This is the third text to video beta server I've joined in the last 2 weeks, but based on my first 2 prompts it might be the most interesting one. Good find!
Just want to take time out to thank you so much for your content! Watching you has helped me greatly with learning to create my own movies. I have just finished a short 5-minute silent movie using Gen-2, now I shall look into this video generator. You got a watcher for life 👍🤗
There is simply no other channel that does what you do, at least that I've found. Yet another stellar AI resource to add to the arsenal - thank you so much, Tim! The guy passing out was too much!
Hey, HoboLobo! Good to see you! It really is crazy...I kind of had to pinch myself when I went into the "archives" to see that the Gen-2 Beta was in April! That's...insane. It feels like years ago!
What I find most interesting is how the AI has gotten out of the uncanny valley. It didn't need to be perfect or even close to it. A lot of these videos have imperfections and mistakes. Aside from the blatantly obvious ones, most of the "mistakes" are just resulting in an interesting style rather than something jarring or uncanny. For example at 8:04, I'm pretty sure the AI is kind of confused on the background and the grunt... but I actually kind of like how it looks. A few adjustments and it might be really cool.
PS, you have a great following... nice to see people appreciating all of this and being inspired as well. Looking forward to what you folks all come up with in your AI journey's! Tim has his finger on the pulse and is def one of the more reliable channels on this subject. Like much of you, I watch his vids all the time. Keep creatin', and keep exploring, but most of all, just have fun with it!!!
Oh man, thank you so much-- and thanks to anyone else reading this comment! It might be slower growth by not going with the "flashy" topics ("How I made $3000 a day using ChatGPT!") or declaring Midjourney dead every 4 videos... But, I think in the end it has been worth it, since the folks who watch and comment are super creative, friendly, and...well, just downright awesome people.
This one has some really extra special sauce. I’m really looking forward when they update with image to video. That’s when I’ll swing back for a proper deep dive. I’m keeping my eye on Moonvalley for sure!
Nice work Tim - this one seems to have much more stability than any others right now. I feel they found success with the stability by limiting the prompts to specific styles. Helps bottle neck any and all prompts to a style they spent time training and not let it run free, getting crazy results.
This looks amazing. I'm gonna have to try out the character I was messing around with in a MJ vs Dall-E 3 vid I was working on. And now you've given me another video I want to work on. Stop it!
@@TheoreticallyMedia And holy crap. I just did my first vid on Moonvalley and it turned out awesome. That's going up on Twitter in a few minutes. Thanks again for bringing my awareness to this.
Haha, it is such a Mullholland Drive or Lost Highway moment! So, no Image or Video...yet. This is like a week old, and the devs are burning the midnight oil to get those features working. But for now, I'm actually kind of enjoying the old school text prompting! And considering how good it is, I can't wait to see what Image prompting is going to be!
wow Tim! as always highly highly appreciated you taking the time to share this without AI clickbait and in your usual calm but cool manner ;) Loving it. And the styles are amazing imho. Cant wait to give this a try.
Hurrah. Another option to play with. Competition is healthy. There is a filmmaker on TH-cam who uses all the generators to complete his flick. I like that system. Thanks again even if it wasn't a deep dive!
I’ll for sure circle back as it evolves! I think I made 3 Pika videos before I finally felt it was time to do that Ultimate Tutorial- haha, and even then they released a new feature like 2 days later! …sigh, can never keep up…
So, I think it’ll go the opposite direction in terms of cost. As the tech moves forward, it’ll actually get cheaper. Granted, there will likely be a super high end pro version that will cost a TON, but the more general models will still be (relatively) cheap. I think of it movie cameras. You can shoot on a 30k camera and get amazing results. But, at this point, you could also shoot a movie on your phone and get stellar results.
Yeah, it isn't quite a bad as Gen-2 tends to be, where like, entire buildings just slide into frame. To be honest, I actually kinda like those weird AI freakouts. One of these days Christopher Nolan is going to use it as an effect in one of his films. But...yeah, being Nolan, it'll be a REAL building they'll slide in...hah, lunatic.
@@TheoreticallyMedia that could be an interesting film conceit; how many times do you have to rewatch it to notice everything that’s going on? Nolan will name it “Latency”.
Excellent to hear! And get in early for sure! Although it might be "limited" in terms of all the other features we've gotten used to-- I think if you just see the base model w/ text prompts, it'll become apparently really quickly that they're onto something here. It's only going to get better!
@@ancientburner I for sure saw a few AI Video Ninja Masters in there today! But the crazy thing is: all text prompts! That throwback aspect is actually super cool! And the outputs are really fantastic. If that’s the clip in a few weeks when the “bigger” features drop? Well, it’s gonna be a real contender to the throne!
2:42 NEURON ACTIVATION No but for real, when this gets to a point similar to where image generation is at now where everything is so coherent and consistent to the point that you might not even question if it was generated or not then its gonna be realllllll scary. there's so much scaaarrryyyy uses this could be used for
You know me Tim, I try every new AI model that comes out, and for beta, this one has been pretty smooth. Simple straight forward prompts tend to work better, I definitely yielded much better results. Motion and coherency have been pretty good, especially for being in beta. Gotta say, it's been fun playing around with all these different vid models, but I still needs more vid to vid in my liiiiife. I love using Blender and live action video to ai video, I have great results when i have control over the comp. That's been the one thing that separates Runway from the rest of the pack. I literally was able to become whoever I wanted to be, from Lawrence Fishburne to Kratos. I have yet to see another platform that got me as close as Runway. Kaiber, works at times, but doesn't really fully transform the vid like gen 1 did for me. But Moonvalley is def one to look for... as always, thanks for another great video my friend. Keep crushin' it brotha!
Always love your insights and really appreciate the time you put into your projects! That's the one thing that sometimes kills me, I get to play with these models and go deep for a day or two, but then it's off to the next one. Don't get me wrong, I love it-- but I only have time for these micro projects. Anyhow-- complaints aside: agreed! I'm dying for a KILLER V2V model. Gen-1 is pretty great, but it REALLY needs a consistency update. Kaiber is also great, but needs some kind of image reference capability. Don't want to complain too much though-- where we are is pretty amazing. And a year ago: I would have never believed any of this was possible!
@@TheoreticallyMedia yeah, absolutely, plus the people at each of those companies have been super awesome to deal with, it surely have been a great journey, wonder if we'll ever see Vid-Journey in the near future
It looks good - thank you for the video. I still animate my Midjourney images manualy in After Effects, but with this pace, those new AI movie tools might be useful in a few months
I'm actually doing a video on Adobe Max right now (well...not RIGHT now-- right now I'm procrastinating by answering comments! ha!!)-- but my overall thought is that we'll be seeing "Gen Fill" in AE pretty soon. For sure Adobe has their own Video Generator (that's been teased for Premiere), so I'm sure there is a workflow to get to AE coming soon. Ok, back to work!!
Thanks, I never seem to get in early on these things until now! Already generated a number. My only beef with pretty much all AI image gen is that it is highly biased towards characters facing the camera. Really hard to get it walking away or across, especially without looking at the camera. Anyway, thanks for the "hot tip".
Haha, as the old TH-cam saying goes: Hit that Bell notification!! I do try to bring these things to you all as soon as I can, since I know betas can close up quick-- It's almost worse for me, since I then get yelled at in the comments for the next 4 months with people saying "Why didn't you tell us earlier!" haha... ...wait, that's going to happen again isn't it? Sigh. Well, at least you got in!
I do love it when you post an exciting new AI Video like this... Although the Moonvalley Discord Server is going the Bing Creator Dalle3 way.... SSS LLL OOO WW! But still what a jump in quality. Like you said in your vid looking forward to aspect ratios and neg prompts. I also like the mad,shonly ness of the vids and 5 secs does give you wiggle room to edit. Thanks again, D Edinburgh
Yeaaaah. That might have been a bit my fault. The servers are apparently slow on account of a rush of new users. They did announce that they’re scaling up, so hopefully it won’t be too long.
I have a comic book video series I’m working on and whenever I get something crazy from the AI prompt, I rewrite the story to fit that crazy out put lol. The guy who passed out would have been beamed up in the spaceship after being knocked out
very cool stuff! :) the rate of improvement we're seeing is just ridiculous. however. gotta admit. i'm still really a big fan of the cursed ai videos. stuff like the sleepy guy and the cell phone guy from Apr 2023? it always makes me LOL :D
Thank you, Richard! This one was a LOT of fun to put together! As I said in the video, I think there is a bit of a lost art of text prompting for AI Video...ha, and by "lost" art, I mean...y'know, over the last few months!
Moonvalley seems to have this thing, where if you prompt with fairly specific character details, it'll keep a pretty consistent character. I think I got over the details in the back half of the video? Sorry-- I make these things, watch them once, and then I'm moving on! Ha-- that and I HATE watching my own videos!
I really do appreciate that so much! Taking a pause to answer this comment as I fire up for the next one! Thank you again-- it really helps keep the motivation up!
So, I tried via a prompt, and it didn't take. I think you can probably do them the old fashioned way: Generate a video, cut it at the midpoint, and then run it backwards. But, not just by a prompt. I'll pop it in the suggestion box for the mods...To be honest, I tried too-- so you aren't the only one who is thinking about it!
Which platform can I add my own content. Picture from my phone to be transformed into anime video or 3D?. I have only went through a few videos but looking for this application
So that's a stylization filter or pass for the most part. If you are trying to run it locally, AnimateDiff and ControlNet are kind of the bleeding edge-- BUT, it also takes a lot work/know how and hardware to pull off. Kaiber is worth checking out. Here's a video I did on it: th-cam.com/video/wVuU2Ee7fks/w-d-xo.htmlsi=qdUnaVrfEVO3a7Rm Gen-1 from Runway does this...kinda. I really like Gen1, but it has that whole consistency problem, meaning you won't really get the same character twice. Another way you can go: Take your Still Image, bring it over to Leonardo.AI and use it as a reference image to create a Anime style character in the same pose as your photo. Then take that photo to Pika Labs and use the /animate command to animate the still. So, yeah-- there really isn't an all-in-one solution just quite yet...but we're getting there.
Thank you for all the intel. I will follow links you have given. Definitely going to support your patreon , your knowledge is valuable thanks for sharing. Benjamin
This one could be a serious contender to Gen2. I suppose they will start charging for it at some point, too. Can´t wait to be able to do this locally for free, using a webUI like Comfy or InokeAI... my monthly AI generator subs are starting to be serious burden on my finances; don´t think I´ll be able to keep up with this "subscription-based service" scheme 😕
They’re kind of turning into the streaming services for me. Amazon for a month, cancel. Netflix for 2 months, cancel. Disney…etc. I’ve got my mainstays, but you’re right: there are so many, it’s impossible to keep up with all of them and not end up bankrupt!
In my first visit to the beta, I can see multiple NSFW warnings for peoples prompts, even on subjects that look pretty benign. It looks like just another case of a company babysitting its users with their morality policing
To be fair, I did see a convo on this at one point. The devs are aware that the language model is a bit trigger happy right now. I actually got flagged on my “dark haired model ordering coffee” at one point, for the life of me, I still have no idea what the trigger word there was. They said they’ll be taming it down soon. I do get it. They’re early in and all public channels right now, so if you popped in and there were no filters… …well, I think you know how that’d go!
The trigger might not be the words, some part of the video generated could have triggered the filter. Many of those prompts work fine in a second or third re roll. By mysterious reasons 😏
@@TheoreticallyMedia I am sure it works both ways. It does not generate a video if the prompt has ehm... a wrong combination of words, but also it does not shows the video if something in the images triggers it. But it's imperfect, that's true.
Thank you so much! It’s a real testament to flying by the seat of your pants! I have to do these videos so quickly that I never have time to ponder! If I did, I’d probably ruin it!
Looks like next week they’ll be introducing some new features (img to video etc) so I’ll be swinging back for another look. I have some pretty cool ideas for it
Of any of the models. I think this one seems to be the most likely to catch on to a name. I think a name and a few character traits would get it right there. Might give that a shot tomorrow!
It is SUPER cool. The really practical stuff I'm seeing is some of the ambient design things-- like at 0:43. Run that at slow speed (or prompt for other designs) and you could use it as video art in a gallery or coffee shop or whatever...
heres what kindof worries me when it comes the the AI's that make photo realistic poeple....what will happen when these softwheres are perfected .....especially if real,specific poeple eventually becom renderable...between this and voice AI which is also coming along strongly...what if poeple start using this for malicious purposes? what if its used to mimic somones likeness and voice to frame them for something or slander them? DA already has a problem with poeple using image AI programs making "nsfw" of photo realistic children... im just worried this is going to get out of hand as well.
So, I think that you can’t spend too much time worrying. Every time a new technology appears, someone figures out a way to use it for ill purposes. AI is a double edged sword for sure, but in this case it also cuts for us as well. In all likelihood the cure for cancer will come out of AI. Generic diseases might be a thing of the past. It will likely be able to predict a heart attack long before it happens… The glass is always half empty and half full at the same time.
Yeeeeeah, that is one thing that tends to be a problem. I think when it gets image prompting, that'll solve that issue, but it seems to be a thing across the board. I have played around with doing a Faceswap on an AI generated Video, and that also seems to help a lot-- but...it's also an extra step.
I think it's cheaper? Plus, Midjourney has shown that its a really good way of developing a community. If you go to a website and use that, everyone is sort of in their own little silos and never interact. With Discord, you actually have the chance that a small community will form. I know many don't love it, but it looks like we're stuck with Discord and creative AI Startups for awhile now...
I am too, first time to a new platform when it’s just starting. Excited to see how it grows, and how I grow with it. Keep posting, love the content, I’ve learned a lot from this page
Thanks so much! I’m going to circle back to this one soon. If you saw the latest Video Face Swap one, I tried it out with our Blue Haired Solider and Jennifer Laurence. It worked…surprisingly well. I just cut it from the video since it was running a bit long, but I’m going to revisit it next week!
Don’t get me wrong, I love pika as well! To me, I feel like they all have their strengths in certain areas. Moonvalley is so new, I’m not sure what this is quite yet, but their text to video is really, really good. The more the better (at the price of free as well!)
I mean, that's common for any "free" service. As I leave this comment on TH-cam, which is owned by Google. I was just talking about this with a friend a few days ago-- It just sort of is how our digital lives are today. The best practice in my opinion is multiple "burner" accounts. Open a few Gmail accounts, a few Yahoo accounts, etc...I mean, there are still ways to track, but-- y'know, make it a little harder.
This is the third text to video beta server I've joined in the last 2 weeks, but based on my first 2 prompts it might be the most interesting one. Good find!
Just want to take time out to thank you so much for your content! Watching you has helped me greatly with learning to create my own movies. I have just finished a short 5-minute silent movie using Gen-2, now I shall look into this video generator. You got a watcher for life 👍🤗
That's so awesome to hear!! Seriously, makes my night! Thank you so much for watching and learning...and most importantly: DOING!
There is simply no other channel that does what you do, at least that I've found. Yet another stellar AI resource to add to the arsenal - thank you so much, Tim!
The guy passing out was too much!
Thanks for the find. I'm definitely trying it out! Crazy how fast this stuff is moving 😮
Hey, HoboLobo! Good to see you! It really is crazy...I kind of had to pinch myself when I went into the "archives" to see that the Gen-2 Beta was in April!
That's...insane. It feels like years ago!
What I find most interesting is how the AI has gotten out of the uncanny valley. It didn't need to be perfect or even close to it. A lot of these videos have imperfections and mistakes. Aside from the blatantly obvious ones, most of the "mistakes" are just resulting in an interesting style rather than something jarring or uncanny.
For example at 8:04, I'm pretty sure the AI is kind of confused on the background and the grunt... but I actually kind of like how it looks. A few adjustments and it might be really cool.
Thank you for introducing Moonvalley AI. I just enjoy it! I even started an AI short film series using it.
PS, you have a great following... nice to see people appreciating all of this and being inspired as well. Looking forward to what you folks all come up with in your AI journey's! Tim has his finger on the pulse and is def one of the more reliable channels on this subject. Like much of you, I watch his vids all the time. Keep creatin', and keep exploring, but most of all, just have fun with it!!!
Oh man, thank you so much-- and thanks to anyone else reading this comment! It might be slower growth by not going with the "flashy" topics ("How I made $3000 a day using ChatGPT!") or declaring Midjourney dead every 4 videos...
But, I think in the end it has been worth it, since the folks who watch and comment are super creative, friendly, and...well, just downright awesome people.
Great video. Looking forward to giving this a try and some play. Thank you for creating this video and for providing the link.
This one has some really extra special sauce. I’m really looking forward when they update with image to video. That’s when I’ll swing back for a proper deep dive. I’m keeping my eye on Moonvalley for sure!
Awesome! I love every new video AI, the more there are the faster we'll be able to really use it for our work.
Nice work Tim - this one seems to have much more stability than any others right now. I feel they found success with the stability by limiting the prompts to specific styles. Helps bottle neck any and all prompts to a style they spent time training and not let it run free, getting crazy results.
This looks amazing. I'm gonna have to try out the character I was messing around with in a MJ vs Dall-E 3 vid I was working on. And now you've given me another video I want to work on. Stop it!
Haha, I know, it's so hard to keep up!! We are all living the Distracted Boyfriend meme when it comes to AI Platforms!
@@TheoreticallyMedia Hah, that's exactly it!
@@TheoreticallyMedia And holy crap. I just did my first vid on Moonvalley and it turned out awesome. That's going up on Twitter in a few minutes. Thanks again for bringing my awareness to this.
@@shadyendeavor fantastic! Drop me an @!!
5:14 🤣
This was great for that alone! Excellent video. Look forward to a deep dive. Cannot believe it's been less than a year. 👍
Until I can use my own images to keep consistency, it’s useless
Yeeeesss exactly! I want to feed my own images/vids, template designs etc..
Well I wouldn't say that it's useless it's definitely cool
The David Lynch moment - cracked me up!
Does this support image to video? Or video to video?
Haha, it is such a Mullholland Drive or Lost Highway moment!
So, no Image or Video...yet. This is like a week old, and the devs are burning the midnight oil to get those features working. But for now, I'm actually kind of enjoying the old school text prompting!
And considering how good it is, I can't wait to see what Image prompting is going to be!
wow Tim! as always highly highly appreciated you taking the time to share this without AI clickbait and in your usual calm but cool manner ;) Loving it. And the styles are amazing imho. Cant wait to give this a try.
Hurrah. Another option to play with. Competition is healthy. There is a filmmaker on TH-cam who uses all the generators to complete his flick. I like that system. Thanks again even if it wasn't a deep dive!
I’ll for sure circle back as it evolves! I think I made 3 Pika videos before I finally felt it was time to do that Ultimate Tutorial- haha, and even then they released a new feature like 2 days later!
…sigh, can never keep up…
🔥This is so cool! 🤩thank you for this video 🙏 very useful, as always 👍
I imagine what this will be like 5 years from now and I think it will be very sophisticated (and expensive probably)
So, I think it’ll go the opposite direction in terms of cost. As the tech moves forward, it’ll actually get cheaper.
Granted, there will likely be a super high end pro version that will cost a TON, but the more general models will still be (relatively) cheap.
I think of it movie cameras. You can shoot on a 30k camera and get amazing results. But, at this point, you could also shoot a movie on your phone and get stellar results.
@@TheoreticallyMedia Who knows, but you're probably correct. Exciting stuff either way!
The soldier didn't just vanish; the entire background changed. It happens in all of the pan shots.
Yeah, it isn't quite a bad as Gen-2 tends to be, where like, entire buildings just slide into frame. To be honest, I actually kinda like those weird AI freakouts. One of these days Christopher Nolan is going to use it as an effect in one of his films.
But...yeah, being Nolan, it'll be a REAL building they'll slide in...hah, lunatic.
@@TheoreticallyMedia that could be an interesting film conceit; how many times do you have to rewatch it to notice everything that’s going on? Nolan will name it “Latency”.
Thanks for sharing will definitely look into this!
Excellent to hear! And get in early for sure! Although it might be "limited" in terms of all the other features we've gotten used to-- I think if you just see the base model w/ text prompts, it'll become apparently really quickly that they're onto something here.
It's only going to get better!
Wow, jumped in the server and already seeing some great vids others are prompting.
@@ancientburner I for sure saw a few AI Video Ninja Masters in there today!
But the crazy thing is: all text prompts! That throwback aspect is actually super cool! And the outputs are really fantastic. If that’s the clip in a few weeks when the “bigger” features drop? Well, it’s gonna be a real contender to the throne!
Looking really promising. Will be checking this out as soon as I clock off work tonight. Cheers for the update.
A moon valley update video would be appreciated. I love your breakdowns and they juat dropped a ton of new features ❤ appreciate you and your content!
Cool! I'll circle back! There are a few that I've been meaning to update, so I might package them all together.
we don't go near discord. Wake up Moonvalley.
Great video Tim. We should keep all out weird generations in AI for when we look back fondly in 10 years.
2:42
NEURON ACTIVATION
No but for real, when this gets to a point similar to where image generation is at now where everything is so coherent and consistent to the point that you might not even question if it was generated or not then its gonna be realllllll scary.
there's so much scaaarrryyyy uses this could be used for
You know me Tim, I try every new AI model that comes out, and for beta, this one has been pretty smooth. Simple straight forward prompts tend to work better, I definitely yielded much better results. Motion and coherency have been pretty good, especially for being in beta. Gotta say, it's been fun playing around with all these different vid models, but I still needs more vid to vid in my liiiiife. I love using Blender and live action video to ai video, I have great results when i have control over the comp. That's been the one thing that separates Runway from the rest of the pack. I literally was able to become whoever I wanted to be, from Lawrence Fishburne to Kratos. I have yet to see another platform that got me as close as Runway. Kaiber, works at times, but doesn't really fully transform the vid like gen 1 did for me. But Moonvalley is def one to look for... as always, thanks for another great video my friend. Keep crushin' it brotha!
Always love your insights and really appreciate the time you put into your projects! That's the one thing that sometimes kills me, I get to play with these models and go deep for a day or two, but then it's off to the next one. Don't get me wrong, I love it-- but I only have time for these micro projects.
Anyhow-- complaints aside: agreed! I'm dying for a KILLER V2V model. Gen-1 is pretty great, but it REALLY needs a consistency update. Kaiber is also great, but needs some kind of image reference capability.
Don't want to complain too much though-- where we are is pretty amazing. And a year ago: I would have never believed any of this was possible!
@@TheoreticallyMedia yeah, absolutely, plus the people at each of those companies have been super awesome to deal with, it surely have been a great journey, wonder if we'll ever see Vid-Journey in the near future
It looks good - thank you for the video. I still animate my Midjourney images manualy in After Effects, but with this pace, those new AI movie tools might be useful in a few months
I'm actually doing a video on Adobe Max right now (well...not RIGHT now-- right now I'm procrastinating by answering comments! ha!!)-- but my overall thought is that we'll be seeing "Gen Fill" in AE pretty soon.
For sure Adobe has their own Video Generator (that's been teased for Premiere), so I'm sure there is a workflow to get to AE coming soon.
Ok, back to work!!
Great!!!@@TheoreticallyMedia
Absolutely love these videos - thanks Tim!
Thanks so much!! And it looks like they’ll be dropping those missing features next week, so I’ll be circling back for another deep dive!
Haha 4:40 made me giggle, she looks super mega stoned AF
Ha, it’s medicinal!
Thanks, I never seem to get in early on these things until now! Already generated a number. My only beef with pretty much all AI image gen is that it is highly biased towards characters facing the camera. Really hard to get it walking away or across, especially without looking at the camera.
Anyway, thanks for the "hot tip".
Haha, as the old TH-cam saying goes: Hit that Bell notification!! I do try to bring these things to you all as soon as I can, since I know betas can close up quick-- It's almost worse for me, since I then get yelled at in the comments for the next 4 months with people saying "Why didn't you tell us earlier!" haha...
...wait, that's going to happen again isn't it?
Sigh.
Well, at least you got in!
I do love it when you post an exciting new AI Video like this... Although the Moonvalley Discord Server is going the Bing Creator Dalle3 way.... SSS LLL OOO WW! But still what a jump in quality. Like you said in your vid looking forward to aspect ratios and neg prompts. I also like the mad,shonly ness of the vids and 5 secs does give you wiggle room to edit. Thanks again, D Edinburgh
Yeaaaah. That might have been a bit my fault. The servers are apparently slow on account of a rush of new users. They did announce that they’re scaling up, so hopefully it won’t be too long.
Haven't tested out Pika Labs Masking yet. Bu that looks crazy cool. You'll have to update your PK PDF.@@TheoreticallyMedia
I have a comic book video series I’m working on and whenever I get something crazy from the AI prompt, I rewrite the story to fit that crazy out put lol. The guy who passed out would have been beamed up in the spaceship after being knocked out
very cool stuff! :) the rate of improvement we're seeing is just ridiculous. however. gotta admit. i'm still really a big fan of the cursed ai videos. stuff like the sleepy guy and the cell phone guy from Apr 2023? it always makes me LOL :D
Thank you for covering new tools for AI. This was a great video.
Thank you, Richard! This one was a LOT of fun to put together! As I said in the video, I think there is a bit of a lost art of text prompting for AI Video...ha, and by "lost" art, I mean...y'know, over the last few months!
what is the space animation from at the beginning of the video?
Just something I quickly rolled up in Moonvalley-- I think it was /create Looping Time, Design or something like that...
Def going to play w this Tim brah! That girl on the scooter scene was def money 🎉
The enormous cell phone is great, but can we talk about his cuff links? 😅
I also love that the woman at 1:36 is talking on two phones at once...which, to be fair-- I have seen people do!
A very stylish man indeed 😄
The guy who falls asleep at 5:15
Just needed a quick power nap before the big meeting!
how did you retained the caracter?
Moonvalley seems to have this thing, where if you prompt with fairly specific character details, it'll keep a pretty consistent character. I think I got over the details in the back half of the video? Sorry-- I make these things, watch them once, and then I'm moving on! Ha-- that and I HATE watching my own videos!
Indeed, a great channel!! Thanks so much for keeping us informed!
I really do appreciate that so much! Taking a pause to answer this comment as I fire up for the next one! Thank you again-- it really helps keep the motivation up!
Great news you always give helpful news which help us to use ai more better 😅❤😊
Thank you very much!
Is it hard to create loop videos?
So, I tried via a prompt, and it didn't take. I think you can probably do them the old fashioned way: Generate a video, cut it at the midpoint, and then run it backwards. But, not just by a prompt.
I'll pop it in the suggestion box for the mods...To be honest, I tried too-- so you aren't the only one who is thinking about it!
@@TheoreticallyMedia thank you
I just went on to the Discord server and they do have a small 'Guide' as a separate chat room
Oh that must be new. I’m keeping an eye out for image to video. Once that drops, I’ll do a full detailed run through
Trying this one after PIka became credit based. Thanks for posting!
Which platform can I add my own content.
Picture from my phone to be transformed into anime video or 3D?. I have only went through a few videos but looking for this application
So that's a stylization filter or pass for the most part. If you are trying to run it locally, AnimateDiff and ControlNet are kind of the bleeding edge-- BUT, it also takes a lot work/know how and hardware to pull off. Kaiber is worth checking out. Here's a video I did on it: th-cam.com/video/wVuU2Ee7fks/w-d-xo.htmlsi=qdUnaVrfEVO3a7Rm
Gen-1 from Runway does this...kinda. I really like Gen1, but it has that whole consistency problem, meaning you won't really get the same character twice.
Another way you can go: Take your Still Image, bring it over to Leonardo.AI and use it as a reference image to create a Anime style character in the same pose as your photo. Then take that photo to Pika Labs and use the /animate command to animate the still.
So, yeah-- there really isn't an all-in-one solution just quite yet...but we're getting there.
Thank you for all the intel.
I will follow links you have given. Definitely going to support your patreon , your knowledge is valuable thanks for sharing.
Benjamin
Thanks Tim, it looks good I will give it a go soon!
It’s a really solid one! Let me know what you think when you get in!
Nice one, Tim! Thanks for sharing!
Can I generate image with speech or host voice ?
Not yet. But I have a video today that will run through how to do that!
Can you add to your own server to avoid scrolling through other users work ?
This one could be a serious contender to Gen2. I suppose they will start charging for it at some point, too.
Can´t wait to be able to do this locally for free, using a webUI like Comfy or InokeAI... my monthly AI generator subs are starting to be serious burden on my finances; don´t think I´ll be able to keep up with this "subscription-based service" scheme 😕
They’re kind of turning into the streaming services for me. Amazon for a month, cancel. Netflix for 2 months, cancel. Disney…etc.
I’ve got my mainstays, but you’re right: there are so many, it’s impossible to keep up with all of them and not end up bankrupt!
In my first visit to the beta, I can see multiple NSFW warnings for peoples prompts, even on subjects that look pretty benign. It looks like just another case of a company babysitting its users with their morality policing
To be fair, I did see a convo on this at one point. The devs are aware that the language model is a bit trigger happy right now. I actually got flagged on my “dark haired model ordering coffee” at one point, for the life of me, I still have no idea what the trigger word there was.
They said they’ll be taming it down soon.
I do get it. They’re early in and all public channels right now, so if you popped in and there were no filters…
…well, I think you know how that’d go!
The trigger might not be the words, some part of the video generated could have triggered the filter. Many of those prompts work fine in a second or third re roll. By mysterious reasons 😏
@@myriampro4973 ah, that makes sense. So the filter is on the backside of the prompt, basically? Like after gen, not before?
@@TheoreticallyMedia I am sure it works both ways. It does not generate a video if the prompt has ehm... a wrong combination of words, but also it does not shows the video if something in the images triggers it. But it's imperfect, that's true.
The moonvalley link says it's expired. I guess I missed the party
You might want to try again? I’m seeing others getting in! Might have been a weird glitch
Thanks for this nice tutorial. Please tell me how you refer same character in multiple scene with an example.
Do these ai videos use faces from the internet or are these faces generated from scratch
i just wish i could get this quality with local text 2 video
Yeah, I mean the amount of local storage you'd need-- well, basically I'm sure your local power grid might be asking you some questions!
sir y are you drawing more power than the local school?
Can the movie be longer than 3 secs?
up to 5 seconds currently
@@TheoreticallyMedia Do you know if they may increase it? Whole lot of work otherwise...
Amazing!! love your ideas
Thank you so much! It’s a real testament to flying by the seat of your pants! I have to do these videos so quickly that I never have time to ponder! If I did, I’d probably ruin it!
The invite link is expired, can you please give us a new one? Thanks
yes, looks, good, I try myself, and if you mix people names, famose people names , as well you can get same faces
Looks like next week they’ll be introducing some new features (img to video etc) so I’ll be swinging back for another look.
I have some pretty cool ideas for it
Interested to know maybe giving a character a name would make a character repeatable, if someone tries send a reply I'm curious
Of any of the models. I think this one seems to be the most likely to catch on to a name. I think a name and a few character traits would get it right there.
Might give that a shot tomorrow!
That outtake! 😂
negative is now supported.
It's so good
It is SUPER cool. The really practical stuff I'm seeing is some of the ambient design things-- like at 0:43. Run that at slow speed (or prompt for other designs) and you could use it as video art in a gallery or coffee shop or whatever...
SWEET!!! 😃👍
Thanks dude
1000%!
heres what kindof worries me when it comes the the AI's that make photo realistic poeple....what will happen when these softwheres are perfected .....especially if real,specific poeple eventually becom renderable...between this and voice AI which is also coming along strongly...what if poeple start using this for malicious purposes? what if its used to mimic somones likeness and voice to frame them for something or slander them?
DA already has a problem with poeple using image AI programs making "nsfw" of photo realistic children...
im just worried this is going to get out of hand as well.
So, I think that you can’t spend too much time worrying. Every time a new technology appears, someone figures out a way to use it for ill purposes.
AI is a double edged sword for sure, but in this case it also cuts for us as well. In all likelihood the cure for cancer will come out of AI. Generic diseases might be a thing of the past. It will likely be able to predict a heart attack long before it happens…
The glass is always half empty and half full at the same time.
The most terrible thing in AI are the mad eyes. It make me very uncomfortable.
Yeeeeeah, that is one thing that tends to be a problem. I think when it gets image prompting, that'll solve that issue, but it seems to be a thing across the board.
I have played around with doing a Faceswap on an AI generated Video, and that also seems to help a lot-- but...it's also an extra step.
Why are they pushing folks to rely on discord? Annoying
I think it's cheaper? Plus, Midjourney has shown that its a really good way of developing a community. If you go to a website and use that, everyone is sort of in their own little silos and never interact. With Discord, you actually have the chance that a small community will form.
I know many don't love it, but it looks like we're stuck with Discord and creative AI Startups for awhile now...
Thank u!!
You’re are 100% welcome!
Addcsome subtitles
Doesn’t TH-cam do that automatically?
@@TheoreticallyMediayt does, he's new to yt i guess
This is dope
It’s pretty sick. And just getting started, which means it gets better everyday. Pretty excited about this platform!
I am too, first time to a new platform when it’s just starting. Excited to see how it grows, and how I grow with it. Keep posting, love the content, I’ve learned a lot from this page
Thank you...againヘ(^_^)ヘ
1000%!!
Good video 😊
Thanks so much! I’m going to circle back to this one soon. If you saw the latest Video Face Swap one, I tried it out with our Blue Haired Solider and Jennifer Laurence. It worked…surprisingly well.
I just cut it from the video since it was running a bit long, but I’m going to revisit it next week!
It just makes very short videos, what's the point? Can't use it for a music video, they're few seconds long.
it just released calm down.
Tuff
Crazy stuff. I think this week they’re going to drop those missing features, so I’ll be circling back for sure
Still prefer Pika labs
Don’t get me wrong, I love pika as well! To me, I feel like they all have their strengths in certain areas. Moonvalley is so new, I’m not sure what this is quite yet, but their text to video is really, really good.
The more the better (at the price of free as well!)
I what use this
Is it okay to make nsfw videos with it?
It is not. They have a filter
@@TheoreticallyMedia thanks, is there a.i program that's okay with it?
🤨
amazing ;)
It is pretty great!! And Freeeeee! (well, for now at least!)
The product is the user, this company is collecting your data to sell to 3rd party advertisers.
Read 👏 the👏Privacy 👏Agreement
I mean, that's common for any "free" service. As I leave this comment on TH-cam, which is owned by Google.
I was just talking about this with a friend a few days ago-- It just sort of is how our digital lives are today. The best practice in my opinion is multiple "burner" accounts. Open a few Gmail accounts, a few Yahoo accounts, etc...I mean, there are still ways to track, but-- y'know, make it a little harder.
AI mean weird. What fuck going on people mind. can AI make something normal.
Depends on your definition of normal! For sure things are still on the weird side, but it’ll get better as time goes on.
Ha Ha gas leak 🤣
999th like
I gotta start replying to the 666th with an Iron Maiden song.
This new fascist international state made possible by AI is great! ...because you can also pretend to have talent.
Not to split hairs, but AI tends to skew more decentralized, so I’d call it socialist, not fascist.