Adobe's NEW AI Video Features are AMAZING! (Let’s Recreate Them)
HTML-код
- Опубликовано: 22 май 2024
- Join Our Newsletter for the Latest AI Filmmaking News: curiousrefuge.com/
Our AI Filmmaking Course Starts Soon: curiousrefuge.com/ai-filmmaking
Check Out AI Advertising: curiousrefuge.com/ai-advertising
Time Code:
00:00 Intro
00:44 Vasa 1
01:55 Adobe's upcoming AI Video features in Premiere Pro.
04:11 How to Use Adobe's future AI video features Today.
13:19 Adobe's new AI video features and the future of AI in creative processes.
16:57 Recap of AI Creativity Tournament and introduction to Meta's new AI tools.
18:14 Features of Meta's AI tool and comparison with ChatGPT.
22:08 Video2Game
22:57 AI Films of the Week
23:00 AI-mee is Heartless
23:23 The Fliffatroo
23:45 The Shipping Forecast
24:07 Outro
Helpful Links:
Film Ideas (Meta.AI vs ChatGPT): docs.google.com/document/d/1g...
VASA 1: www.microsoft.com/en-us/resea...
Adobe's AI Updates: x.com/Adobe/status/1779916565...
Pika Labs: pika.art/home
RunwayML: runwayml.com/
Meta AI: www.meta.ai/
Video2Game: video2game.github.io/
AI Films of the Week:
AI-mee is Heartless: curiousrefuge.com/ai-film-gal...
The Fliffatroo: curiousrefuge.com/ai-film-gal...
The Shipping Forecast: curiousrefuge.com/ai-film-gal... 
Развлечения
Such a great time at the event! Congrats again to everyone who participated!! You made history!
We couldn't have done it without you! We're pumped for next year :)
@@curiousrefuge @TheoreticallyMedia you guys are turning into quite the AI double act ;-)
Keep the collabs coming, guys!
Thank you so much for the update
Any time!
Love the Films of the Week! Thank you!
Our pleasure, thanks for watching!
YAAAAS baby!! Thank you dude!!
Any time!
Thanks for the tips.
Happy to help!
Thank you so much for the clever depiction of Adobe tools and, guidance for the best use of Pika and Runway. Haiper seems Great and also it is free, but just doesn't seem quite ready yet.
Your videos are always a wealth of information. So many tools are similar in their capabilities. I would love to have your recommended list of the best tools to build a complete A.I. film with dialogue, effects, and music; along with the total cost in subscriptions.
Glad you like them! We'll be making a tool page very shortly to help out with this!
One doesn’t exist yet
I am working on the idea of generating graphics through AI and creating a book trailer. So far i have created two videos and i think it turned out pretty good. 😛 These new features would be really helpful for me, I think. Thank you so much for this ❤
Sounds great! We can't wait to see!
Awesome 👍💪
Thanks for watching!
Thank you!!! 🤖
Our pleasure!
Hello! How does one go about submitting their work for the Ai Films of the week section. Thanks for everything you folks do :)
You can visit our gallery and submit. Alternatively, if we catch it on social media we may reach out to feature it too :)
film of the week is the best part keep going
So glad you enjoy that!
Darn, missed you guys! We went but didn't see you on the showroom floor.
Aw nooo! It's okay, we'll be back again next year :)
Thanks for what you are doing with this. I am looking forward to taking your AI Filmmaking course soon. L.HUNT
Is there a link to the e-sports event? I'd love to see a video of the "games"!
Congrats on organizing it and creating a new genre!
Once we have one we will share it :)
Heres an idea- do a video with tools we can get today!
We cover that in our free discord and also in course!
I am so happy they put a friendly face on Skynet...
cue the terminator 2 music!
Literally had to scroll for a while to get to a real person. Worth the thumb exercise.
Glad you enjoyed the video!
Curious - is there currently a good AI option for converting live action video into animation?
You can do style switching with tools like Runway or Leonardo :)
What's the closest thing to VASA-1 that is already released with consumer access? I.e. lip syncing with head, eye, and expression motion?
Likely something like Viggle but you would need to do your own acting.
@@curiousrefuge Viggle is very interesting. Am I correct that it's more about mapping a photo onto the full body and fitting to the physical motions? As opposed to lip-syncing?
I Other than VASA-1, I recall a Chinese research team that also put something out that was amazing at not only lip synching, but expression interpretation and head motion.
Have you found any good tools for AI powered Colour Grading?
Great question - not really. Typically most people would suggest the usual things like Davinci Resolve.
Curious -- Midjourney is frquently referenced as a gold standard in text to image programs, yet to my understanding it has not joined hands with any other ai groups -- in the same way that Dall E3 has partnered with open ai chat GPT. Why not? Or is there an MJ partnership on the horizon that is not yet visible to us? Anyone know? Anybody...???
Midjourney is in conversation with Grok actually. So we’ll see what comes of that.
NOT in Brazil Yet... 😞
Oh no! Good point, hopefully soon!
Microsoft did not "release" a tool. They specifically stated that Vasa 1 is not a product and there is no release time.
correct! It was a demo that is not released and says such on their paper.
22:25 Tony Hawk realground
Nice catch!
good video, please use my custom chat GPT, its create movies, books, and also generate the images. so you can later do a video
We'd be happy to test it out first!
vasa 1?
Vasa 1 :)
thanks advanced a.i. host.
Beep boop beep - human has found out our secret!
Just kidding - we are real!
Creatives have no idea how many people this will put out of a job
Creatives that utilize these tools will win.
@@curiousrefuge at what point does a tool replace its user?
It will kill the entire video, film and tv industry.
I made the cut
Woohooo, nice work!
I hate that every time you show something on her its nevver available for use and then you use borderline click-bait thumbnails to lure me in...Seriously "How To Use Adobe AI Video Features" The answer is wait til its realeased later this year! That headine is a bit misleading... Its odd thst you dont have more subs when your focused on AI when similar channels Ive seen got 100k-250k in less then a year, Justr a suggestion but the difference I see is they Dont use any misleading thumbnails and the RUclips Algorithm sets them up better, this is just a theory as noone seems to really know exactly how it chooses, maybe its a coincidence but I woulsd like to think that it leans away from anyrthing misleading. maybe just try it for a month and check the numbers.
We appreciate your feedback. Did you see in our title "(Let's Recreate Them)"? The AI community is very large, we focus on a subsection of it.
Fantastic video and highlights
Just sorry Microsoft is behind this next level photo to avatar, one of my least fav companies ever and don’t trust anything they do.
thanks for watching and we appreciate the comment!
Exports out of AI tools like pika Or runway is terrible so far. They need to give better render, there should be even raw export option. Resolution isn’t everything.
That would be an interesting feature!
microsoft isnt releasing anything.
go back and re-read the information.
"Given such context, we have no plans to release an online demo, API, product, additional implementation details, or any related offerings until we are certain that the technology will be used responsibly and in accordance with proper regulations."
they probably won't release it all,
its an internal tool.
later
||| | || | |
ne gative 1
Good catch - although this technology may be released very soon via another competitor.
@@curiousrefuge actually the other competitor is in the same boat. no plans to release it publicly, or api either.
If you're talking about the EMO project from alibaba.
This project is intended solely for academic research and effect demonstration.
This page was built using the Template which was adopted from the Nerfies project page.
They 're basically trolling companies and people to fund future developments.
They have no interest in releasing these to the public.
It's all a shell game, like crypto and NFTs.
later
|| || || | | | |
ne gative 1
Or fake videos that fule us or steal money 😢
Or inspire creatives to do more!
@@curiousrefuge I believe that handcrafted things will experience a renaissance as generated content becomes more prevalent in the media in the coming years. Similarly, real actors in movies will be increasingly valued. There's a significant difference when you know that a real person-a star who ages alongside your generation-is behind a character, as opposed to just a generated face. Empathy plays a crucial role in films that are driven by emotion, as it adds a layer of authenticity and relatability that is hard to replicate with artificial creations
3:21 I hope we don't offload therapy, teaching, etc. to AI. AI is way too unreliable with basic facts. I've been using Pi, Bing, ChatGPT and others for over a year now and whenever it gives me info, I often double check with Wikipedia, Google, etc. and I often find the info given is just flat out wrong, particularly with dates and numbers and it also confuses facts, conspiracy theories, and on and on. And it can't do math very well. Way too many hallucinations to rely on AI to teach our kids.
It's fine for casual searches where the exact truth doesn't really matter - like asking for an actor's birthday or age - but I've run across these AI answers being wrong as well (when I bother to check, lol).
It's true - while these are simply examples, we do not recommend that people seek out AI as for therapy needs at this time.
I hope you understand at some point AI will be more reliable than asking a human or searching online. There is a reason with each AI model release, they show the results of some tests and they get better.
@@berkertaskiran not with the current methods. They will always be an unreliable human analog.
@@High-Tech-Geek There is no "current method". They are evolving all the time. I've just seen Mustafa Suleyman's TED talk and people are already describing LLMs as "large multimodal model", and not "large language model". Bear in mind that a "human analog" with unlimited compute can very easily become an ASI and removing bias from it would not be the world's hardest task.
@@berkertaskiran multi-modal just means designers allow it to interact with outside controls like your Uber app or the local nuclear power plant's control interface. The method is still just choosing a likely word/action that would come next in most sentences it has been fed.
Using this current method, there is no ability for AI to comprehend or understand or rationalize or think.
And guardrails simply tell it to compare it's output to a list of forbidden rules and not proceed if there's a "match". But unless humans create guardrails that match the infinite combinations of all bad behavior, people guide AI around them fairly easily.
Unless humans come up with a new method for AI, it won't "become more reliable" in the sense that we can rely on it fully. It will always be flawed.
The data we feed AI is flawed and full of bad information (the Internet). AI should never be trusted as factual for critical situations. It confidently tells me false information ALL THE TIME, but most people don't bother to check.
I have seen no proof to suggest the host isn't an AI avatar just reading AI news to me. :)
Hahah! Beep boop beep! Alert, alert! Human has caught on to our mischievous plans!
Bypass alert! In todays video (within a 60 second span).. Vasa starts to bypass the need for (and devalue the work of) actors, animators, therapists, girlfriends, teachers and RUclipsrs. Let's see... Image... check... voice sample... check... Llama 3 for latest AI film news script... check... Curious Refuse channel... check... This self fulfilling prophecy brought to you by just kidding.
At same we start to see more often a lot of oddities and trial and errors so creative artists seems to have a bright future to fix and put everything together
@@hombacomto advance the training of the ai thus eventually making them obsolete. Fully ai cinema will never be taken seriously, I make shorts with ai and I’m just now understanding the depth of its soullessness. Honestly, I literally make my living as an ai video marketer and it truly cannot create anything “new” but it can Frankenstein out some cool shit. CR is honestly pushing the idea of escapist content over intentional personal art which is what I believe the intention behind generative video is partial to.
Everything will be fine! Don't worry :)
Sure, we currently give these tools the seed of an idea (or inspiration) to set them into motion, but we won't really be "needed" for any paying gigs in the very near future. For now, we might be able to give ourselves fluff job titles like "prompt engineer" and "content curator", but the real name is babysitter. Babysitting the work of AI. AI is a baby, but all babies and toddlers eventually grow up and no longer need to be babysat.
Speaking in broad strokes, I think the real crime of using the creative arts as a "harmless testbed" and "proving grounds" for AI, is that this is a large chunk of the people of the world that have a passion for what they do. I don't really know of any passionate accountants, but it would be a scarier thing for the people in power to allow AI to be in charge of the financial system (to bypass their power). Personally, I think that AI is eventually (not too long from now) going to remove a lot of our feeling of purpose in the world. Will my kids spend all their time with no careers, being entertained by AI with little else to do and reason to engage with others and living on a meager UBI while the world runs on an AI autopilot? There has to be some passions, purpose and goals to strive for in life. What will be our purpose if all the hurdles in life have been removed? I know from my childhood that you can only spend so much time hanging out with friends before that is no longer fun (think the last couple of weeks of summer vacation). I hope this is not the kind of future we are handing them, but no one really knows that yet. In the mean time, we probably need more focus on shaping how our post AGI ASI future will look, before we lose our ability to influence that.
BTW... I realize this is your thing currently. You are a really engaging presenter and I wish you well.
@@JGeo1 have you noticed how bored we get with a new AI tool and want more? We have competing tools because there is not one way to do things. Just use the tools and understand their limits. Don’t look down on the knowledge to use them because it’s not only about prompting, you need knowledge to understand the result. We don’t know how to use all tools yet and a machine can’t answer that because it doesn’t understand our needs, it’s just trained on historical data and we change our minds.
what a horrible time to be alive lol
A great time to be alive :)
You need to use some AI on your audio , your mic sounds terrible
No it don’t 😂
We appreciate the feedback - we'll see what we can do!
guys and gals def don't pirate adobe products, nope, don't do it!
Don't do it!
Unfourchnuly, i will get fire at the end of the year
How so?!
I’m starting to think Onlyfans is the only winner in Ai 😂
LOL