正在加载视频...

视频加载失败

Finally I can announce that Luma just released their first video model DREAM MACHINE! I was one of the first to get access and reimagined my Dismal Swamp Trailer. This is image to video only! I’m blown away. And you can try it now! SIGN UP HERE: #LumaDreamMachine #sora...

150,097 次观看 • 2 年前 •via X (Twitter)

11 条评论

Dustin Hollywood 的头像
Dustin Hollywood2 年前

@LumaLabsAI Looking sharp as hell Dave! Fucking love that everyone is about to drop movies 🫶💯😈🔥🎥💻

Dave Clark 的头像
Dave Clark2 年前

@LumaLabsAI Yessir. It’s game time.

Aikura.ai 的头像
Aikura.ai2 年前

@LumaLabsAI Wow so cool! Look at mine!

Chikai 的头像
Chikai2 年前

@LumaLabsAI WOW!!! So well done Dave! I've was so excited to see what filmmakers like yourself would do with Luma as I played with it myself!

Dave Clark 的头像
Dave Clark2 年前

@LumaLabsAI Thanks @lifeofc yeah this was a test to learn how to wield. It. Can’t wait to experiment further!

TheNextHit 的头像
TheNextHit2 年前

@LumaLabsAI Wow, the movement is incredible!

Dave Clark 的头像
Dave Clark2 年前

@LumaLabsAI I was blown away!

Alex Patrascu 的头像
Alex Patrascu2 年前

@LumaLabsAI Top stuff, Dave! 🏆 I'm still editing mine, can't wait to share the trailer.

Dave Clark 的头像
Dave Clark2 年前

@LumaLabsAI Can’t wait to see!

Aze Λlter 的头像
Aze Λlter2 年前

@LumaLabsAI that last shot is gold. Wow.

Dave Clark 的头像
Dave Clark2 年前

@LumaLabsAI Thanks brother. Def the money shot!

相关视频

It's official. OpenAI just launched SORA 2... and an entirely new social app. Truthfully... I wanted to hate this. The idea of an endless turbo slop scroller is just… nightmare fuel to me. BUT. The reason I think Sora (the app) might actually just work, is that it’s a multiplayer meme-creation machine. It’s fun. It doesn’t take itself seriously. I think of it like a super advanced evolution of JibJab. The group chats will love it. Now… personally, I'd guess that I will be churning after the novelty fades… but I wouldn’t be surprised if they built some meaningful network effects here. But let’s talk about the model. Sora 2 is an instant Veo 3 alternative. It actually produces better physics, better audio, and deeper realism imo. It can also generate up to 15 seconds of video. Where it falls short is that it’s limited to 720p right now. We can assume 1080p and 4k exist, but the compute cost associated would probably not be economically viable for OpenAI. Okay back to the app. You can create a model after your own likeness, for yourself OR other people to use as a “cameo.” You have control over who and how someone can use your likeness. I’m “rpn24” on the app. If you have access to the app (it’s invite-only for now) – go ahead and create whatever you want using me (for now). I'm not too worried because OpenAI leaned heavily into moderation for launch. Which is probably smart because I expect the media and general public to blast this at first. They’ve also taken further precautions to protect your likeness. You can not export or save a video with someone else's likeness on it, and screen recording is disabled. Also Sora is testing (first I know of its kind) an algorithm you can prompt.

Roberto Nickson

147,341 次观看 • 10 个月前