The Wait is Over! Gen-3 is OUT! - First Testing & Impressions
28,069
Published 2024-07-01
▼ Link(s) From Today’s Video:
Gen 3: runwayml.com/blog/introducing-gen-3-alpha/
Gen 3 Prompt Guide: help.runwayml.com/hc/en-us/articles/30586818553107…
Emad's Post: twitter.com/EMostaque/status/1807839613014159700
Dor Bros Post: twitter.com/thedorbrothers/status/1807782470164590…
Gen 3 Text: twitter.com/blizaine/status/1807584743069790650
Always Generating: twitter.com/notiansans/status/1807629033028407754
► MattVidPro Discord: discord.gg/bQgcbjs2Sg
► Follow Me on Twitter: twitter.com/MattVidPro
► Buy me a Coffee! buymeacoffee.com/mattvidpro
-------------------------------------------------
▼ Extra Links of Interest:
AI LINKS MASTER LIST: www.futurepedia.io/
General AI Playlist: • General MattVidPro AI Playlist
AI I use to edit videos: www.descript.com/?lmref=nA4fDg
Instagram: instagram.com/mattvidpro
Tiktok: tiktok.com/@mattvidpro
Second Channel: / @matt_pie
Let's work together!
- For brand & sponsorship inquiries: tally.so/r/3xdz4E
- For all other business inquiries: [email protected]
Thanks for watching Matt Video Productions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!
All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.
Timestamps:
00:00 Introduction to Runway's Gen 3 Alpha
00:41 Exploring the Gen 3 Alpha Interface
01:00 First Video Generation Attempt
02:14 Complex Prompts and Their Results
06:29 Using AI to Enhance Prompts
09:06 New Prompt Ideas and Generations
14:15 Settings and Custom Presets
18:45 Community Reactions and Comparisons
21:27 Conclusion and Upcoming Live Stream
All Comments (21)
-
JOIN ME FOR A LIVESTREAM TOMORROW WHERE I GEN YOUR PROMPTS! https://www.youtube.com/watch?v=mksEAr6MRc0&ab_channel=MattVidProAI
-
Hollywood artists breathe a sigh of relief. Safe for another day
-
conclusion : let's wait for GEN-4
-
I love your channel. I have been following you for years, and the thing I love most is that you are always quick to post news and always put links in the video description.
-
"Imagine being able to play a video game where what you see is generated in real time". This is exactly what the CEO of Nvidia says is the future. Pixel by pixel generation processed by a GPU, no game assets, pre-recorded audio / video. Purely generated and I assume also stored in the background in case you wanted to "play" that particularly generated game, given each new generation would likely be wildly different and any story line would be also wildly different.
-
Real question. What am I supposed to do with this? The footage is unusable except to look at and say "wow, neat". I'm not going to pay $15 for 60 seconds of semi-coherent video that isn't even what I asked for.
-
Really disappointing. Yeah the quality has improved but the prompt understanding is still absolute crap, it ignores like 90% of my instructions. And you cannot import images so .. yeah .. you can make fun little clips by accident or whatever but I would not know how to be able to get what I have in my mind with such a bad prompt interpreter.
-
Luma has a huge advantage, being able to import midjourney/dalle images. Both are extremely expensive.
-
Come on Matt, it's utter shit and you know it
-
This is very disappointing when you see the ad for it Especially when it's so damn expensive
-
I've been getting some great generations but the pricing is rough, nearly $15 for a minute of footage if everything you get is usable which is unlikely unless you get a more expensive plan or subscribe for a year.
-
I honestly do not agree that this is the best we have access to. I feel like we got in general alot more coherent and usable shots with Luma from what I have seen so far. It seems you can get some really good shots that look better than Luma, but those seems so incredible rare that there is no way you could do that without a unlimited plan and alot of rerolls. What I hope is that Kling releases their model to the world soon, as that one actually seems to be alot better than both of these video models.
-
I think it would be interesting to use a suite of a.i tools. Train a GPT on 3D objects, and use another GPT trained on skins and textures and use another GPT trained on animation. Put them all together and automate anything.
-
Seems like another let down vs the teaser videos they've been putting out
-
12 dollars a month for 62 seconds of video? laughable
-
Thank you so much for the shoutout Matt! 🙌
-
This is really exciting!!!
-
waiting for open souce community to build up something to beat it
-
Instructions - if only it was able to follow them…
-
That alien handshake 😂 is still cracking me up. Awesomeness. The 2nd one was nice too. Great hand nuke too. Great experiment.