The AI 'Genie' is Out + Humanoid Robotics Step Closer
Вставка
- Опубліковано 25 лют 2024
- First text-to-speech, text-to-video and text-to-action, and now text-to-interaction? Let’s take a look at the new Genie paper from Google DeepMind, and set it in the context of new developments regarding Sora and Gemini. We’ll hear what Demis Hassabis has to say about Altman’s $7 trillion dollar chip ambitions and touch on some recent notorious missteps. We’ll also learn more about Gemma, fully automated fabs, Elevenlabs integrated into Sora, and AI cheating unleashed.
AI Insiders: / aiexplained
Genie Paper: arxiv.org/pdf/2402.15391.pdf
Elevenlabs Sora: / 1759287778268025177
Sora Takes Time: www.bloomberg.com/news/newsle...
Samsung Humanless Fab: www.tomshardware.com/tech-ind...
Nvidia Gear: / 1761052023821369639
Sora Job ‘Loss’: www.hollywoodreporter.com/bus...
Huang Hardware: www.wired.com/story/nvidia-ha...
Ted Xiao News: / 1760591701410799682
Optimus Gets More Fluid: / 1761450961409573077
RT-2-X: deepmind.google/discover/blog...
Hassabis Interview: www.wired.com/story/deepmind-...
Gemma Paper: storage.googleapis.com/deepmi...
Full 5-star Video: / this_seems_relevant_wi...
AI Shark: www.techradar.com/computing/p...
AI Monitor: www.tomshardware.com/monitors...
Google Redteaming: / s. .
Figure AI Valuation: www.bloomberg.com/news/articl...
AI Insiders: / aiexplained Non-Hype, Free Newsletter: signaltonoise.beehiiv.com/ - Наука та технологія
Hey, thanks for covering our work! Great video as always.
Which one were you part of?
I'm the first author of Genie
Thank you guys for defining "scaling" in a non-buzzword way.
Oh wow, am honoured Jake. I don't think people realise how significant the achievements are of you and your team.
No, you're too kind. We love doing the work and are happy to see other people like it too! I've been watching your videos since the early ChatGPT ones, easily some of the best AI coverage on UA-cam. Great stuff!
Plz dont follow the "shocking" train like all other ai clickbait youtubers, i love your content.
It's not clickbait if it really is shocking. Sora is pretty shocking.
I will SHOCK you by not doing so
@@aiexplained-officiallmao
This comment is STUNNING
This revolutionary, groundbreaking AI assistant rewrites all headlines and regenerates all thumbnails to remove sensationalism from your feed! Everything can be boring now!
New AI development SHOCKS THE INDUSTRY. EVERYONE IS STUNNED, EVERYWHERE!
Every man, woman and child in paroxysms of shock, unable to stand, talk or leave UA-cam comments
Weekly
I'm so tired of that malarkey. Channels that have that in every other title are being lazy and irresponsible. Even past the elbow of the curve it would be bad practice and I don't think we're there yet.
@@aiexplained-official 😂
*Matt Wolfe wants to know your location*
The non-clickbait titles / not expressing things with needless over-hype is such a breath of fresh air! Thank you so much for optimizing for efficient science communication and treating your viewers with respect and not as nearly illiterate, focus deprived, ad revenue generating meat bags :)
We need to punish those who do with dislikes and misleading reports. Only way they'll learn not to insult their viewers.
“Don’t recommend channel” is the best feature on youtube, tbh.
Philip appeals to a more knowledgeable audience.
He's kinda like AI's Patrick Boyle. =)
but @parkhess don't you like it when CONSEQUENCES WILL NEVER BE THE SAME and THE FUTURE IS NOW
😂😂😂😂😂😂😂😂
People need to get a grip. Imagine being so triggered by titles 😂😂
I'm subscribed to a decent amount of "AI UA-camrs" at this point and I've found that almost all of them trend towards sensationalism and clickbaity content at this point. And if I've watched one of their videos on a topic I find that I can often skip most of the others without losing out on much.
This channel is the exception. You always manage to provide insightful and exciting content without hyperbole and it feels like your takes are actually your own and tend to add something to the conversation and not just regurgitating the same talking ponts as everyone else.
Keep doing your thing my guy, you're the best on the platform at this.
Thanks so much, noone
I’ve only found 2 channels worth subscribing to. This one for the “nuts and bolts”, and Matt Wolfe for “hey look at this cool stuff you can do!”
@@chrism1503 same here!
@@chrism1503Matt Wolf is kinda sensationalist too. Not too much but he does it.
Fireship is the other one id recommend. Funny and informative
Man, I feel kinda numbed by everything. When Dalle came out, I was in awe and started playing around with it. When GPT-4 came out, I was in awe and started playing around with it and really thinking about what it means to be intelligent. It excited me and made me think about deep questions. But, somehow I don't get that those kind of reactions anymore to the new stuff. Even though the new stuff is crazy. Sora is a great example. When it came out my reaction was more like "oh, that's crazy". But, in a numb way. I don't think I'm prepared for what's coming, at all.
The Singularity is what is coming, something is secure, this world gonna be very entertaining since 2026 with all the possibilities Ai will bring, movies, music, video games made by yourself, ultra and semi realistic, immersive reality, Neural chips implants, chatbots smarter than 99% of world population, the same chatbots incarnated and walking in animatronic bodies, metal 3D printers, organs 3D printers, drones expending food and medicines in conflict zones, nanobots with IA attacking only cancerous cells, and of course all that will have their military counterpart too.
I think it's because the first stuff was truly novel, the newer stuff isn't completely novel to us; it's a predictable improvement on the previous novelty no matter how good it is. Humans adjust to novelty pretty quickly.
This is exactly how I've been feeling lately. I mean, I'm very glad to see things progressing, but it's quickly accelerating out of this human's ability to comprehend or care. I think we're nearing the point where this stuff enters the zeitgeist even for people who aren't into AI and don't follow the news (y'know, the vast majority). Think of your average person's extraordinarily cartoonish understanding of how computers and most technologies work, now imagine those same people trying to make policy decisions and vote on laws about a nascent superintelligence...
We will all become somewhat desensitized over time, and I would worry about AI taking over completely as it overstimulates and lulls to “sleep”
Nobody is
Imagine a procedurally generated Dungeon and Dragons experience with a similar model in the future,
Exactly
Imagine . . . dragons
@@gamblerofratsRadioactive…radioactive
Yeah... that's the problem ironically.
I imagine a strategy video game with scale models of civilizations with 500 NPC villagers each with their own AI and humans have to manage their economy, technology and prepare them for cataclysms or invasions from other empires. Enveloping!
I'm changed shocked and stunned by this video.
Are you left on the floor, in a puddle of piss?
...I've never pissed so much in my life.@@aiexplained-official
We don't need to be shocked, just hear qualitative reports
*quality
get zapped
Words or worms, who really cares
@@assgoblin3981 🤭🤭🤭
Exactly 3 ai developments have actually shocked me. Alpha fold, my first interaction with Chat GPT and the Sora demo
This video just INFORMED VIEWERS ABOUT SOME OF THE INDUSTRY!!
And the almost the ENTIRE WORLD ignored THIS VIDEO
I think the Google stuff shows how profound an effect the views of the people handling the training prompts has on a model.
I get that the click bait tiles are a meme at this point.
But that was how you could easily dismiss a channel without watching a single video or even read the whole title.
Now that everyone is doing it ironically, you have to actually be careful not to permanently block someone who makes "real" videos.
So thanks for keeping the title clean :-)
Doing it "ironically" is just an attempt to get away with doing it unironically.
Haven't even watched it yet, but thank you for not being SHOCKED! Haha, you just said it!
Not that I would expect less from you, just need to let you know your high class doesn't pass unnoticed.
I wish this didn't become a meme because I want them to be able to say it's shocking because I am shocked by the advancement of ai
@@mattmaas5790 I know what you mean (and it IS amazing), but we can be shocked in our own free time. 😛
@@jvlbmeyou can shock me anytime you want lol it’s okay I can take it
it's kind of funny because I assumed I was at a point where I couldn't be truly SHOCKED anymore by AI announcements. But the text to interactive video caught me off guard. Man. What a time to be alive. I can't even imagine what things will look like by the end of the year, much less in 15-20 years.
15 - 20 years.
😂
Yeah. I saw it coming, but not today. Not this year, either...
Indeed. At this stage, is there even any point trying to imagine what things will be like in 20 years?
(OpenAI didn’t even exist ten years ago. 16 years ago there was no such thing as a smart phone.)
One of my favorite things about AI progress is just the fact that most people interested have had most of the ideas of what AI could do, just not that it would happen so damn fast. My late dad talked about switching perspectives in movies, replacing the actor for yourself and fun stuff like that, sad that he lived just short to see all this.
Never thought of it that way, but it's true. I'm still young, but if you told me about current AI when I was graduating high school, I would've been very skeptical. That wasn't even 10 years ago.
I already knew AI had capabilities to generate some OK sounding music, but I figured all the video and image stuff would've taken longer. I'm a musician, so I knew generating music wouldn't be too hard. Just pick a key, pick a chord progression, and have the AI generate MIDI within those parameters. A lot of DAWs already had features similar to this.
@@urphakeandgey6308Hi, yeah I am also a musician, studied production and as an audio engineer. I agree, the smart daw midi stuff. I'm wierdly scared and exited to get blown away by music AI
Sorry bro
@@Iam_inevitabIeThanks, was a cool dude. Helped me not be so freaked out about human ingenuity and how fast it can be
Sorry for you loss. Let’s hope the afterlife is much more fascinating than anything this world could ever produce or create.
Job unpredictability is a very relatable point. I'm 21, so at an age where I definitely think about that, and there are many jobs that generally seem like a good option right now, but I don't wanna bet on them, since for all I know they might need 10 times less people in just a year.
Do something with your hands like plumbing, be awhile (hopefully) before the robots come along.
@@ollydix Maybe I should, but it just feels like I would waste my potential in maths... Also I'm already half done with my physics bachelor.
4-5 years @@ollydix
@donit. I'd suggest staying in the domain and use your knowledge to expand the area of physics neural networks. I personally think it's an incredibly interesting field I would be focusing on if I knew more about physics.
I@@donit.If you are in physics, you good bro. Just go in research
BABE WAKE UP NEW AI EXPLAINED JUST DROPPED, GRAB THE POPCORN
EDIT: LOL - the world is neither shocked nor stunned. But coming from a Londoner, it would take a lot, wouldn't it?
True true
@@aiexplained-official LOL :D
1. There are already robots that walk really close to how humans walk.
2. Google's fiasco is clearly not them cutting corners or lacking something, rather the opposite the added too much, they put too much effort on "certain things". What we saw is not a bug but a feature (in their opinion).
Its been One Week since Sora.
All these days and still i get amazed to see those videos sora created the first time, truly a master breakthrough it was.
just so long as the industry that created it was shocked... 😂😂
Sora is gonna look like a joke when you see what they have in store for us
Honestly, it's crazy that we can even ask the question about whether or not we are entering a new era. It seems like era is now being measured in months lol.
When I read the thumbnail my heart honestly skipped a beat lol
Thanks for the disclaimer in the beginning.
We've hit a "Future Shock" scenario in which the changes keep coming at a pace faster than (many/most/nearly all) institutions plan.
We can't simply draw a trajectory and figure that we'll have AI with certain capability levels in a defined sequence or by a date; recently, each major release has been significantly surprising to the point of upending plans for exploiting/adapting/coping.
to mirror what a few others have said... I watch your videos because you provide thoughtful analysis and avoid hyperbole. I'm tired of being told what to think. I'm happy to hear what you think, just don't tell me that I should be shocked or stunned or whatever.
You do really well with this. please don't change your presentation style.
I like the way you mentioned it's not SHOCKING, huge respect for avoiding clickbaits!
What you are describing here might be an animated comic book where instead of having static frame the comic book as animated frames.
As such all the myriads of comic books made to date could then be animated and manipulated as well.
There may be a big potential market there!
Yeah I could see this being used in business
For the story about Tyler Perry I think we have to be quite cautious about the conclusions we draw from these kinds of events. Personally I'm not just going to take him at his word that Sora was the thing that made him pull out - it seems plausible to me that he could have had an ulterior motive. Corporations pull this kind of shenanigans quite often.
That's not to say we won't see impacts of generative AI in the economy, only that we should look to data not the words of self-interested agents to see those effects.
Ahh nice, phew, no shocking news here. Everywhere else I've been looking, I was shocked about how shocking it feels to be shocked by reading about all the shocked people also being shocked about being so shocked to read about other shocked people
Automating music, automating pictures/ photography, automating art/painting, automating code/ programming... Now automating game development with AI. Everything will be automated away with AI one day. I do in fact believe one day there won't be a need for jobs/ work and in turn money. AI will bring in the fundamental paradigm shift for the economy and society as a global whole, and it will all happen in the next 5 to 10 years. Definitely within all our own current lifetimes. Pretty cool.
Yup, totally sounds like fun, more garbage to consume
will ill still keep making songs on the piano because I love piano :) who knows maybe people will get bored of perfected ai generated music and come to my concerts :)
If they will have free time ...why not?@@Ah__ah__ah__ah.
This channel will go down in history books 🙌🏼
For real
Hopefully in a good way?
Best intro on all of UA-cam. You are just too cool for CAPITAL LETTERS.
Haha thank you
I must admit, this is pretty stunning. And I wouldn't have known it if you has said so in the title, I'm avoiding that trend as much as I can. So thank you for that!
Remembered me of the music video a-ha “Take On Me” .
Pencil drawings come to life.
Thank you so much for talking at a more moderate speed! It makes all the difference for the hearing impaired in terms of how much we can digest.
Beyond that, wonderful content as usual!
This is by far the best AI news channel I've found, thank you!!
(plus I'm already sick of all the clickbait titles haha)
Thanks so much Dave
8:00 You touched upon something interesting. I used to work for boring traditional software development and I noticed that with normal software, the bank backoffice workers can process 100x as many paperwork than if they did it by hand. It doesnt mean that we did layoffs, it just meant that no new hires will be necessary in the next 100 years or practically forever. Unless someone quits or retires of course.
All automation must steal jobs, otherwise nobody would buy that automation.
Thank you for sharing your time and work Phillip, in the description for the podcast you reference text to action, or text to interaction, but the group at Sanctuary AI have already moved on to voice to action, I pointed this out to another individual yesterday on X, really appreciate all the work you're doing and information you pass on to cavemen like myself, peace
Great point, so much more than text input
This video deserves a like because you delivered more than the title suggested.
Always love your videos! Thanks for the work you put into these!
thank you for explicitly mentioning that I won't be shocked. the reason that I was hesitant to try out your excellent channel in the first place was because of fatigue from all of the other clickbait AI UA-cam channels. Even two minute papers is on the hype train and he has a doctorate in light transport simulation.
Great video! I heard you mention a video on ai agents is in the works; hopefully MultiOn AI is part of it. *rubs hands together cleverly*
Please keep the level headed approach to AI news and explanations. It is something I appreciate highly from you! Good video as always!
Will do tuams!
I can’t imagine anything beyond Thought to IoT. It’s just a matter of scaling to include programmable matter at that point. Nice work, once again.
Thank you for acknowledging the meme without giving in, and for this and every excellent video :)
(Edit: typos)
:)
Thanks again! Your videos are always a breath of fresh air for my day.
Ah, I always love it when a new ai explained video comes out!
:)
No matter if they are shocking or not - thank you so much for keeping up with the latest AI developments! I wouldn't know how to get reliable information about these topics if it wasn't for your channel.
Thanks guest !!
I swear every time you post my heart sinks. I get more scared but have to watch 😂
Thanks for not being one of the "SHOCKED" Mor*ns
Great content as always! Love your videos and your insights into this rapidly evolving landscape.
Thank you smith!
Can you please add a link to the 1 minute clip in the description? I would love to send it to a few people
Awesome vid as always
Will try to fix links!
What excites me about Genie is the implications when combined with the Vision Pro and similar advancements. It brings us much closer to Star Trek's holodeck. I look forward to feeding it whole books and voluminous histories, and walking around inside.
robotics is where my interest in ai is these days, apart from its other scientific applications, of course (CLIMATE!!). looking forward to hearing a lot more about motion via a GAN where the adversary part is physics itself!
i think we would be remiss not to remember how much simulation waymo/google is doing to train its fleet of robot cars. i always keep that in mind when im trying to understand the timescale for practical robots, but also how waymo still seems to be inching along, steadily. thx for the vid :)
I'm definitely seeing a ton less help desk jobs... Great timing just getting certified.. I think everything's a little slower right now.
Thanks for not SHOCKING the ENTIRE industry like all other AI-themed videos as of late. Although I do wonder how much of a conversion boost you would get if you'd add in those magic words lol. The others seem to find it worthwhile enough to add it to the title of every single video they put forward.
Yeah it would probably triple views but some things in life are more important
8:20 I 100% agree with this take. This is why I trust your videos more than anyone else's. You understand how the world works on a deeper level.
Another amazing video. Thanks Philip! (havent finished watching it yet lol.)
The pronunciation slip in tortoise reveals you to be an AI... had me fooled for a long time AI explained well done
Haha, is that not how you pronounce it?
Good points in this video and I'm glad to hear that you also bring up the realistic advantages rather than "shaking" the world with the progress of AI development. Part why AI and robotics won't take over all jobs is due to the "Automation Paradox" meaning that if no one works there no money to buy products that AI and robots produce. Sure you will have to change and learn new skill sets.
Second is that "AI is not taking over, It's those who leverage AI who will". So make sure to keep up as AI develops and how you can use it to your advantage in the line of work you do because AI is coming whether you want it or not...
Before we get carried away I'm already shocked and stunned
I saw this coming like a year ago, I was imaginating that at some point image generation would be so fast and advanced to the point of being able to kind of simulate a game with all of its rules or simulate a real time environment of sorts (that could simulate anything in it, using sora that would be a possibilty maybe if mixed with Genie)
Thanks! Amazing content, as always. 🙏🏼
Thanks Stephen, for your kind comments over the months
Once again the best AI coverage out there! Can I ask you a question with some background? For at least a decade, I have reasoned that if it’s possible to create thinking machines that are smarter / faster / cheaper to hire than humans in every way, our economic value will collapse. You hint at this issue in the vid. It seems we have 3 choices: (1) accept automating everything and everyone lives on UBI, (2) fight to “keep our miserable jobs” that bots can do better and create neo-fascist / communist dystopias, or (3) do nothing and watch societies collapse. Currently (2) and (3) seem most likely, because (1) is a paradigm shift, but (1) seems by far the best for the average person. Would you care to do a video on this subject under the assumption that AI powered bots could automate ANYTHING (including even the new jobs!) within say 20-30 years?
Your title is so true! Ai is here and it’s still in its infancy. It’s all just a matter of time. The ‘Genie is out of the bottle’ The train has left the station’ yes it’s no going back and we can’t say stop cos too many players are after the same rewards or at least some of the rewards, they know that not being in the game isn’t an option unless they want to be a ‘Nokia’ or ‘Kodak’ ! So the race is on and there is a limitless budget. It’s Play All, play everything or face redundancy and humiliation. AGI is here, they are just waiting for the right time. First to announce AGI must have all the buttons working and be able to stand the inevitable scrutiny, but the reputation that will follow will be a new level of respect in an already respected technology industry.
"The inability to plan your career". The Technological Acceleration Paradox is going to become a known thing very soon. It is the principle that the faster you accelerate the less likely you will arrive at the future you envision. The things you dream about become obsolete before you reach your destination. I wrote something recently on this topic.
As a retired cartoonist I can highly recommend no job with free money.
enjoy your retirement! may we all join you soon :)
Very interesting - thank you for update:)
Thank you for not using the stupid SHOCKS the entire industry click bait ❤
Bloody ell I forgot to even this of it. We'll have ai games in the next few years. I imagine you could write pretty much any flash game with just a prompt, and easily refine it to whatever you please. What a time we live in
and im only 30 seconds in!
This is amazing for the Video Game and Simulation industries
For the past 10 years if someone wanted to get into software I would have told them ignore algorithms and go straight to frameworks and libraries and cloud platforms as 95% of job postings want those. Going forward however I feel like it will become increasingly important to know basics VERY WELL, as you want to be able to smoothly transition from one tech stack to another more and more frequently as the tech changes more and more rapidly.
For 10 months now I am amazed by how good ole' programmers just won't admit programming is toast. They just won't 🙂
The compute optimal training strategy discussed in the chinchilla paper is only concerning itself with training compute, right? But if you plan on widely deploying a model, inference compute is not insignificant. A smaller model trained on more tokens (i.e. inefficient according to chinchilla) would be cheaper to run than a bigger model with the same performance. If your service is popular enough, there's a break even point where the smaller, but longer trained model is the better choice.
yes it's training compute
Exactly
It is hard to believe how jaded I am.
I struggle to remember how impossible it would have been to believe if someone had showed me this a couple years ago.
Yeah we're going to build all the sci fi stuff
This was predicted last year.I was waiting for this to hit gaming as well as VR
Great video! Some of the links are not working for me though, they seem to be truncated. In particular, I would like to see the Hassabis interview.
Will be updating them! Not sure what happened
Thank you for your quality videos. Love your content!
Another element of the AI in jobs discussion that is often missed is that initially it will be interns and lower echelon employee positions that get replaced first. I only ever hear people talk about how their high level position is safe and that no AI could ever hope to match them, but they completely miss the point about how their own paralegals or assistants end up replaced. Who exactly did you plan on carrying the torch after you retire?
Nice to see UBC on that Genie paper.
I instantly gave you a like when you said the part about not being shocked etc etc!!!
Bro if genie blew your mind. You must check out some of the research on triplane representations(true 3d representations). Not 3d models, but iterative 3d environments. More compelling to me for obvious reasons, especially for robotics.
Fyi you're top three AI UA-cam channels. Period. Great content as always.
What do you mean top 3?! :)
@@aiexplained-official 😂. True true.
“ *Societal impact* Genie could enable a large amount of people to generate their own game-like experiences. This could be positive for those who wish to express their creativity in a new way, for example children who could design and step into their own imagined worlds. We also recognize t with significant advances, it will be critical to explore the possibilities of using this technology to amplify existing human game generation and creativity…”
So we have this technology that, even just a few years ago, would have seemed nothing short of miraculous and what these people at Google come up with for “societal impacts” has mostly to do with _games?_ (I get that in our current economic system, such as it is, pointing to something that might be most profitable most quickly is a kind of imperative.) I dunno-I find this relentless techno emphasis on games, CGI animations for entertainment and virtual reality-in other words, escaping from the “real world”-disconcerting. It’s like the 21ˢᵗ-century digital equivalent of soma from Aldous Huxley’s _Brave New World._
You have a point.
The mainstream opinion before 2022 was that AI at first will target low skill, repetitive jobs. However, when text to image models came out, I got a bad feeling that this whole endeavour took a very unwise turn...
I don't think it is a lack of testing, the issue is that Gemini is behaving exactly as intended.
Nah... It's somewhere in between.
The most likely thing what happened is that they finetuned it to "diversify" its results but overdid it and didn't thought to test it for a not "diversified" prompt.
Great video. I just want to say it's Sad that AI is now going to be used for cheating in video games, I am hoping AI will eventually be used to Wipe out cheating from video games. AI Spectators that can watch competitive multiplayer games and immediately detect cheating or hacks and stop them right away. That's one aspect of the AI future I want for the gaming industry
Me too gensteps
Thanks for keeping it real Philip.
You missed the part where they are trying to generalize automation of interacting with internal enterprise business apps such that it could take action across several web & application interfaces visually, this is the real threat to low-skill tech adjacent jobs.
Also somehow they talk about this helping with real world interactions & processes involving multiple interactions that it can piece together to achieve some goal.
I can imagine Genie will have a huge and helpful impact for architects and urban planners to show clients how people might interact with a static design
great video Philip, as always
Great idea
My belief is that sora and genie are play things. They demonstrate capabilities when you scale compute 100x(according to emad). And scaling works contrary to belief of most. Image to video or game is useful up to a degree. But they gain world models as emergent property. And using that world model they can for example make a blender scene. Write code for a unity game. And have editable environment. Thats what happens when you scale models. Apply that to gpt 4 and we are so back.
Thank you for remaining more grounded. It'll prove a smart decision I think.
Thanks temporal
I’m shocked and stunned you’re not using clickbait titles ;) ;)
I know, right
Genie was my immediate thought for one comes next the moment I saw Sora. Software developers were already walking a tightrope with how good LLMs are at coding, but this could essentially make coding itself unnecessary. Why develop a backend and a frontend interface using code when AI can simply create the whole thing on the fly? I'm sure fundamental new hardware paradigms are already in the works to take the humans entirely out of the loop.
I'm pretty sure you seen it but have have you heard of the large world model, ring attention, long ROPE and the new mistral large
Is the Discord membership tied to contintued Patreon subscription at the AI Insiders tier or would a single monthly subscription be enough to be allowed to stick around on Discord?
At the moment former members have the general discussion and announcements
still shocked me best ai channel ❤
I'm tired of being shocked by thumbnails just to get subpar news. Good video title 😊
10:32. I don't know, at this point, I don't think I'm comfortable making any hard bets as to what scale cannot accomplish. But, if nothing else, I'm sure other architectures can probably achieve these goals at lower scale levels.
What's exciting about this is it demonstrates our ability to train AI is increasing into more open-ended scenarios. If that makes sense.
I hope this year is going to be "text to task" and all about robotics integration
Another great episode 👍
I love the infinite wolf pup hallucination. ;-)
They were multiplying like bacteria :)
Great video as always! and thx for not jumping on the clickbait bandwagon!
When I hear that AI movie trailer narrator voice all I can think of is a Turbodong video... 🤣