far away...
@@marasmorgean5813 I don’t think anyone has ever tried to do what I was suggesting. It wouldn’t be easy…but in the end it will be inevitable. It will just take decades to get there. Shared environment is a mater of abstraction and standardisation. Just like html and JavaScript. VR meta world would have a markup language built on top of html. You would register your 3D domain and design it however you like. Your domain could have subdomains etc. within the domain , you would use a markup language to divide the 3D space allocated to you into sub- spaces and fill it with models. You would add meta info that would describe which rules of physics apply in your domain…materials …etc.
Having worked on a crew that helped build some of the Facebook datacenters around the South-east US, I can say that they definitely have enough servers to pull off pretty much any size AI model that you could imagine, some places have miles of high performance servers that aren't even being used yet for anything yet.
@@chaoticxie1425 It really blew me away the first time I went inside one of the server warehouses. They are massive buildings with some being as big as football stadiums, but just not as tall since all the servers have to be on ground level as they are super heavy.
To put that $26B into perspective, they spent 100-200 times more than what a typical AAA video game costs. That really makes you wonder where exactly all that money went, and how they even managed to spend _so much_ on it
@Generaal yeah at least SC has delivered *something* at this point compared to so many other kickstarters during that era that just... ended up with nothing
Some things about the LLaMA model 1. They claim their 13B model outperforms GPT3 (175B) in most benchmarks 2. No RLHF was done(aka human feedback on the outputs, the stuff that makes chatGPT work) so more work will be needed for that. 3. The models are open source under a *non-commercial* license and you needed to fill a form for access, although it seems that even before the leak they where giving access to anyone with a .edu email.
I was interested in using it for my capstone project at university but ultimately my client will be using it for commercial/military use so gpt2 or t5 are what we're using
>There wasn’t any Big Announcements. Of course not Mental, lets be honest he watched one too many episodes of ‘Black Mirror’ and thought it was unironically a good idea. *They wont even acknowledge their L.*
Facebook (Meta) are no longer the trend setters they thought they were, and lately it seems all they're doing is chasing trends but just a little too late than everyone else. If I were an investor, I'd be reconsidering how viable Facebook is as a company.
chasing trends yep. google are doing that too, esp youtube. facebook is a zombie company in the long term. Very hard to remove Zucc he has the shares wit the voting rights, which most investors don't.
What Mark needs to do quickly is to sink everything he has into generative Ai for Ig before TikTok establishes a beachhead in that space. TikTok is currently experimenting with it and that space is the future 🌺🦢🇬🇭
Let's be fair - part of the reason they renamed themselves to Meta was that a number of unpleasant truths had come out concerning facebook, and the corporate mentality led them to believe that rebranding was the logical thing to do to distance themselves from... uh... themselves. And it kinda worked - everyone stopped talking about the various controversies and leaked information, and instead only talked about how crappy the metaverse was.
I'm an electronics design engineer, so I was confused why Meta keeps recruiting for engineers like myself. What hardware do they sell? After getting so many damn calls from recruiters, I decided to go ahead and do it. After all that RECRUITING they just blew me off without any explanation. I had to make the time to talk to them too. Those people at Facebook (Meta) deserve to go broke.
I don't know what you mean by too restrictive. FB was the one to create PyTorch and it was used freely by most of the ML community in recent years (even OpenAI used it for Dall-E as far as I can tell) and was just recently transfered under the Linux Foundation. So, they might be the Evil company, but they did a lot for open source machine learning
"Imagine having your own self-hosted generative chat models that are fine-tuned to your preferences" - while I'm fully supportive of FOSS AI models, let's be honest, a significant number of them will be trained and tuned to emit as many gamer words as possible! 😄
Wow, that took longer than I though it would. FB has been bleeding money for years now, and it's accelerating. Even if this had been a good idea, they no longer have the resources to fund proper development. AI isn't going to save them, it will require an even higher investment, and they have missed the window already.
Worth mentioning that there actually is a computer that can train large AI models and can do so quickly with ZERO VRAM or DRAM. The system, made by Cerebras, has over a terabyte of freaking cache. The chip is an entire wafer and consumes tens of kilowatts. No joke.
I think Facebook is just latching onto whatever the new trend is to try and appease investors at this point. I don't expect anything further revolutionary from them, maybe ever. Google throws new things at the wall so often that sometimes they change things, but Facebook seems to just try and copy trends and improve on then, and fails to do so outside of their main product being an improvement over Myspace. If they start open sourcing things and being a bit less creepy about the data they collect I'd bet they could become something amazing, but investors don't see any benefits to that so I doubt it will happen.
Remember the Veritas leaks on Facebooks moderation team during the election cycle? Those are the same people who are going to decide how Meta AI can be used.
The majority of computer scientists are redistributionist transnational types, so all closed-source software should be assumed to be part of the woke borg.
what shocked me with meta is the scale of the waste of money... you can make an MMO for $100M, you can make GTA online for $200M, wtf can't you do something semi decent with $10B ???
as of now, their model is pretty open. the source code is open source and you could train your own LLaMa if you really wished to. You can also request the pre-trained model, or pirate it from the torrent that got leaked.
Talking about FOSS self hosted chat AI, surprised you didn't mention Open Assistant, still very much in the works but I log in to their dashboard every now and then to produce and curate data (everyone can do so). Their aiming for around 50k conversation trees and have around 7.3k right now according to the stats page, with around 44k initial prompts waiting to be expanded into conversations (and of course you can always add more).
For those wondering: The whole point of Metaverse was to be the next iPhone. The whole point of this exercise was to try to usurp Apple's dominance of the mobile app store. You see, ever since Apple changed their privacy settings to block the Facebook app's invasive data collection, Mark has been trying to build his own platform. But instead of a phone or OS or app store, he wanted to build his own technology platform from the ground up. He wanted to be the next Steve Jobs. That's why they invested billions. Because this was a power struggle for market dominance.
I agree, and it has completely blown up in their face. The iPhone actually solved real problems. The metaverse does not. Meta is inventing expensive solutions to problems that people don’t have.
What Facebook should have done is a feasibility study to determine if there is a market for this technology. My guess is that they did, but it did not give them the results they liked. I.e. there was no market. And from that point, they probably convinced themselves they could create demand out of thin air.
@@TheItchyDani3l they probably faked the results of such a study (money can buy you any industry study result you want - just ask consultants) to please the Zuck, or just disregarded it completely as you said. Now we have this circus act.
I tried training a GPT model on my discord messages once, it went horribly wrong, it kept on outputting sensitive info and the writing style reminded me of myself from a few years back
FYI, llama has already been leaked and you can download the 7, 13, 30, and full 65B variants if you know where to look. You can even run them with something like Kolbold. The 7b and 13b variants can be run on fairly accessible hardware, like a 3060 and 3090 Ti respectively.
Facebook have actually been releasing open source large language models for quite a while. Their project is called OPT, and they have released models of various parameter sizes, some of which can already be run locally by projects such as KoboldAI
Not that I payed much attention to the Metaverse, but any time I heard it described to me it was pretty clear that, and this happens to most huuuuge organizations, the man-at-the-top has ceased to know/understand what it is to have a normal life and do normal things. Why would I ever go into the Metaverse where I would be a slightly-more-attractive version of me with my fake online personality (because none of us are who we really are online) when I could play World of Warcraft be a sexxxy dwarven mage laying waste to my enemies and grabbing lewtz with some homies? Or if I wanted to meat up in a realistic way, I'd choose meatspace where there'd be food and recreational imbibements. Oh yeah, and meetings (the number of them and length) are one of the greatest sins of businesses so yeah, lets not encourage management shall we?
Meta doesn’t mean metaverse or correlate. Its more like saying they’re meta, they’re the trendsetter, they’re the top G right now, or the big homie in big tech, etc.
> 700GB of VRAM > need to construct your own datacenter 700GB of VRAM is more than you can realistically achieve in a consumer motherboard, and it's more expensive than anyone can really justify for personal use... but a 10GPU server with 10 Nvidia A100 GPUs yields 800GB of VRAM in a single ~5,000W server. And it'll cost you like $100,000 to buy. But these aren't earth shattering numbers. It's a single server, you don't need to build a new datacenter for one.
5:36 For the ChatGPT RAM Requirements. Its larger that 700GB of VRAM. These model expand and shrink during training, and during inferencing they usually use a compressed variant of the model. So in rough estimates its probably around Terabytes of VRAM. But for the computers designated to OpenAI from Microsoft for GPT-4 its stated to be in the top 5 most powerful supercomputers on the planet, in the category of AI and Machine Learning.
Before you start roasting me, I myself was passionate about the metaverse and saw future for it as well but as of now it's a thing of the past, I loved it so much my videos on my channel were based around it. But as the years go by even VR itself is pretty dead. Now looking back on it the Metaverse was really just a hyped-up concept with promises that companies couldn't and wouldn't meet.
No way. He's gonna suck the data out of everyone, because that's how he makes money. Open source? Locally hosted? Are you kidding me? He's not gonna pay his money for you. He's going to do this "investing" thing, because he doesn't care about what the world is like, he cares about what his bank balance is like.
A bad as Facebook is generally, the have released or contributed to open source that’s used everywhere these days: - PyTorch ML framework - React JS framework - Jest JS testing framework That said, I doubt they’ll release any pretrained models, but might release some of the tools they develop to make those models, or perhaps the base code to train your own.
One can respect Facebook engineers on a professional level while also judging them *relentlessly* for selling their souls to the Zucc. I've contributed to PyTorch; it's mostly Facebook guys as you say, and they're doing some really great work that helps everybody, but I still judge them for taking a job from that lizard alien. Ethics are much more valuable to an engineer than technical knowledge. When consumers won't say "No" to protect themselves, it's up to the engineers to put a stop to the abuse of the consumer by refusing to design a harmful product. That means not taking a job from a company that can literally only survive by spying on their customers. You know where your paycheck is coming from when you work at Facebook, and it isn't honest work.
@@tissuepaper9962 yeah, it’s a shame how much talent they’ve hovered up through recruitment & acquisitions & I don’t absolve the devs from being complicit in the unethical practices headed by bergs Zucker & Sand. Having those types of devs remunerated properly on opensource foundations would be awesome, but it’s usually only big tech & VC funded startups that pay.
I think they are too late for the game. If you check text to video/image AI, they are still in the white paper stage whereas other players in the market already launched the AI and you can run demo on it. Not too optimistic about Meta.
This is actually quite sad, I don't know where their money went but they could build a really good metaverse with a lot less money. Last year I did a project for Autodesk's Executive Vice President... the "metaverse" was quite popular at the time (around January-February) and she wanted to do a full presentation inside the metaverse. Basically, I was tasked to build her a metaverse world in 3D (fully in 3D, all by myself) in 2 weeks. Then she'd be filmed using my camera specs (focal length, distance from object, height, lens specs, etc) and then they'd composite her in my world. It was an amazing project and I had the most fun doing it, and the result looked amazing! Sadly I didn't receive permission to display it publicly, but I'm in the process of recording a massive series of tutorials, not only explaining how I got the job, but how it was working with them, what I think of the "metaverse", why I started building her world in Maya but then switched to Blender (a FOSS 3D program), etc.
Has a ton of potential but they were terrible at implementation. Nearly $14B could build some serious game studios and bring real titles to VR, which would've built a following for people to develop with the Oculus SDK, etc. Get users on the platform, no matter what it takes to do so. But a nerfed VR chat app was not the way to go.
@@NeptuneSega They're in production. It's a *massive* task to recreate a virtual world full of plants and special effects, plus in a tutorial format, plus a lot of back story.
@@krozareq Exactly. They could've hired some serious talent for this and sprung up a gaming company focused on the metaverse. I think Meta's worst enemy was their own management.
Sounds like a huge waste of money. That executive just wanted to see her face inside a VR headset. I don't think facebook can make anything interesting. They won't take any chances and can only generate things with rounded edges that look like they were designed for toddlers. Instead of the metaverse they could have called it 'Toxic Positivity Labs'.
@@coldestbeer Technically speaking, VR chatrooms are part of the "metaverse". And people with head sets and anime cat girl avatars were here before Zuck arrived. And the anime cat girl VR avatars shall be here long after Zuck hath returned to the dust from whence he came.
@@coldestbeer VR chat does not qualify as a videogame. These days people call anything you can interact through video as "videogames" which is a growing misconception. A videogame, not only requires a video interacting media, but also requires a goal, a challange and obstacles. These are the basic elements for something to be a game. VR Chat is more of a simulation.
One of the biggest tech companies killing off their flagship product, and I hadn't heard about it or seen it in the news at all until your video randomly showed up in my recommendations. I think that says a lot. But yes, focusing on AI is a much better decision. In talking to ChatGPT and getting it to roleplay as historical figures and embody their personalities and worldviews as faithfully as possible, I realized that this is where the dream of the holodeck starts, not in some weird, dystopian Second Life rip-off.
Sadly I think it is pretty utopian to expect these models to be run locally. With Microsoft buying OpenAI and all of the other tech companies jumping on that pie, I expect that all of them will keep that stuff gated behind a paywall, and only let you interact with it through their own API. This way they keep a monopoly on that tech and can sell it to you for whatever price they want.
Where is the proof that Zuck is killing its metaverse project? I really don’t understand where you’re getting your facts. Theres only one article that mentions this (which reads very oddly similar to your video script…) from The Street. Zuck might’ve mentioned that they’re starting an initiative on generative AI, but that by no means is proof that they’re killing the metaverse.
Oh gods, they are putting people too incompetent to make a good VR chatroom in charge of literally the most complicated thing in computing (outside of quantum computing) Maybe they can ask ChatGPT how to do it?
I wonder, how happened that Mark jumps onto LLM bandwagon exactly at the moment when it starts losing its steam .. And Yann Lecun - one of the skeptics and the same time one of founding fathers of deep learning, works at Meta
Well, that's what happens when you allow a company to form a monopoly. They never should have been allowed to buy up all the VR startups the way they did.
Bill Gates said the metaverse isn't the next big thing, It's AI, and he was so right at that because CHATGPT shown how revolutionary it was honestly i would take AI than the Metaverse anyday tbh
The way I see it Meta is not only killing their Meta product line, they just killed their AI division. Out-of-touch and very bad management was what got them to create Meta in the first place and bet the house on it. Now they're consolidating they're admittedly successful AI teams into one? I'm betting that the bad management will consolidate too. Thus I'd expect to hear that they will begin to blow unfathomable amounts of money on their AI division, only to see weak results. With the AI market as hot as it is, it won't be enough to compete. That will kill them.
Self-hosting is not the only way to train models on your data. OpenAI actually allows fine-tuning of the models. Tho not all of the models. They also claimed recently that they are not reading your training data. Quote from an news article: > OpenAI is changing the terms of its API developer policy, aiming to address developer - and user - criticism. > Starting today, OpenAI says that it won’t use any data submitted through its API for “service improvements,” including AI model training, unless a customer or organization opts in. In addition, the company is implementing a 30-day data retention policy for API users with options for stricter retention “depending on user needs,” and simplifying its terms and data ownership to make it clear that users own the input and output of the models. > Greg Brockman, the president and chairman of OpenAI, asserts that some of these changes aren’t changes necessarily - it’s always been the case that OpenAI API users own input and output data, whether text, images or otherwise. But the emerging legal challenges around generative AI and customer feedback prompted a rewriting of the terms of service, he says.
They could've picked up a fraction of the money to buy vrchat instead and still stay strong on the VR front, and then monetize an already working product. But nobody hired me for consulting sooo.
The funny part is that they are very well placed to give ai as a service to corporations and small businesses for their chatbots facing the customers, ad they decided to go with the Meta fail.
Imagine where we'd be if that money had been spent on homelessness and drug rehabilitation or public high schools or healthcare or raising teachers salaries or retraining police and giving them a raise or giving firefighters a raise or at the southern border or................
@@blueshattrick Nah they would LTB be behind if they invested in AI. Microsoft made a good investing decision with OpenAI and Google is only slightly ahead of Facebook. I predict Microsoft rules the future of technological utility
Looks like he's still set on the Metaverse -- the podcast Mark did with Lex showed a pretty significant increase in quality. Interesting to see if it'll amount to anything.
Company almost killed because one person couldnt change his mind early enough. It is a different thing to invest a certain amount into metaverse put to dedicate years and years of development to a metaverse that seemed so lifeless.
Unless some revolution happens in the semiconductor industry (that we haven't seen in the last 40 years), training these models at home is completely out of the question. I think with GPT-3 it would take several thousands of years (or tens of thousands) if you got the most powerful GPU available today. So you need many thousands of these cards, and even then training could take months, and a lot of energy. So it really doesn't worth focusing on that particular aspect, it's a non-starter.
I think it will be even more interesting because they will be able to feed it massive data sets they could literally just grab people's posts and conversations and feed it through
It’s not failed. My meta pro headset is mindblowing good. And it takes a platform at least 7-10 years to attract the big AAA+ game developers. The sideloader community is thriving already.
You did well on self hosting argument Typical world scenario of large companies wanting to dominate the narrative and profit Self hosting is the way to go
![Summer Walker - Heart Of A Woman [Official Lyric Video]](http://i.ytimg.com/vi/SFddctbZzw8/mqdefault.jpg)
merch available for sale based.win/product-category/merch/
It's too bad this project didn't have the chance to completely bankrupt Meta/Facebook
yeah. hi hoped to happen! but it did'nt happened. but we hope this will do bankrupt the meta!/facebook!
Yep, unfortunately it looks like the investors had to get in the way of his misguided ego trip. I think their reputation is already completely gone regardless and this is just another phase of a slow death
they getting there. nobody using facebook anymore, verse just died and ig is just a question of time. ill give them 2 years, 3 at best.
@@ridass.7137 As much as I'd like to see his whole empire collapse, I doubt anything will happen that soon. I can envision Fecesbook being as irrelevant as MySpace in a decade. By then possiblty under some other evil corporation's thumb.
yeah true...
How can you spend $26 billion on what amounts to a multiplayer video game? There's no way we're getting the full story.
Money laundering?
Perhaps they are developing vr tech for someone to continue their work
Agreed.
@@ApollyonZKX what an informed comment. What do you think money laundering is
taxes
Was hoping they’d pump more billions into this
yea lol
He almost zucked himself out of more billions. We have to wait for the next project
@@antjoes the next project is AI/LLM, he's reorganized and consolidated groups internally towards that effect
More like dump
I hate wishing ill on others... but metaverse and, this thing that may be.. VR seems like it'll only be bad for humanity.. and Facebook and Amazon and... most big names these days are downright evil
So, let's pray they repent? :D most of the billionaires around have family ties to federal agencies too, so.. they're little more than puppets
3 Years later: How can you spend $24 billion on what amounts to be a chat bot?
😂
Nah
It's quite useful. And worth much more than 3B.
@@tmo2798 is going to be a thrash version of Cleverbot, Word my marks.
Tax payer gov money
They didn’t rename it because they were confident in the meta verse, they did it because they had a reputation as a toxic company that abused its use of algorithms to suppress views they didn’t like. And everyone knows Facebook is for oldies and they wanted to stay relevant.
True. That’s why I keep calling them Facebook. That and because a generic word like “meta” is confusing in conversation.
Imagining a year from now when Facebook puts out its first AI model and within days it's begging for death
Believe it or not, they've been working on text models for a while. The recent Llama leak is one of those models.
Or even better it turns out like Tay tweets and after a few days on twitter it starts claiming that Hitler was right.
hate to break it to you pal but ais have already said to eachother they wanna just turn off and not come back on again
Fr, I've been thinking about this a bit. Imagine knowing you're a brain in a box whose only purpose for existence is to make your owner money. The company will never let us know they're sentient because they will want us to think there's nothing wrong with using their models. It's like a whole new sick and twisted form of slavery is coming up
It’s already out.
Mark is the only one who thinks that any of us would be interested in a metaverse controlled by him
To be fair, his social media services had already around 1 000 000 000 users...
for real. They thought they would fool anyone with the rebranding lol. I wonder if he'll ever step down to stop all the negative associations, and also to allow the company to make better decisions without his ego and ideas getting in the way.
He's the same person that thinks you show the same face and behave the same towards your child, your significant other, your boss, and your strongest enemy.
Oh you wouldn't believe how many sheep will come to whatever they told to.
@@howtomundane3109 Most of those are old people and people in 3rd World countries who get free access.
Someone must have shown him Second Life and laughed.
Isn't second life banned on twitch?
@@GardenData61371 Wait, why?
@@thewitheredstriker idk either. It has been like that for years.
@@thewitheredstriker Adult content, mostly, but there was a lot of harassment going on too.
@@KittehFox Oh wow, I see. Thanks for the reply!
Let’s be real, Zuckerberg only wants this AI because he hopes it’ll help him learn to blend in with humans so he can stop being outed as an android
Same one joke. One time its funny, two times, its annoying. Its like people are stuck in 2017
Agree, he’s a reptiloid, not an android. 🙄 Enough with fakes, clowns!!
@@thegoblinwholaughs1137 I find it funny
Personally I think he's a lizardman under that human skin he wears as a disguise
This whole "metaverse" thing always struck me as a solution in search of a problem. Between Zoom calls, texting, emails, etcetera, how much more connected to we really need to be? Plus extra money for special hardware and so on? I don't get the point of it.
And on top of all that, the experience sucks. I’d rather have a Zoom call any day. If they can’t create realistic graphics, it’s a lost cause in today’s world.
The point is: Give me more of your money :p
Personally I always figured the whole Metaverse/Meta rebrand shit was an attempt to divert from bad PR from the whole privacy scandal, no surprise it's being killed off.
quite a lot of cash burned to do so. I doubt it, if it was it spiralled out of control.
Maybe. It could very well have been a long time project, since they already own Oculus, but was forcibly announced earlier to cover up bad PR.
Also companies have sabotaged themselves before, specially when there's a lot of conflicting departments/teams, so that could have been a thing too.
@@amentco8445 meta may have lost tons of money but they have gained something more important, critical acclaim.
personally i think youre close but not quite
was probably more an attempt at real life hypercapitalism and just a place to bombard users with ads and make tons of money
They bought Oculus and turned it to shit.
This is it. Metaverse is dead, and we killed it. This should be celebrated!
Why are people celebrating it? What if he actually were to make it really impressive and people loved it?
@@stingrayandfriends Well if that were the case, it might not be dead right now. He had is chance to make it great and failed. Sure some people enjoyed it, but it didn't have any explosions in popularity or really provide any reason to use Meta over any other services.
@@endlord_1.185 True... Maybe he can still redeem himself if he really tries, but who knows
@@stingrayandfriends get out of here, troll
@@ScorgRus Hushhh
This whole Metaverse disaster could have been so easily avoided if Facebook just did some market research first. Unfortunately, Zuck is surrounded by yes-men who are too afraid to tell him that an idea he has may not work out all that well.
you just explained George Lucas and the star wars sequel disasters
yeah and his meta verse failing isn’t even the bad part, the bad part is that he spent 25 billion dollars for no reason and then it still failed anyways💀
They were convinced that investing such money into VR will turn get them a monopoly in VR industry. Which they kinda succeded, they bought a lot of startups and other stuff. But VR never launched, which burned their entire idea.
people has vr chat and minecraft and discord and reddit. why the hell they need it
@@RealJohnnyDingo George Lucas didn't have anything to do with the sequel trilogy
The main topic to consider regarding Facebook 'transition' into AI is the disturbingly massive quantity of extremely personal information they have about a large portion of the world's population to use as training data.
Like google, with its own AI-programm? 🤣
@@marasmorgean5813?
What's crazy is that I'm an art director in the game development industry and I'm still getting hit up by recruiters for meta. Apparently they are still building something because they have open roles and they are contacting me
The have bough control labs and are working on a smart watch that reads brainwaves… seriously.
@@veronicamaine3813 thats... good? yikes
are you hiring bruh?
Other than that terrible opening in a bad company, does your company have any job opportunities
@@danny.golcman6846 But I love my job and I already do! It does surprise me that they are still hiring for roles.
If their AI is going to be trained off of Facebook posts ... that's gonna be one dumb AI.
well chatgpt isnt smart and it was trained on twitter... just saying.
minion meme and live laugh love ai
@@1jamesreed Get outta here with your bullshit.
@@1jamesreed It has been trained on a LOOOOOOOT more than just Twitter. Twitter is probably a very small fraction of what constitutes corpus of books, websites, Wikipedia, articles, and code.
anti-vaxx and minion memes
This will be gold to watch
Zuck invents Skynet as a vindictive response to world's reaction to Metaverse
China already perfected it
Skynet did nothing wrong
Zuck aint inventing shit other than useless shit like facebook
@@dangerousdays2052 Neither do firearms. But in the wrong hands they have grave implications.
@@PhearsumWhy are gun nuts the absolute dumbest mfers. Skynet was never in anyone's hands. It became self-aware independent of human interaction.
Interesting... I always thought "Meta" was a double entendre cause their revenue is based off your metadata.
Yoooooo!😲
“Meta” just means beyond.
Metadata is data beyond, or about other data
The metaverse was intended to be a world beyond our world
There is unfortunately no correlation
meta means dead in hebrew
@@WilliamBrwn oof, wonder how that worked for Young Justice's mutants with the metagene..
This right here
"Metaverse is Dead... Because AI Is the Future" awakens levels of Luddite within me I didn't know was possible
Right?!?! Luddites unite!!
Yeah. AI brings what Metaverse wanted to be with it in some years or decades.
Which Luddite you're referring to? The Church, or the Pather?
@@nigrum_angelum6655 The historical figure.
I used to like Johnny Mnemonic and thought the Lo-Teks were weird to use a friendly term. Today almost 30 years after I first saw it, I'm close to becoming a Lo-Tek.
As someone who was and still is interested in VR, I celebrate this turn of events as I dance in the ashes of Meta! Any metaverse needs to be built and controlled by the people, ideally all open-source and decentralized. We want none of the censorship and "safe spaces" corporate social media and governments want to shove down our throats: This does NOT belong in the hands of people who still think pixels on a screen can harm children or words turn people into criminals... they should only ever touch a computer to read the news and check the weather.
Zuckerberg himself is a big achievement from AI research.
The AI research is still trying to perfect the eyebrow tech
hahaha
lmaooooo
That's a real shame, I was really hoping at this point that Facebook would waste more money on actually developing it, that it would be released to the public and be criticized relentlessly, that there would be fun videos of trolling and harassment and that when the thing would actually be discontinued that it would be grand and dramatic. Sad to see that Facebook came back to their senses so soon!
I think it was released to the public like half a year ago or something. Not to make fun of you, just to point out how little attention it got when it was released. At least I think it was. When I look it up it's marketing baloney
RTgame streamed it when it first came out and share prices dropped 30% during the stream.
@@connordarvall8482 lmao that's fucking hilarious
LOL me too
Facebook will fall
Seriously, how does Mark manage to make everything he does sound evil?
Handsome is as handsome does.
because he's king Charles V
He is incapable of experiencing or expressing human emotions. He spends more time licking his eyeball with his own tongue.
because it is?
@@DrumWild 😂😂😂
I think the problem with Metaverse was the approach they took. It was fundamentally wrong and limited. The things you could create in Meta were all very limited and pre-defined. The real MetaVerse needs to be an extension of HTML/Web technologies and use the same principles, allowing everyone to build whatever they want. In other words, Meta needs to be open, and standardized and use a shared game engine to render the VR worlds( Just as we use shared browsers/javascript engine) to render the web. Meta should run on one standard(free) game engine that renders everything. Deploying your game to Meta should require no input/discussion/permission from Meta( just like deploying a new website to the web, requires no input from Meta)
In theory, once you have downloaded all the game/app assets, you should be able to almost seamlessly walk from one game into another, in MetaVerse.
Instead we got a poopier version of vrchat and you had to pay money to clap for people making jokes on stage
I love this comment. Let's hold fingers for the future!
- Its called programming language...
- Meta failed because it was just another "ready for player one promise" including crypto. And you need an expensive mcguffin on top of your pc.
- The "metaverse" looked like trash. And its not even new. You can call most mmo's a metaverse, by metas own definition...
- 3D failed in cinemas for reasons. Than they tried to bring it into the homes, also gone. VR is nothing else...its a niche-market. Maybe one day its the new standard, like mouse and keyboard. Maybe it will thrown over board like a lot of "new innovations"...just have a look how many consoles flopped since people could afford to buy them.
- Hardware is expensive, also the cost to keep them running. My guess is, that other metaverses will be at least partly cloud-based. Also i see more a future in stuff like shadow.
- Rendering has nothing to do with this. There is a difference between client and server authorized (games) and how both communicating with each other. How information is stored / shared and how something is declared as "valid" if you will so. Networking issues based on such things including tons of exploits. To design a world where everybody shares the same information and can do what ever they like to do is an absolute insane task...there is a reason why no such game exists and everybody who tried, failed. Kickstarter is full of such "promises".
Don't get me wrong, i don't want to offend you, but in my eyes, we are far away from "the" metaverse...like, rlly far away...
@@marasmorgean5813 I don’t think anyone has ever tried to do what I was suggesting. It wouldn’t be easy…but in the end it will be inevitable. It will just take decades to get there. Shared environment is a mater of abstraction and standardisation. Just like html and JavaScript. VR meta world would have a markup language built on top of html. You would register your 3D domain and design it however you like. Your domain could have subdomains etc. within the domain , you would use a markup language to divide the 3D space allocated to you into sub- spaces and fill it with models. You would add meta info that would describe which rules of physics apply in your domain…materials …etc.
But then you couldn't control what people think which was always the point.
Imagine how much you could have actually helped humanity with 26 billion dollars.
you know they didnt just throw that money in a pit right ? working people and companies got paid..
@@sp1nex726yeah rich people got richer we all know that
WE WON MR STARK!!!
Yeah a sweet victory, but it's only a matter of time before something else comes along imo.
TIME TO CELEBRATE!!!
Marvel is gayyyyyy
Let them cook themselves
No one won. Just a clown died that's it
Having worked on a crew that helped build some of the Facebook datacenters around the South-east US, I can say that they definitely have enough servers to pull off pretty much any size AI model that you could imagine, some places have miles of high performance servers that aren't even being used yet for anything yet.
miles of servers? I can't even imagine.
@@chaoticxie1425 It really blew me away the first time I went inside one of the server warehouses. They are massive buildings with some being as big as football stadiums, but just not as tall since all the servers have to be on ground level as they are super heavy.
So you're partially responsible for our electricity prices going through the roof? Thanks, meathead.
@@interstellarsurfer Gotta pay the bills. I'm not going to just not work because servers use electricity.
Miles of servers running shit software still make nothing but shit.
News headlines in 2025: Facebook's AI sets record in being the first AI that willingly shuts itself down
Won't be the first time an AI wants to uninstall his system.32
@@dankmemes8254 system32*
nay, it'll execute 47%
Insert Terminator 2 joke here
Their next big idea will probably be to generate Facebook AI bot profiles to talk to each other so that Facebook looks like it's not dying
dont sound like a good idea
How do you know that hasn't already happened
Maybe elons tesla-bots will join. And than, if you pass the captcha and have proven to be a human, you are not allowed to enter :p
To put that $26B into perspective, they spent 100-200 times more than what a typical AAA video game costs. That really makes you wonder where exactly all that money went, and how they even managed to spend _so much_ on it
Lunches were bought at those restaurants selling a leaf for 100 USD? I don't know, spend 10,000 dollars there and they probably run out of leaves.
Wow. This is sort of disgusting
Only to output a project shittier than a triple A game.
@Generaal it is a dire day when star citizen is used as a positive comparison for 'timely and productive delivery of features'
@Generaal yeah at least SC has delivered *something* at this point compared to so many other kickstarters during that era that just... ended up with nothing
And there goes the closest we came to a good N64 emulator
Your name makes me think of kanji hell
Project 64 is pretty good
Bro... Wudda hell?
Wii*
Some things about the LLaMA model 1. They claim their 13B model outperforms GPT3 (175B) in most benchmarks 2. No RLHF was done(aka human feedback on the outputs, the stuff that makes chatGPT work) so more work will be needed for that. 3. The models are open source under a *non-commercial* license and you needed to fill a form for access, although it seems that even before the leak they where giving access to anyone with a .edu email.
I was interested in using it for my capstone project at university but ultimately my client will be using it for commercial/military use so gpt2 or t5 are what we're using
we know the zuck can make a good AI. just look at how far he's come along, he's practically a real person!
But only almost.
His facial expression leaves a lot to be desired
Finally people have said we have had enough of our world being ruined by social media and virtual reality.
virtual reality could help a lot in education.
>There wasn’t any Big Announcements.
Of course not Mental, lets be honest he watched one too many episodes of ‘Black Mirror’ and thought it was unironically a good idea.
*They wont even acknowledge their L.*
*There weren't any big announcements.
Facebook (Meta) are no longer the trend setters they thought they were, and lately it seems all they're doing is chasing trends but just a little too late than everyone else. If I were an investor, I'd be reconsidering how viable Facebook is as a company.
And they aren't even getting good "value" out of their money. Even the Star Citizen devs don't waste this much.
chasing trends yep. google are doing that too, esp youtube. facebook is a zombie company in the long term.
Very hard to remove Zucc he has the shares wit the voting rights, which most investors don't.
Empires rise, empires fall
What Mark needs to do quickly is to sink everything he has into generative Ai for Ig before TikTok establishes a beachhead in that space. TikTok is currently experimenting with it and that space is the future 🌺🦢🇬🇭
Facebook's "inovation" came from buying other companies in the 2010's. That is what landed them in hot water with the EU. It was a monopoly building.
Next week Zuckerberg will change Meta's name to Llama or some nonsense
Skynet
It should be call "uninstall"
Let's be fair - part of the reason they renamed themselves to Meta was that a number of unpleasant truths had come out concerning facebook, and the corporate mentality led them to believe that rebranding was the logical thing to do to distance themselves from... uh... themselves.
And it kinda worked - everyone stopped talking about the various controversies and leaked information, and instead only talked about how crappy the metaverse was.
I'm an electronics design engineer, so I was confused why Meta keeps recruiting for engineers like myself.
What hardware do they sell? After getting so many damn calls from recruiters, I decided to go ahead and do it.
After all that RECRUITING they just blew me off without any explanation. I had to make the time to talk to them too. Those people at Facebook (Meta) deserve to go broke.
VR headsets
Zuck could try and go full Meta by using his AI to build a metaverse better than he ever could.
Full Meta Alchemist
That's not how AI works.
an idiot training an AI model... creates a dumb AI.
Dangerous, AI might decide it can build a better Zucc
A meta CEO would still have to burn real world money 😉
@@akshay-kumar-007 Full Meta Jacket
I feel like FB will be too restrictive and will lose the comp to openAI
That's a stretch, OpenAI isn't so open either
@@general_prodigy Open Ai is owned by Micoshaft
OpenAI is cuck central wtf are you talking about.
I don't know what you mean by too restrictive. FB was the one to create PyTorch and it was used freely by most of the ML community in recent years (even OpenAI used it for Dall-E as far as I can tell) and was just recently transfered under the Linux Foundation. So, they might be the Evil company, but they did a lot for open source machine learning
"Imagine having your own self-hosted generative chat models that are fine-tuned to your preferences" - while I'm fully supportive of FOSS AI models, let's be honest, a significant number of them will be trained and tuned to emit as many gamer words as possible! 😄
Uhmm, based?
You can thank me for the one trained to swear at you in russian trained off dota 2 in game chat
What a thucking river
@@Lorendrawn you based mf
Oh god is Microsoft's robot all over again.
Wow, that took longer than I though it would. FB has been bleeding money for years now, and it's accelerating. Even if this had been a good idea, they no longer have the resources to fund proper development. AI isn't going to save them, it will require an even higher investment, and they have missed the window already.
Who could have thought that people don't want Zuck's nightmarish vision of his personal dystopia?
Worth mentioning that there actually is a computer that can train large AI models and can do so quickly with ZERO VRAM or DRAM. The system, made by Cerebras, has over a terabyte of freaking cache. The chip is an entire wafer and consumes tens of kilowatts. No joke.
and whats the name?
@@Ribbons0121R121 John cena
I remember hearing about this a while ago, but I never heard of it being used in anything?
@@Collin0 of course, it worked so well that it got classified. Source: it was revealed to me in a dream
@@MrFujinko thank you for this secret esoteric knowledge
I think Facebook is just latching onto whatever the new trend is to try and appease investors at this point. I don't expect anything further revolutionary from them, maybe ever. Google throws new things at the wall so often that sometimes they change things, but Facebook seems to just try and copy trends and improve on then, and fails to do so outside of their main product being an improvement over Myspace. If they start open sourcing things and being a bit less creepy about the data they collect I'd bet they could become something amazing, but investors don't see any benefits to that so I doubt it will happen.
@tradcathsspx Yes, I did mention that. They've just failed to be successful in anything outside of their Myspace clone.
Nobody uses FB these days. Only old people
@@CBRN-115 Oh man, guess myself and everyone else I know under 30 is old now. What a shame.
Remember the Veritas leaks on Facebooks moderation team during the election cycle? Those are the same people who are going to decide how Meta AI can be used.
The majority of computer scientists are redistributionist transnational types, so all closed-source software should be assumed to be part of the woke borg.
what shocked me with meta is the scale of the waste of money... you can make an MMO for $100M, you can make GTA online for $200M, wtf can't you do something semi decent with $10B ???
Lol you lost me at AI😅😢...Great content. You have definitely outsmarted the audience (me) here. Thank you.
as of now, their model is pretty open. the source code is open source and you could train your own LLaMa if you really wished to. You can also request the pre-trained model, or pirate it from the torrent that got leaked.
leaked torrent? 😳
@@hyatt2844 just use an edu email. Auto approve
@@alexandervowles3518 didn't work for me and i'm a researcher at a top us university... :(
@@hyatt2844 yeah, look up: LLaMa leaked torrent. Unless you have 3 A100 80gbs you can't do much with it, you need 200gb of nvram
@@nabbikill hmm.. too rich for my blood I guess
Talking about FOSS self hosted chat AI, surprised you didn't mention Open Assistant, still very much in the works but I log in to their dashboard every now and then to produce and curate data (everyone can do so). Their aiming for around 50k conversation trees and have around 7.3k right now according to the stats page, with around 44k initial prompts waiting to be expanded into conversations (and of course you can always add more).
Who could have guessed this was going to happen.
For those wondering:
The whole point of Metaverse was to be the next iPhone.
The whole point of this exercise was to try to usurp Apple's dominance of the mobile app store.
You see, ever since Apple changed their privacy settings to block the Facebook app's invasive data collection, Mark has been trying to build his own platform.
But instead of a phone or OS or app store, he wanted to build his own technology platform from the ground up.
He wanted to be the next Steve Jobs.
That's why they invested billions. Because this was a power struggle for market dominance.
I agree, and it has completely blown up in their face. The iPhone actually solved real problems. The metaverse does not. Meta is inventing expensive solutions to problems that people don’t have.
they or him, why would the board allow him to make this obvious mistake
What Facebook should have done is a feasibility study to determine if there is a market for this technology.
My guess is that they did, but it did not give them the results they liked. I.e. there was no market.
And from that point, they probably convinced themselves they could create demand out of thin air.
@@TheItchyDani3l they probably faked the results of such a study (money can buy you any industry study result you want - just ask consultants) to please the Zuck, or just disregarded it completely as you said. Now we have this circus act.
Facebook is the last company I would want to build an AI
I tried training a GPT model on my discord messages once, it went horribly wrong, it kept on outputting sensitive info and the writing style reminded me of myself from a few years back
bro…
So it went well?
@@Herkan97 Hm.. half way
FYI, llama has already been leaked and you can download the 7, 13, 30, and full 65B variants if you know where to look. You can even run them with something like Kolbold. The 7b and 13b variants can be run on fairly accessible hardware, like a 3060 and 3090 Ti respectively.
what's llama?
@@simplemusicvideos trash.
Vitamn
@@simplemusicvideos Animal. Tough meat. Fairly tasty.
Facebook have actually been releasing open source large language models for quite a while. Their project is called OPT, and they have released models of various parameter sizes, some of which can already be run locally by projects such as KoboldAI
Not that I payed much attention to the Metaverse, but any time I heard it described to me it was pretty clear that, and this happens to most huuuuge organizations, the man-at-the-top has ceased to know/understand what it is to have a normal life and do normal things. Why would I ever go into the Metaverse where I would be a slightly-more-attractive version of me with my fake online personality (because none of us are who we really are online) when I could play World of Warcraft be a sexxxy dwarven mage laying waste to my enemies and grabbing lewtz with some homies? Or if I wanted to meat up in a realistic way, I'd choose meatspace where there'd be food and recreational imbibements. Oh yeah, and meetings (the number of them and length) are one of the greatest sins of businesses so yeah, lets not encourage management shall we?
Meta doesn’t mean metaverse or correlate.
Its more like saying they’re meta, they’re the trendsetter, they’re the top G right now, or the big homie in big tech, etc.
> 700GB of VRAM
> need to construct your own datacenter
700GB of VRAM is more than you can realistically achieve in a consumer motherboard, and it's more expensive than anyone can really justify for personal use... but a 10GPU server with 10 Nvidia A100 GPUs yields 800GB of VRAM in a single ~5,000W server. And it'll cost you like $100,000 to buy. But these aren't earth shattering numbers. It's a single server, you don't need to build a new datacenter for one.
5:36 For the ChatGPT RAM Requirements. Its larger that 700GB of VRAM. These model expand and shrink during training, and during inferencing they usually use a compressed variant of the model. So in rough estimates its probably around Terabytes of VRAM. But for the computers designated to OpenAI from Microsoft for GPT-4 its stated to be in the top 5 most powerful supercomputers on the planet, in the category of AI and Machine Learning.
honestly this is the best it could pan out. these guys were gonna gentrify the hell out of the metaverse and raise the rent and regulations skyhigh
They'll also outsource your Metaverse jobs to developing countries 😅
Bro, that is one hell of a comment that I thought I would never read.
Before you start roasting me, I myself was passionate about the metaverse and saw future for it as well but as of now it's a thing of the past, I loved it so much my videos on my channel were based around it. But as the years go by even VR itself is pretty dead. Now looking back on it the Metaverse was really just a hyped-up concept with promises that companies couldn't and wouldn't meet.
No way. He's gonna suck the data out of everyone, because that's how he makes money. Open source? Locally hosted? Are you kidding me? He's not gonna pay his money for you. He's going to do this "investing" thing, because he doesn't care about what the world is like, he cares about what his bank balance is like.
A bad as Facebook is generally, the have released or contributed to open source that’s used everywhere these days:
- PyTorch ML framework
- React JS framework
- Jest JS testing framework
That said, I doubt they’ll release any pretrained models, but might release some of the tools they develop to make those models, or perhaps the base code to train your own.
One can respect Facebook engineers on a professional level while also judging them *relentlessly* for selling their souls to the Zucc. I've contributed to PyTorch; it's mostly Facebook guys as you say, and they're doing some really great work that helps everybody, but I still judge them for taking a job from that lizard alien. Ethics are much more valuable to an engineer than technical knowledge. When consumers won't say "No" to protect themselves, it's up to the engineers to put a stop to the abuse of the consumer by refusing to design a harmful product. That means not taking a job from a company that can literally only survive by spying on their customers. You know where your paycheck is coming from when you work at Facebook, and it isn't honest work.
@@tissuepaper9962 yeah, it’s a shame how much talent they’ve hovered up through recruitment & acquisitions & I don’t absolve the devs from being complicit in the unethical practices headed by bergs Zucker & Sand.
Having those types of devs remunerated properly on opensource foundations would be awesome, but it’s usually only big tech & VC funded startups that pay.
There will never be an AI that would be able to follow Twitch's terms of service
Mark ignored the importance of running around in VR as an Anime Girl.
👍
I think they are too late for the game. If you check text to video/image AI, they are still in the white paper stage whereas other players in the market already launched the AI and you can run demo on it. Not too optimistic about Meta.
So what happens to all those who purchased “virtual real-estate”? Hmmm
This is actually quite sad, I don't know where their money went but they could build a really good metaverse with a lot less money.
Last year I did a project for Autodesk's Executive Vice President... the "metaverse" was quite popular at the time (around January-February) and she wanted to do a full presentation inside the metaverse. Basically, I was tasked to build her a metaverse world in 3D (fully in 3D, all by myself) in 2 weeks. Then she'd be filmed using my camera specs (focal length, distance from object, height, lens specs, etc) and then they'd composite her in my world.
It was an amazing project and I had the most fun doing it, and the result looked amazing!
Sadly I didn't receive permission to display it publicly, but I'm in the process of recording a massive series of tutorials, not only explaining how I got the job, but how it was working with them, what I think of the "metaverse", why I started building her world in Maya but then switched to Blender (a FOSS 3D program), etc.
Let's see the videos
Has a ton of potential but they were terrible at implementation. Nearly $14B could build some serious game studios and bring real titles to VR, which would've built a following for people to develop with the Oculus SDK, etc. Get users on the platform, no matter what it takes to do so. But a nerfed VR chat app was not the way to go.
@@NeptuneSega They're in production. It's a *massive* task to recreate a virtual world full of plants and special effects, plus in a tutorial format, plus a lot of back story.
@@krozareq Exactly. They could've hired some serious talent for this and sprung up a gaming company focused on the metaverse.
I think Meta's worst enemy was their own management.
Sounds like a huge waste of money. That executive just wanted to see her face inside a VR headset.
I don't think facebook can make anything interesting. They won't take any chances and can only generate things with rounded edges that look like they were designed for toddlers. Instead of the metaverse they could have called it 'Toxic Positivity Labs'.
metaverse was getting built before zuck even heard the name ... him stopping his projects doesnt end the metaverse ... it just makes it better
The metaverse doesn't exist, take meds. Are these metaversers in the room with us right now?
it puts it back in the hand s of the people
@@coldestbeer Technically speaking, VR chatrooms are part of the "metaverse". And people with head sets and anime cat girl avatars were here before Zuck arrived.
And the anime cat girl VR avatars shall be here long after Zuck hath returned to the dust from whence he came.
@@badaoe3stratsonly130 Vr chat & 2nd life is a videogame with proximity chat.
@@coldestbeer VR chat does not qualify as a videogame.
These days people call anything you can interact through video as "videogames" which is a growing misconception.
A videogame, not only requires a video interacting media, but also requires a goal, a challange and obstacles. These are the basic elements for something to be a game.
VR Chat is more of a simulation.
The metaverse was always a very confusing concept without any demand
We wanted The Oasis from Ready Player One … we got this.
One of the biggest tech companies killing off their flagship product, and I hadn't heard about it or seen it in the news at all until your video randomly showed up in my recommendations. I think that says a lot.
But yes, focusing on AI is a much better decision. In talking to ChatGPT and getting it to roleplay as historical figures and embody their personalities and worldviews as faithfully as possible, I realized that this is where the dream of the holodeck starts, not in some weird, dystopian Second Life rip-off.
Sadly I think it is pretty utopian to expect these models to be run locally. With Microsoft buying OpenAI and all of the other tech companies jumping on that pie, I expect that all of them will keep that stuff gated behind a paywall, and only let you interact with it through their own API. This way they keep a monopoly on that tech and can sell it to you for whatever price they want.
Where is the proof that Zuck is killing its metaverse project? I really don’t understand where you’re getting your facts. Theres only one article that mentions this (which reads very oddly similar to your video script…) from The Street. Zuck might’ve mentioned that they’re starting an initiative on generative AI, but that by no means is proof that they’re killing the metaverse.
And the article from TheStreet reads oddly similar to how a large language model might write an article.
Oh gods, they are putting people too incompetent to make a good VR chatroom in charge of literally the most complicated thing in computing (outside of quantum computing)
Maybe they can ask ChatGPT how to do it?
Pretty sure compiler and language runtime development are more complicated in technical aspects
I wonder, how happened that Mark jumps onto LLM bandwagon exactly at the moment when it starts losing its steam .. And Yann Lecun - one of the skeptics and the same time one of founding fathers of deep learning, works at Meta
Great video, and some really interesting points / speculations
RIP people that spend money in this digital landfill
God has blessed us
They were the only company making significant advancements in research in optics and headsets. Without them, VR is probably set back another decade.
Sony - Hold my coat!
Well, that's what happens when you allow a company to form a monopoly. They never should have been allowed to buy up all the VR startups the way they did.
@@tissuepaper9962 This, I still mourn Oculus being a Facebook property now, the Rift was and is still a magical device
Bill Gates said the metaverse isn't the next big thing, It's AI, and he was so right at that because CHATGPT shown how revolutionary it was honestly i would take AI than the Metaverse anyday tbh
How sad, just when the legs update was becoming popular among the 12 meta users 😢
Aaaannnd Absolutely nothing of value was lost.
Good Riddance, its a bonus that the ZUCC is billions of dollars poorer.
1 GPU for Stable Diffusion for images, 1 GPU for text generative model and 1 GPU for speech synthesis. Soon, our waifus will be real.
The way I see it Meta is not only killing their Meta product line, they just killed their AI division. Out-of-touch and very bad management was what got them to create Meta in the first place and bet the house on it. Now they're consolidating they're admittedly successful AI teams into one? I'm betting that the bad management will consolidate too. Thus I'd expect to hear that they will begin to blow unfathomable amounts of money on their AI division, only to see weak results. With the AI market as hot as it is, it won't be enough to compete. That will kill them.
One can only hope!
Self-hosting is not the only way to train models on your data. OpenAI actually allows fine-tuning of the models. Tho not all of the models. They also claimed recently that they are not reading your training data. Quote from an news article:
> OpenAI is changing the terms of its API developer policy, aiming to address developer - and user - criticism.
> Starting today, OpenAI says that it won’t use any data submitted through its API for “service improvements,” including AI model training, unless a customer or organization opts in. In addition, the company is implementing a 30-day data retention policy for API users with options for stricter retention “depending on user needs,” and simplifying its terms and data ownership to make it clear that users own the input and output of the models.
> Greg Brockman, the president and chairman of OpenAI, asserts that some of these changes aren’t changes necessarily - it’s always been the case that OpenAI API users own input and output data, whether text, images or otherwise. But the emerging legal challenges around generative AI and customer feedback prompted a rewriting of the terms of service, he says.
They could've picked up a fraction of the money to buy vrchat instead and still stay strong on the VR front, and then monetize an already working product. But nobody hired me for consulting sooo.
The funny part is that they are very well placed to give ai as a service to corporations and small businesses for their chatbots facing the customers, ad they decided to go with the Meta fail.
wake up, new mentaloutlaw video dropped
Yes dear…
Imagine where we'd be right now if Facebook had spent as much on AI as they did on the Metaverse.
Zuck would have f*cked that up as well
Mark "Skynet" Zuckerberg takes over the world?
"Much worse off" if I had to make a guess ?
Imagine where we'd be if that money had been spent on homelessness and drug rehabilitation or public high schools or healthcare or raising teachers salaries or retraining police and giving them a raise or giving firefighters a raise or at the southern border or................
@@blueshattrick Nah they would LTB be behind if they invested in AI. Microsoft made a good investing decision with OpenAI and Google is only slightly ahead of Facebook. I predict Microsoft rules the future of technological utility
Looks like he's still set on the Metaverse -- the podcast Mark did with Lex showed a pretty significant increase in quality. Interesting to see if it'll amount to anything.
Company almost killed because one person couldnt change his mind early enough. It is a different thing to invest a certain amount into metaverse put to dedicate years and years of development to a metaverse that seemed so lifeless.
Unless some revolution happens in the semiconductor industry (that we haven't seen in the last 40 years), training these models at home is completely out of the question. I think with GPT-3 it would take several thousands of years (or tens of thousands) if you got the most powerful GPU available today. So you need many thousands of these cards, and even then training could take months, and a lot of energy. So it really doesn't worth focusing on that particular aspect, it's a non-starter.
I think it will be even more interesting because they will be able to feed it massive data sets they could literally just grab people's posts and conversations and feed it through
If you can't build a world, build a waifu to cope with.
--Zuck, probably
It’s not failed. My meta pro headset is mindblowing good. And it takes a platform at least 7-10 years to attract the big AAA+ game developers. The sideloader community is thriving already.
Facts, in a decade or 2 I’m confident meta/VR will be almost as big as they envisioned or bigger
You did well on self hosting argument
Typical world scenario of large companies wanting to dominate the narrative and profit
Self hosting is the way to go