It appears OpenAI’s announcement of their new text-to-video model Sora has spooked the industry.
The photorealism of the examples provided by OpenAI has left many scared that the technology will be disastrous for the creative community, but few have paused to reflect what this might mean for billion dollar companies whose entire lunch is at stake.
Adobe, being the biggest player in this field, appears to have been caught pants down. The $244 billion dollar behemoth has been the go-to provider for video and animation software for years, and the emergence of Sora and other similarly capable models could present an existential threat to this side of their business.
Not to mention that even their non-video business like Photoshop are also under attack from startups like Midjourney.
Is this death by a thousand cuts, or is Adobe going to react?
It turns out they have a plan. They are going to compete directly with Sora and similar upcoming platforms in Generative AI offerings.
Gautham Mysore, the Head of Audio and Video AI Research at Adobe, on Tuesday introduced CAVA (Co-Creation for Audio, Video, & Animation) Research organization in Adobe Research.
The team is made up of 50 researchers working on all aspects of video creation – video, music, speech, sound effects, animation, and interaction design.
Text-to-video and text-to-audio models out there are getting dramatically better, as expected. In addition to developing models, we focus deeply on control of these models for real use cases by leveraging our deep understanding of video creation at Adobe. It enables us to both effectively augment workflows of today and invent workflows of tomorrow. This control will enable creators to craft their story and iterate on it from high level story arcs to fine-grained detail, to truly co-create with AI. Having worked on audio for a long time, I understand how critical high quality audio is for great video, how it brings the story to life, and how hard it is to get it just right. Speech, music, and sound effects each have their distinct roles in video and we work on all of them. We aim for it to be a given that all video will have amazing audio. Gautham Mysore
Although this is still early days, we are excited to see what Adobe will cook up, and whether they will survive the upcoming AI take over.
