Live AI assistants, INSANE 3D models, Gemini 2.0, SORA is out, AI video to audio, full AI comics

แชร์
ฝัง
  • เผยแพร่เมื่อ 15 ธ.ค. 2024

ความคิดเห็น • 251

  • @CodyCha
    @CodyCha 19 ชั่วโมงที่ผ่านมา +82

    Trellis 3D models are insanely good looking

    • @Allplussomeminus
      @Allplussomeminus 13 ชั่วโมงที่ผ่านมา +5

      Need to see the topology of them though.

    • @jamesjonnes
      @jamesjonnes 11 ชั่วโมงที่ผ่านมา +1

      It only allows one image as reference, so more specific models such as people don't come out as good.

    • @NimVim
      @NimVim 11 ชั่วโมงที่ผ่านมา +2

      @@Allplussomeminus Manual clean up will be necessary, just like with 3d scans, sadly.

    • @JesusPlsSaveMe
      @JesusPlsSaveMe 11 ชั่วโมงที่ผ่านมา

      ​@@Allplussomeminus
      Where are you going after you die?
      What happens next? Have you ever thought about that?
      Repent today and give your life to Jesus Christ to obtain eternal salvation. Tomorrow may be too late my brethen😢.
      Hebrews 9:27 says "And as it is appointed unto man once to die, but after that the judgement

    • @lehsu
      @lehsu 9 ชั่วโมงที่ผ่านมา

      3D modelers will need to look for a new career soon

  • @ShifterBo1
    @ShifterBo1 20 ชั่วโมงที่ผ่านมา +97

    Funny how Sora came after so much hyping and still wasnt good enough to satisfy most of us, truly pathetic.

    • @BoyFromNyYT
      @BoyFromNyYT 20 ชั่วโมงที่ผ่านมา +20

      Oh yeah with the price tag😂

    • @hdfsgervda
      @hdfsgervda 20 ชั่วโมงที่ผ่านมา +14

      Wow, the open-source Hunyuan beats Sora for many things and you can run it locally. If I was OpenAI (or even the makers of Flux, Black Forest Labs) I would be very worried

    • @CodyCha
      @CodyCha 20 ชั่วโมงที่ผ่านมา +16

      I tried Sora...and it's a total garbage

    • @bigdaddy5303
      @bigdaddy5303 19 ชั่วโมงที่ผ่านมา +3

      Yes I can't get noods

    • @xenn2996
      @xenn2996 18 ชั่วโมงที่ผ่านมา

      @@bigdaddy5303are you trying to blackmail someone?

  • @hdfsgervda
    @hdfsgervda 20 ชั่วโมงที่ผ่านมา +42

    Gemini 2.0 looks good. The Hunyuan open-source model is very exciting too

    • @orac229
      @orac229 14 ชั่วโมงที่ผ่านมา

      Gemini 2 is junk. I had a conversation with Gemini (argument) and it admitted that it is NOT capable of generating original content or ideas and is only able to draw on existing content from the web. No real difference from previous versions other than some more intutive interactions. NOT worthy of the AI label, just an advance web search engine

    • @DoubleRainbowXT
      @DoubleRainbowXT 13 ชั่วโมงที่ผ่านมา +3

      @@orac229😂 your really funny

    • @PhillipJohnsonphiljo
      @PhillipJohnsonphiljo 11 ชั่วโมงที่ผ่านมา +3

      @@orac229 user error: please replace user

    • @JesusPlsSaveMe
      @JesusPlsSaveMe 11 ชั่วโมงที่ผ่านมา

      ​@@orac229
      To everyone in this chat, *Jesus is calling you today!*
      Repent and turn away from your sins today to obtain eternal salvation🤗

  • @rhadiem
    @rhadiem 16 ชั่วโมงที่ผ่านมา +10

    I love that you feature open-source apps. Definitely earned my sub. Thanks.

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา +1

      thanks for the sub!

  • @davidmangold6167
    @davidmangold6167 15 ชั่วโมงที่ผ่านมา +3

    Just wanted to say thank you for these weekly videos. I love watching them on Sunday mornings!

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      You are very welcome!

  • @IceMetalPunk
    @IceMetalPunk 15 ชั่วโมงที่ผ่านมา +12

    It's important to note that Gemini 2.0 Flash's "Stream Realtime" in the AI Studio is not currently natively multimodal. It's still doing the old style "speech transcription -> text-to-text LLM -> text-to-speech" pipeline. You can confirm this by trying to get it to alter its voice or tone in any way, which it will be unable to do, or by asking it to listen to non-speech sounds and seeing that it won't even *send* the audio to the model that way. (And if you include non-speech sounds with some speech, it can't hear the non-speech bits at all.) I suspect it'll be updated in 2025 when their announcement video said public rollout of multimodal image chat would happen.

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา +1

      thanks for sharing!

    • @Bolidoo
      @Bolidoo 10 ชั่วโมงที่ผ่านมา

      It could be natively multimodal, but it's training data being from google's TTS and STT system's. So they get the latency and speed of 1 model, while having similar limitations to TTS and TTS.

    • @IceMetalPunk
      @IceMetalPunk 4 ชั่วโมงที่ผ่านมา

      @@Bolidoo No, but it literally does not even send your audio unless you have words in it.

    • @Bolidoo
      @Bolidoo 4 ชั่วโมงที่ผ่านมา

      @IceMetalPunk well a good quick system like that is needed no matter what, because the intent is for the ai to not stop talking due to random sounds, but the user talking.
      It can't be just TTS because you can ask it to tell a long story and it starts to talk very quickly, just like any other prompt. I don't think it's the most probable explanation they generated the text that fast. Rather when you hear the first word, the generation hasn't yet finished.

    • @IceMetalPunk
      @IceMetalPunk 3 ชั่วโมงที่ผ่านมา

      @Bolidoo TTS models can also stream their output and take streaming text input. Low latency doesn't imply native modality.
      It can speak faster for longer text, but it can't modulate the speech audio in any way. Even if you ask it to and it says it will, it doesn't. That is a huge indication (along with the rest) that what the model says and what audio is produced are happening in two separate places.

  • @XEONFX23
    @XEONFX23 15 ชั่วโมงที่ผ่านมา +3

    lol no 3D artist would call those " insane 3D models " since there impractical. But it is interesting ill give it a try for base meshes

    • @XEONFX23
      @XEONFX23 15 ชั่วโมงที่ผ่านมา +3

      possible background assets with a remesh

  • @joannot6706
    @joannot6706 16 ชั่วโมงที่ผ่านมา +3

    I'm never going to be tired of this guy saying:
    "This is what you get"

  • @chariots8x230
    @chariots8x230 19 ชั่วโมงที่ผ่านมา +5

    1:52 I don’t want the AI to “predict” what’s behind the object though. I want more control, so I can instruct the AI on exactly what the back of the 3D object should look like. At the very least, I want the ability to use AI to edit the back of the 3D model so I can make it into exactly what I’m looking for.
    Additionally, it would be cool if we could use AI to automatically rig the 3D characters and objects we generate.

    • @ac7969
      @ac7969 19 ชั่วโมงที่ผ่านมา

      That's the thing, if it can imagine it without your input - it can surely imagine it eithbjudt your text input, as you wish, master.

    • @globurim
      @globurim 18 ชั่วโมงที่ผ่านมา +3

      Baby steps, researchers are perfecting prediction first before you will be able to edit it. If it werent able to predict, even editing would be difficult.

    • @ge2719
      @ge2719 16 ชั่วโมงที่ผ่านมา

      Surely it'd be better for it to create the model using multiple photo angles of the reap object. Rather than a text input.

  • @iso-
    @iso- 6 ชั่วโมงที่ผ่านมา

    To be able to generate that audio that accurately is incredible

  • @teamredstudio7012
    @teamredstudio7012 10 ชั่วโมงที่ผ่านมา +2

    Truly amazing!! I can't wait for proper open source multimodal.

  • @TheRoyalSkies
    @TheRoyalSkies 20 ชั่วโมงที่ผ่านมา +14

    AMAZING!!!!!!!!

    • @leandrogoethals6599
      @leandrogoethals6599 13 ชั่วโมงที่ผ่านมา +1

      When Trellis video?

    • @SatvaMoliya
      @SatvaMoliya 13 ชั่วโมงที่ผ่านมา

      👀

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา +1

      😃

    • @juanjesusligero391
      @juanjesusligero391 9 ชั่วโมงที่ผ่านมา

      Nice to seee you here! I hope you also make a Trellis video in your channel soon! :D

  • @zengrath
    @zengrath 19 ชั่วโมงที่ผ่านมา +4

    that 3d modeler is cool, I kind of wish i had a use for it. Imagine making a cool creature using AI image generation and tweaking it until you get it just right, then using that image to then make a 3d model to then be used in your animations, movie, game, or whatever. wow. Indie developers will be making cool niche games that big developers are too scared to make while being able to make graphics far more efficiently, even if not AAA quality still free time saved and money saved to focus purely on gameplay is huge.

    • @TheMaxPirat
      @TheMaxPirat 14 ชั่วโมงที่ผ่านมา +2

      It's very far from usable for animations/movies/games or anything animated. Output is pretty bad and topology is unusable, you can't rig it and animate. You can use it as static object not to close to camera at best or as a reference for modeling.

    • @Langenbacher
      @Langenbacher 13 ชั่วโมงที่ผ่านมา +1

      3D modeller here. The mesh is its greatest weakness. You need to rework it entirely, and at that point, you could just as well build it by hand.
      Don't get me wrong, it is impressive, and I am sure at one point it will have ai retopo as well, making it usable in a professional setting.

    • @ianwatson5767
      @ianwatson5767 9 ชั่วโมงที่ผ่านมา

      ​@@LangenbacherAs a 3d artist I agree. What i don't understand though is why people care about generating a 3D model to animate in render in say, Blender, when you could just use AI image/video generator and just skip CGI entirely. VFX are already starting to invest in AI generated VFX over CGI

    • @HELL_Media
      @HELL_Media 5 ชั่วโมงที่ผ่านมา

      “It’d be cool to have a dream where people like you, I’m so lame I can’t even imagine being cool”

    • @zengrath
      @zengrath 5 ชั่วโมงที่ผ่านมา

      @@Langenbacher Yea i agree it's likely not in a super usable state yet, but the fact it's this well in essentially 1.0 is amazing. Doesn't blender and all have a way of optimizing models too? like reducing unnecessary faces and such with losing as much definition as possible? I am sure that this will be figured out in the future. I did boot it up and played with it a moment, i was shocked how it makes a model in only seconds, even if i bump steps all the way up, but i did notice it failed to transfer a lot of details. But again it's 1.0, surely a better version that takes it's time getting all the details right, then can be imported and cleaned up in a program like blender could eventually use it for something. I dunno. It's all still so early in this AI tech

  • @menosproblemos6993
    @menosproblemos6993 15 ชั่วโมงที่ผ่านมา +2

    The Sora Yoga reminds me of Terminator II :)

  • @TioElias-f2j
    @TioElias-f2j 12 ชั่วโมงที่ผ่านมา

    amazing to see all this new stuff coming frequently ,
    thanks for the contend

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา +1

      you're welcome!

  • @tozrimondher4250
    @tozrimondher4250 18 ชั่วโมงที่ผ่านมา +1

    Yes ❤ Finally I can almost skip the sculpting part in my workflow for creating 3D characters

  • @plumzstudios2500
    @plumzstudios2500 19 ชั่วโมงที่ผ่านมา +4

    How did you get it to generate images? I've tried numerous times and it doesnt do it, let alone edit images

    • @LacunzaOriginal
      @LacunzaOriginal 18 ชั่วโมงที่ผ่านมา

      me too doesnt work

    • @peterlustig7064
      @peterlustig7064 12 ชั่วโมงที่ผ่านมา

      I guess it's not rolled out yet

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      it might not be available for everyone yet

  • @connor.chan.jazzman
    @connor.chan.jazzman 7 ชั่วโมงที่ผ่านมา

    That video audio model is absolutely nuts.... Wtf. It seems like not only music producers such as myself are in danger, but sound designers too. It can reliably get realistic audio on clips... Its only a matter of time itll be able to add creative touches and more cinematic sound design to clips, real or animated..

  • @gameboardgames
    @gameboardgames 4 ชั่วโมงที่ผ่านมา

    Wild! Man. Amazing. .
    Trellis as a game making guy, is mind blowingly amazing. And can't believe its free right now as well. Total game changer.

  • @senju2024
    @senju2024 10 ชั่วโมงที่ผ่านมา +1

    When you show your camera to Gemini 2.0, you only need to show it like 1 second as it has memory. This means even if you are not currently showing the chinese label, you can still ask questions about it as it already saw it. 🙂

    • @yoagcur
      @yoagcur ชั่วโมงที่ผ่านมา

      Time limit is 10 minutes, I heard

  • @phdost3267
    @phdost3267 18 ชั่วโมงที่ผ่านมา +1

    Just tested Trellis. It's getting there but def still a way to go before this is really usable. The examples look nice but custom images are 3/10 at best from what I tried

    • @ge2719
      @ge2719 16 ชั่วโมงที่ผ่านมา

      Does it not allow you to input multiple images of an object to get a better result?

  • @constantinosschinas4503
    @constantinosschinas4503 8 ชั่วโมงที่ผ่านมา +1

    30:10 "Everything looks SUPER consistent": Railings traveling inside of the train window.

    • @aggressiveaegyo7679
      @aggressiveaegyo7679 5 ชั่วโมงที่ผ่านมา

      You can't even begin to imagine the rage and hatred you will feel towards the younger generation, because of their attitude towards the absence of diseases and death from old age as if it were the norm, while you lived in those difficult times.
      That's how now old people over 29 are irritated by the complaints of 25-year-old children, about the speed of the Internet, the need to work a lot, that their feelings are not taken into account or the low quality of AI.

  • @ZiaUlhaqZia
    @ZiaUlhaqZia 12 ชั่วโมงที่ผ่านมา

    Great and informative video, keep it up. Subscribed!

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      Thanks for the sub!

  • @marwanzayn6455
    @marwanzayn6455 10 ชั่วโมงที่ผ่านมา +1

    I tried using Gemini 2.0 and it doesn't edit pictures when i used it , i tried many times and everytime the respond was that it can't do it

  • @Entity303GB
    @Entity303GB 20 ชั่วโมงที่ผ่านมา +5

    I KNEW YOU WOULD DO NOW A VID (4 am for me)

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      😃😃😃

  • @zauliqe1783
    @zauliqe1783 15 ชั่วโมงที่ผ่านมา +2

    The polygon mesh and UV: 💀💀💀💀💀💀💀💀💀💀💀💀💀💀

  • @jzajzz
    @jzajzz 7 ชั่วโมงที่ผ่านมา

    27:30 how do you start using that tool .. there's nothing on the website but demos it

    • @Tyloron
      @Tyloron 7 ชั่วโมงที่ผ่านมา

      Exactly. How can we use the demo or download it locally if there are no working links to do so? Please help.

  • @VintageForYou
    @VintageForYou 14 ชั่วโมงที่ผ่านมา

    Another great video Swift Edit Looks amazing.👍🏼

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      thanks!

  • @zyuzikoff
    @zyuzikoff 2 ชั่วโมงที่ผ่านมา

    How did you do harmonic play?

  • @outlierz1796
    @outlierz1796 20 ชั่วโมงที่ผ่านมา

    I don't understand where you did the swift edit demo. Can you please explain?

  • @beastash3003
    @beastash3003 19 ชั่วโมงที่ผ่านมา +3

    Sora is something in high actions scenes i laughed so hard

  • @stl-xx5rq
    @stl-xx5rq 18 ชั่วโมงที่ผ่านมา

    I think it's worth noting that Sora isn't out in UK and EU yet, the same thing happened with Advanced Audio.

  • @MrRandomPlays_1987
    @MrRandomPlays_1987 15 ชั่วโมงที่ผ่านมา

    7:13 - WHOA, that's what I've been imagining we would get in the future (real time diffused environments affected by mouse clicks and movements ETC) and it's already a reality right now what the heck, incredible how fast the AI field progresses.

  • @Vojec9
    @Vojec9 10 ชั่วโมงที่ผ่านมา

    the part of video that you explain within Gemini 2.0 image manipulation is not clear, because I tried and got the answer: I cannot manipulate directly into image pixel... is there some tweak. please what exactly you set in car image...thank you...

  • @yikesAnTuAn
    @yikesAnTuAn 14 นาทีที่ผ่านมา

    Is there any photo composition ai apps? For example if I want to put a photographed portrait of a real model dressed up as an alien from a white backdrop to a different planet? Is swift Edit able to do that? Or any other ai sources you know of?

  • @mbwolk
    @mbwolk 5 ชั่วโมงที่ผ่านมา

    At 14:32 you demonstrate the use of Gemini 2.0 Flash for prompt-based image editing. I tried to do the same, but Gemini responded "I understand your request, and I appreciate you wanting to push the boundaries of what I can do. However, I must reiterate that I cannot directly manipulate the image file." How did you get it to work?

  • @Elefantoche
    @Elefantoche 10 ชั่วโมงที่ผ่านมา

    Trellis and SwiftEdit look very promising. Hope the later doesn't get the same xportrait 2 treatment...

  • @yikifooler
    @yikifooler 12 ชั่วโมงที่ผ่านมา +1

    3D Model : Prompt -- 'Increase the size of the Melons"

  • @GrandelZeon
    @GrandelZeon 19 ชั่วโมงที่ผ่านมา +3

    2026 gonna be sick i be with new ai gain speed we only gonna go faster and faster with ai to the point whin ai solve DNA to give us immortally i just hope i be alive to that point even if i have to down pay for 500 years to pay off my immortality .

    • @akosreke8963
      @akosreke8963 17 ชั่วโมงที่ผ่านมา

      Believe in Jesus and you will have your immortality free.

    • @codepython-k4c
      @codepython-k4c 9 ชั่วโมงที่ผ่านมา

      There is no morality, ever living being dies , that’s life pal

  • @mohegyux4072
    @mohegyux4072 10 ชั่วโมงที่ผ่านมา

    where can i try SwiftEdit? can't find a demo or code

  • @austinhaynesmusic
    @austinhaynesmusic 16 ชั่วโมงที่ผ่านมา

    Google AI Studio doesn't host its version correctly so I can't get it to modify an photo like your examples show. I'm not sure what I'm doing wrong. It is trying to use imgur for hosting and when I use the link it just says image not found.

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      image generation might not be released to everyone yet

  • @921Ether
    @921Ether 9 ชั่วโมงที่ผ่านมา +1

    we're living in the singularity

  • @dimitrishow_D
    @dimitrishow_D 14 ชั่วโมงที่ผ่านมา +2

    Its always huge news and it always disappoints

  • @matrixxx3662
    @matrixxx3662 14 ชั่วโมงที่ผ่านมา

    According to Chat llm (which Includes multiple chat bots) It will take about 5-10 years to reach perfection with these video models.

  • @vi6ddarkking
    @vi6ddarkking 19 ชั่วโมงที่ผ่านมา +4

    0:26 Last year I commented on a sillyTavern video that we'd be able to do Full consistente AI comics before the end of 2024.
    This is far from the only prediction I got to the month in the last few years.
    If you're interested in the equation I am using.
    Go back to your 2020 self just before this AI craze started and think about when you would have believed all this madness would happen. Then divide the time frame by 10.

    • @AGoodHeartedSoul
      @AGoodHeartedSoul 13 ชั่วโมงที่ผ่านมา

      dividing by 10 is wild

  • @devEveryWhere2542
    @devEveryWhere2542 14 ชั่วโมงที่ผ่านมา +1

    Brick of meta Verse on the Way

  • @YeNLol10
    @YeNLol10 17 ชั่วโมงที่ผ่านมา

    Wassup bro I was Waiting for you videos!

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      hi!

  • @AceLive_
    @AceLive_ 15 ชั่วโมงที่ผ่านมา

    OpenAI but is it OpenSource? No? Then Hunyuan will forever win this competition. Open source outlives subscription as it is limited to the people that pretty much trained the AI where as open source it has potential to be more improved by community. Of course it could also fall into subscription base within someone's wrong hands when someone decides to make money of it. Wishing that hunyuan will truly be open source, can't wait for image to video for comfy.

  • @NeoRelic-o8p
    @NeoRelic-o8p 7 ชั่วโมงที่ผ่านมา

    Last step!!! Here we go!!! 🥳🎉

  • @jblahut3591
    @jblahut3591 3 ชั่วโมงที่ผ่านมา

    "It does understand the physics of real world things..."
    He said as the AI video showed the window frame on the inside of the train scroll past like it was part of the landscape. Whoops.

  • @badashphilosophy9533
    @badashphilosophy9533 14 ชั่วโมงที่ผ่านมา

    mmaudio seems really cool and inovative, thats one tool we will need soon with our ai video making, ahead of the game kudos to those guys and gals. and with sora, soon we wont be dissapointed with shows like lord of the rings and marvel end game fully and completely skipping the giant wars that theyve lead up to

  • @colewilder6196
    @colewilder6196 18 ชั่วโมงที่ผ่านมา +1

    MMA audio is...spot on!
    (for an ai anyways)

  • @timbomb374
    @timbomb374 14 ชั่วโมงที่ผ่านมา

    The 3d model one is gonna be great for 3d printing mini figures

  • @mikkoturunen1893
    @mikkoturunen1893 13 ชั่วโมงที่ผ่านมา

    Gem 2.0 shows me only the html code of the image location instead of the actual image.

  • @esentakos
    @esentakos 13 ชั่วโมงที่ผ่านมา

    When asking Gemini 2 to create avideo. It instantly gives an expired sharing website url.

  • @tedfire
    @tedfire 14 ชั่วโมงที่ผ่านมา

    I just tried AI Studio with Gemini 2.0 and its refusing to produce images ? Its says its just an LLM and can't generate images? Am I missing something?

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      image generation might not be released to everyone yet

  • @edmitry
    @edmitry 19 ชั่วโมงที่ผ่านมา +1

    When do you have time to test all these neural networks?😅

  • @HER0S0L
    @HER0S0L 14 ชั่วโมงที่ผ่านมา

    I get the YT enthusiasm, but some of the news start to sound like commercials for example Diffsensei missed the task on several fields on 23:10 where everything is wrong but you ignore it.

  • @NatjoOfficial
    @NatjoOfficial 18 ชั่วโมงที่ผ่านมา +2

    Welp, that’s my dreams of being a professional 3D artist out the window I guess…

    • @RoyMaya
      @RoyMaya 7 ชั่วโมงที่ผ่านมา

      I learned to 3D model 15 years ago. I knew this day would come but I honestly didn't think it wold come this fast!

  • @chariots8x230
    @chariots8x230 19 ชั่วโมงที่ผ่านมา

    21:47 It would be useful if we could edit each of the panels though. I’d like to have the control to decide exactly what each of the panels will look like.

  • @Khann_2102
    @Khann_2102 6 ชั่วโมงที่ผ่านมา

    how to use the gemini ai image tho, it never produces an image for me

  • @magistrcooldayn233
    @magistrcooldayn233 16 ชั่วโมงที่ผ่านมา

    Picture motion look amazing, hope they will opensource it

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      I hope so too

  • @AKU666
    @AKU666 15 ชั่วโมงที่ผ่านมา

    I wonder if motion prompting can lead to advances in the field of video compression

  • @enadegheeghaghe6369
    @enadegheeghaghe6369 17 ชั่วโมงที่ผ่านมา

    I was greatly impressed by Gemini 2.0 flash

  • @lordmo3416
    @lordmo3416 15 ชั่วโมงที่ผ่านมา

    Image output is not available on gemini yet. such a bummer

  • @Kinos141
    @Kinos141 8 ชั่วโมงที่ผ่านมา

    Going to try Trellis on Hugging Face, but this is cool. Love FOSS.

  • @shubham-yd7xm
    @shubham-yd7xm 17 ชั่วโมงที่ผ่านมา

    Lol Trellis 3D is generating the best quality model

  • @badashphilosophy9533
    @badashphilosophy9533 14 ชั่วโมงที่ผ่านมา

    the 3d model generator is intricate so might be good for large things but i think hyperhuman or whatever its called is still the best and first of all this kind of amazingness, this new one looks like its a copy basically, but if its free then it wins coz hyperhuman is about 30 per month

  • @doodlydodo
    @doodlydodo 13 ชั่วโมงที่ผ่านมา

    i cant get sora to make simple coherent videos to save my life. I have to remake a video 30 times to get one usable video. even the simplest shit...

  • @eeyoretriple6
    @eeyoretriple6 14 ชั่วโมงที่ผ่านมา

    Can you please explain to me how you got Gemini 2.0 native image output model I've been sitting here for half an hour on this thing and it's telling me up and down that it can't do this at all? You also didn't say how to activate it in your video. Seems like you left it out for a reason..... is Google paying you?

    • @kirlianpictures
      @kirlianpictures 12 ชั่วโมงที่ผ่านมา +1

      Yep, it doesn't work

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา +1

      image generation might not be released to everyone yet

  • @Ali_Alhakeem
    @Ali_Alhakeem 12 ชั่วโมงที่ผ่านมา +2

    Trellis 3D is insanely good !!
    best Ai tool so far after gpt.

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      yes 😃

    • @ianwatson5767
      @ianwatson5767 9 ชั่วโมงที่ผ่านมา

      It's cool but you can't animate them unless you bring them into 3d program and rig/animate everything. 3D generators are cool but if the end goal is just to make digital media why not just use ai image/video generators?

  • @WisdomTooth1987
    @WisdomTooth1987 17 ชั่วโมงที่ผ่านมา

    finally the future is here. indie films and game development is going to be wild.

    • @TacticsTechniquesandProcedures
      @TacticsTechniquesandProcedures 14 ชั่วโมงที่ผ่านมา

      Topology is still bad. We have a way to go before models are usable in a game engine but it’s promising.

    • @WisdomTooth1987
      @WisdomTooth1987 14 ชั่วโมงที่ผ่านมา

      @ topogun. there is a lot of remeshing tools too. and you can also just use it as a better reference to redesign it yourself but use it as a proof of concept.

  • @dutchthenightmonkey3457
    @dutchthenightmonkey3457 17 ชั่วโมงที่ผ่านมา

    I opened your channel page.. and apperantly I was only recommended the “normal” videos, everything else you upload on ai is… something else

    • @Rudgged
      @Rudgged 9 ชั่วโมงที่ผ่านมา

      Yeah this is a gooner account trying to use AI for their weird fantasies.

  • @LocalizedDownpour
    @LocalizedDownpour 19 ชั่วโมงที่ผ่านมา +1

    Made in Abyss jumpscare @ 22:39

  • @zengrath
    @zengrath 18 ชั่วโมงที่ผ่านมา

    Swift edit and motion prompting are two i am most interested in trying, figures it's the ones not out yet. let us know when they are :)

  • @yikifooler
    @yikifooler 12 ชั่วโมงที่ผ่านมา

    Consistence character AI is a prank nowadays, the clothes are always different, It's a joke I really wanna see a Character from a netflix show in different position

  • @shubhuwithkotaropfp
    @shubhuwithkotaropfp 19 ชั่วโมงที่ผ่านมา +1

    What y'all think of Nvidia Meshtron (it was revealed yesterday)

  • @DanIel-fl1vc
    @DanIel-fl1vc 19 ชั่วโมงที่ผ่านมา +3

    Trellis is interesting but only for a substitute for 3d scanning. Probably most useful for generating 3d models like rocks. Maybe it would also work on faces, ask an AI to generate you 20 faces. Then this converts the faces into 3d, then you project a face with good topology around the trellis model. But you need 16GB of ram, unless the demo plans to stay up for awhile. Something useful from AI for once, saving you time. Put it in an installer so people don't have to mess about with the console to install or run it, that's a time sink.

  • @CHRIS-ELID
    @CHRIS-ELID 13 ชั่วโมงที่ผ่านมา

    Can I use Gemini 2.0 and add Photorealistic Avatar 3d? Please make a tutorial

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      i dont know of a quick & easy way yet

  • @MarxOrx
    @MarxOrx 9 ชั่วโมงที่ผ่านมา

    TLDR: SORA won the breakdance Olympics

  • @Memes4__U
    @Memes4__U 19 ชั่วโมงที่ผ่านมา +1

    2:02 i think this is an old printing machine

  • @CoupleFiesta-sy1vz
    @CoupleFiesta-sy1vz 13 ชั่วโมงที่ผ่านมา

    Cant generate image in gemini 2.0 flash

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      that feature might not be out yet

  • @A_HUMBLE_MONK
    @A_HUMBLE_MONK 7 ชั่วโมงที่ผ่านมา

    Rayguns cousin at the end has gone too far...🤣

  • @JDPxkumagawa
    @JDPxkumagawa 15 ชั่วโมงที่ผ่านมา

    Thank for the video there are a lot of AI that I want to try out this week from your video if it's ok with you can you give a tutorial on how to install diffsensie on pc/laptop please if you don't mind and thanks for your videos it really helps a lot with discovering new AI tools.

  • @MeowtualRealityGamecat
    @MeowtualRealityGamecat 12 ชั่วโมงที่ผ่านมา

    I’m sure this is like what PlayStation used to give players on the rewards program.

  • @LuciousKage
    @LuciousKage 18 ชั่วโมงที่ผ่านมา

    Soo menu 3D modelers loose jobs, did not expect that so soon

  • @greengreens9936
    @greengreens9936 8 ชั่วโมงที่ผ่านมา

    I cannot get google ai studio to make images

  • @fred-n6f
    @fred-n6f 9 ชั่วโมงที่ผ่านมา +1

    nice! thank you

  • @adammakagassama5645
    @adammakagassama5645 17 ชั่วโมงที่ผ่านมา

    WTF!!! This morning I Wake up as an expert 3D artist an old mangaka and a filmaker 😅 and Tony stark 😮

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      🚀🚀🚀

  • @zdspider6778
    @zdspider6778 13 ชั่วโมงที่ผ่านมา

    36:04 "And it would generate a real person."
    Weird Science.

  • @GrandelZeon
    @GrandelZeon 19 ชั่วโมงที่ผ่านมา +1

    why make models for a game whin players in game can generate there own models styles that what games gonna look like real deal player one shit

  • @Mrme-cn9je
    @Mrme-cn9je 16 ชั่วโมงที่ผ่านมา

    Great job! I was wondering if there is an AI that creates characters with walk or jump spritesheets that can be used to make 2D games!

  • @goliathonscave9834
    @goliathonscave9834 7 ชั่วโมงที่ผ่านมา

    My first 3D prompt: Cortana naked

  • @TheBeanGreen
    @TheBeanGreen 18 ชั่วโมงที่ผ่านมา

    Ai vision has gotta be great for blind people

  • @itssagar4212
    @itssagar4212 9 ชั่วโมงที่ผ่านมา

    Hi can you make full tutorial on git installed and miniconda to use them to install ai model and delete ai model that I install using them and how to update the ai model that I install using them. Sorry for the bad english

  • @WhisperingWoodsCamp
    @WhisperingWoodsCamp 33 นาทีที่ผ่านมา

    Could you please add a Turkish subtitle option? I am very interested in the videos, but I have difficulty watching them because my English is not good.

  • @SS801.
    @SS801. 16 ชั่วโมงที่ผ่านมา

    Sora finally 🔥🔥
    Lol but I can't access it because of high demand 😔

    • @theAIsearch
      @theAIsearch  11 ชั่วโมงที่ผ่านมา

      are you able to create an account?

    • @SS801.
      @SS801. 11 ชั่วโมงที่ผ่านมา

      @@theAIsearch i already have one

  • @psyraproductions
    @psyraproductions 17 ชั่วโมงที่ผ่านมา

    I consider Gemini to be nothing... unless they give out the native multimodal outputs.

    • @xenn2996
      @xenn2996 17 ชั่วโมงที่ผ่านมา

      What does that mean?

    • @psyraproductions
      @psyraproductions 17 ชั่วโมงที่ผ่านมา +1

      @xenn2996 this model is nothing yet until they actually add the native generation also audio abilities

  • @FaHiM58
    @FaHiM58 7 ชั่วโมงที่ผ่านมา

    I am amazed by MMAudio

  • @MarioTGP
    @MarioTGP 6 ชั่วโมงที่ผ่านมา +1

    it's a gundam!