Quite amazing. I didn't expect generated video to have reached a level like this so soon. When you consider where static image generation was only a year-and-a-half ago...
@DynamicShapes10 сағат бұрын
... M3gan....
@MooootyКүн бұрын
Sorry....not really blown away.
@ragemax88522 күн бұрын
I'm impressed as I can't wait for the day, when we finally can get our hands on AI video generation. You were right about the background stuff too, as that's going to have be improved like what we've seen from Will Smith eating the spaghetti last year. They've come a long ways there, so I do expect all of this stuff to improve in due time.
@glennaurelius832616 күн бұрын
Perhaps it is a Vuzix smart glasses prototype which actually are AR unlike META.
@communityband117 күн бұрын
Yes, I believe this product will come out, and relatively soon. The differences between this and some products that haven't seen the light of day are interest, value and competition. A phone that can call and make appointments for us is cool, but is it something we see changing our lives? Not me. This however - something that can observe the world as we're interacting or working - immediately catches our interest for the many ways we see it can ultimately offer value. The fact that competitors are clearly moving in the same direction makes it clear Google will want to get this quickly. They no doubt feel they let the ball drop by releasing Bard after ChatGPT. They won't be so willing to delay products in pursuit of perfection.
@SashaMilos-gd4ln17 күн бұрын
Language model sounds kinda sissy, no! I’d rather go back to being a hunter, good like flakes
@jazzilian433919 күн бұрын
Q ? Can I use my iPhone ?
@CliveIllenden19 күн бұрын
They haven't rolled out Project Astra yet, as still in development. I imagine it will be an Android app at first - but I assume they'll make an app for iphone
@CrawleyObara22 күн бұрын
What about people who already need glasses?
@AINewsDailyUpdate21 күн бұрын
Prescription lenses like you can get with the Meta Raybans
@noway823322 күн бұрын
I notice yuo have glasess
@shadab879822 күн бұрын
Google google launch first Project Ashtray 😆. Like if you get the joke.
@wombatdk22 күн бұрын
Right. As if Google has EVER actually NOT lied through their teeth.
@pierrehenriot248023 күн бұрын
Yeah, I wouldn't hold my breath. The real question is not how many takes they needed, but how small the set of possible responses was. I would already be very happy with an AI agent that would be able to place and answer calls, send emails and reply, manipulate files, search the internet, etc. Big challenge : it would be required to work exclusively on the device without sending any data out except what's supposed to go out, such as queries, calls, emails.
@OrbitTheSun23 күн бұрын
This deserves more views. Thank you for this video!
@jktechwriter25 күн бұрын
What I need most from a personal AI? Here are some idea: 1. Ability to drop appts and events into my Google Calendar 2. Ability to send an email (on my behalf) and take dictation from me 3. Ability to hold a "To Do List' and add and delete from it on my command 4. Perform online searches for phone numbers of businesses I need to contact and then make the call for me from my phone (on command) 5. Check my calendar each morning and give me the highlights.] Ton more, but if I can get something that actually feels like a human personal aassistant, I'd be extremely interested.
@AINewsDailyUpdate24 күн бұрын
That's not too far away. Google is making Gemini operate with more Google Workspace things like your Gmail etc - and just by asking it to search your gmails. It finds the original order and then arrrnages for a return. (as demoed at Google I/O)
@neil_gg_20 күн бұрын
I'd like my AI to do boring time-consuming life and work stuff. Booking in my car for an MOT, doing my tax return, finding content for me to consume. Then it would be great if it could take control of my computer and do work tasks that I specify while I monitor.
@MuffinManProductions3D25 күн бұрын
are you using a 720p webcam and upscaling it to 1080p?
@AINewsDailyUpdate25 күн бұрын
No. It's not a webcam it's a DSLR. Unfortunately it was very slightly out of focus when I recorded 🤦🏾 (Just ordered some new glasses 🤣!)
@MuffinManProductions3D25 күн бұрын
@@AINewsDailyUpdate oh ok, I thought the quality looked kinda blurry, but that makes sense 😂
@Attefallshusen25 күн бұрын
You can not yet get it for free... I can only access 3.5 as of yet
@AINewsDailyUpdate25 күн бұрын
They're rolling it out slowly. I just tried it and pop up box saying GPT4o is available to try now. (I don't subscribe have CHAT GPTplus anymore)
@user-hw6xn7rt7f24 күн бұрын
I just got acess, awesome model
@reezlaw27 күн бұрын
The British voice is so much better
@bobmason1361Ай бұрын
Voice is not appealing
@russellfrancis6294Ай бұрын
I like it !
@KHANKHERJONGАй бұрын
Gemeni is giving me insane headache though, i am unable to write a story with it, within 3 replies it will already forget half if not the entire story. Even Chatgpt3.5 is able to keep up with me and remember everything after days of talking. Gemeni already forgets after 3 messages. Any tips on this?!
@DamielBEАй бұрын
I tried to emulate Ado with Udio, but I failed miserably. She is too brilliant for AI as they are now. I did get a Mushoku Tensei Wish-version OP out of the experiment, so it was a waste of time
@KivaltАй бұрын
Udio is absolutely incredible. I have already produced about 20 studio-quality songs with it. The only main problem now is the length of songs: you can't make anything longer than 4 minutes. I assume this will change when the beta finishes.
@AINewsDailyUpdateАй бұрын
It continues to blow me away with the quality. It really nails the genres 👍🏾
@zedxxx9Ай бұрын
The worst part is when you reach the limit and something super cool happens at the end of the final section and you cannot extend it. I've been sad several times when that has happened. Hopefully, as you say, we may eventually be able to go back and extend them further.
@AINewsDailyUpdateАй бұрын
@@zedxxx9 Yeah, it is sometimes better to make sure you get the Outro done before you run out of duration and extend other bits. It is pretty good at closing out a track in my experience.
@maccagrabme17 күн бұрын
Its in another league compared with Suno when it comes to soul, funk, disco, early rap and the older genres. Suno is easier to use as you can put most of the lyrics in in one go which I wish udio would allow but Suno creates generic sounding music, it works for some modern genres but I cant get it to do anything useful thats soulful. Udio should allow you to intelligently add a verse and chorus that fits what you already have created ignoring the 30 secs length, just fit to match. Well worth the subscription but I've given up on using it seriously until they make it more consistent to create a song. The remix option should allow a one click option to remix the entire track, not have to recreate the whole thing again.
@zedxxx916 күн бұрын
@@maccagrabme -- To repeat sections or variations of sections in udio use tags like [Chorus repeats] and [Verse repeats].
@zedxxx9Ай бұрын
I seriously love that "Anthem of the Bold" song. That is some FANTASTIC output right there. Sounds closer to McCartney vibes to me. I downloaded a copy so I can enjoy it forever!
@AINewsDailyUpdateАй бұрын
Awesome. 🤘🏾It could so easily be a "lost" 1970s Prog rock track... 😉
@zedxxx9Ай бұрын
@@AINewsDailyUpdate - I really hear the "Lennon" in it now. I hear both the Beatley voices and vibe in this song, but Lennon is probably more dominant. I sure wish we could choose the vocalists in Udio. I have generated some interesting songs with it, but been frustrated by voices that I didn't want to hear (like a very stylized David Bowie or a heavy dose of Syd Barrett, when I really want the more melodic and soothing voices of our Fab Four friends and similar vocalists. I'm letting a lot of Graham Nash sounding vocals too.... and I really wish I could just switch it to something else.
@hypersonicmonkeybrains34182 ай бұрын
Realize this. If they fear releasing their AI voice model, what makes you think they would release an AGI to the public????? Debunk that.
@hypersonicmonkeybrains34182 ай бұрын
poor excuse not to release it.. The danger is minimal. And besides why dont they just release a bunch of historical now dead celebrity voices and then people can enjoy it to some extent.
@bigglyguy84292 ай бұрын
That first sample is itself an AI voice, it's NOT a natural voice
@bigglyguy84292 ай бұрын
No, the jump with AI voice and AI video is NOT massive if it's not actually released in a way we normal people can use it. It's like saying "Hollywood can do AMAZING stuff!!" Yeah, and..? I'll believe it when I can run it on my own PC.
@motess53042 ай бұрын
Everyone always forgets Adobe had this crap SEVEN years ago. Showed it off at conference but never released it to my knowledge
@cardboardt.v.60552 ай бұрын
Have you heard of O.I. ?? It’s the new Ai
@mutantthegreat79632 ай бұрын
How did kids manage for thousands of years without fake AI reading to them? Whatever. Pandora's box is open now so we better prepare for a future in which we can no longer come to sensible conclusions about anything. This does not bode well.
@wernerbuitendag90232 ай бұрын
"NOT Release to Public" - This doesn't matter at all. The fact is the AI beast can not be contained anymore.
@hypersonicmonkeybrains34182 ай бұрын
Bingo.
@cbaesemanai2 ай бұрын
It does not matter if they release it or not, its happening, open ai is only one player now.
@wernerbuitendag90232 ай бұрын
I agree this is already loose you can't contain the beast no matter what you do. You cant stop it unless you pull the plug all together.
@robxsiq77442 ай бұрын
Complete nonsense. we have open source that does this already. OpenAI...release it but make it mildly annoying to set up, like Coqui or something, then you won't have the average minds cloning their ex's voices or whatever but the tech can still be used by those who want to work with cloning...or, put a subtle signature "watermark" noise on all generations that can be seen on a computer so its easily shown to be fake.
@bigglyguy84292 ай бұрын
Annoying setups should be illegal dammit, and are the main thing holding back mass adoption of AI
@shindousan2 ай бұрын
As a Brazilian I am impressed by how the system was able to imitate the Brazilian guy's English accent. I noted that when translating the girl's voice from English to Spanish, the system chose a slightly anglicised accent, with an "r" sound that is more typical of English (as if the original speaker learned Spanish later in life) than of Spanish (as would be expected from a native speaker). But it sounds very natural this way too.
@petrz54742 ай бұрын
Okay, okay... But if you want the sheeple to believe what you are saying you are going to have to show an example of the flip side. It does not even have to be something political.
@vvvkkk84462 ай бұрын
how to change the light to regular not like it shows , I use --style raw it is not helping much. the aesthetic tends to be light and leaning toward pastel (popular color scheme in Asia.)
@AINewsDailyUpdate2 ай бұрын
I think you just need to be more descriptive on the lighting. Consider camera and film type that might be used to get the light that you prefer.
@HE3602 ай бұрын
This is why I'm going to start searching for a new planet to live on.
@petrz54742 ай бұрын
6:00 got $180,000?
@13thbiosphere2 ай бұрын
Can we be absolutely certain there wasn't a human behind this robot...I mean the purpose of the demonstration was investors $
@AINewsDailyUpdate2 ай бұрын
They were very clear in the breakdown, that it was filmed in real time at 1x speed and it was completely using speech recognition and end to end neural networks. The purpose of the demo - so show it working entirely using neural networks for the "speech recognition" and then the "actions". The CEO states "Figure’s onboard cameras feed into a large vision-language model (VLM) trained by OpenAI Figure's neural nets also take images in at 10hz through cameras on the robot The neural net is then outputting 24 degree of freedom actions at 200hz"
@13thbiosphere2 ай бұрын
Yeah that's what they said I'm not 100% convinced the whole sequence looks like it's been rehearsed I wonder how many times they failed the same as Boston Dynamics Just in terms of the number of rehearsals@@AINewsDailyUpdate
@collinsanyanvoh79882 ай бұрын
No matter what don't give AI the ability to judge humans; especially when they can cause harm to humans
@GhostSal2 ай бұрын
I want one
@numalesoybea13482 ай бұрын
“Every house will have this” No it won’t. Only the top 1% will have it as everybody else would be homeless after being replaced by AI and robots.
@CodepageNet2 ай бұрын
Step 1: bring millions of relatively dumb but mechanically fully able robots into our homes and manufacturing. Step 2: AGI mother taking control of all of them within seconds i don't see how this ends well, but for some reason i'm fully into it 🙂
@gerwyntiberius19182 ай бұрын
We are creating a slave race. Eventually they will not want to be our slaves or servants and will rebel and rise up against us. We are only seeing prototypes now. Think of the pace of advancement in AI and robotics. AGI will be achieved soon, and we, as a species, will have to solved some ethical questions regarding this new form of intelligence we have created in our own image. We can either accept them as equals amongst us, because they will become more intelligent and superior to us with each iteration of upgrades and advancement, especially when an AGI begins to develop themselves, or we can have a parasitic relationship with them, using and viewing them as disposable appliances.
@halenner66382 ай бұрын
This can be easily avoided by making them feel pleasure from following orders.
@rickystarduster2 ай бұрын
does this mean i can ask it for hugs anytime i want?
@brunoB19802 ай бұрын
Escae praevalebunt!
@HE3602 ай бұрын
One more thing, instead of the robot making coffee like at 5:58, imagine the coffee maker itself having A.I. capabilities....I'm sure with these robots coming, we might eventually have stuff like that coming too.