OpenAI Announces Sora 2 AI Video Model

I'm sure you guys have heard by now but OpenAI announced their 'Sora 2' video model yesterday. I watched their live stream (https://www.youtube.com/live/gzneGhpXwjU?si=2ubb0q6OfvgpLPh8) and some crazy stuff. This is the big competitor to Google's Veo 3.

Here are some highlights from this page (https://openai.com/index/sora-2/)

Stronger physics and realism: Sora 2 models the real world more accurately, including natural mistakes like a ball rebounding instead of teleporting.
Audio generation: adds realistic speech, sound effects, and background soundscapes alongside video.
Personalized cameos: users can record themselves once, then drop their likeness and voice into any Sora-generated video.
New Sora app: a social iOS platform where people can create, remix, and share videos with friends.

The new Sora app looks like a TikTok clone and I'll be honest I think this is a net negative for society when we have millions of people doom scrolling AI slop.

What do you guys think about Sora 2?! I need an invite code to access Sora 2 so when I get that I'll plan on making a YouTube video for it.

4 comments