ChatGPT maker OpenAI just announced a new "12 days of OpenAI" event, revealing a new model or feature every weekday for the next two weeks.
The AI lab has remained tight-lipped on specifics, but this could include any number of previously hinted at or revealed technologies that have remained hidden away due to anything from misinformation risk to the cost of running it.
Most predictions on social media point towards the AI video generator Sora, the full version of the o1 reasoning model and some form of improved control over the voices in ChatGPT Advanced Voice using Voice Engine.
CEO Sam Altman wrote on X: "Each weekday, we will have a livestream with a launch or demo, some big ones and some stocking stuffers. We’ve got some great stuff to share, hope you enjoy! merry Christmas."
What can we expect from OpenAI?
12 days.12 livestreams.A bunch of new things, big and small.12 Days of OpenAI starts tomorrow.December 4, 2024
OpenAI hasn't exactly been sitting on its hands for the past year. The company shipped GPT-4o, Advanced Voice, o1, ChatGPT Canvas and more since January last year and announced even more.
Sora
o1 model
Advanced Voice Vision
GPT-4o image generation
Canvas upgrades
Video analysis in ChatGPT
Voice Search
What we're likely going to see over the next two weeks, as Sam-ta Clause puts on the big red suit, is that some of those announcements will become products we can use.
Of everything OpenAI could announce during its "12 days" event, Sora is probably the most hotly anticipated. Less so than it would have been six months ago, as we've now got models like Hailuo MiniMax, Runway Gen-3 and Kling, but it's still a big deal, and the recent leak points to it still being at the top of its game.
We're also expecting the full release of the reasoning o1 model. This will likely allow it to get access to ChatGPT memory, GPTs, and live search data, making it much more useful in planning a project or working through complex ideas.
GPT-4o can generate images, but it's a feature nobody can use as OpenAI locks it off. This may be made available over the next 12 days. Early demos of this feature suggested impressive results and control.
Advanced Voice is powerful but can be improved. This could include giving it live search access and other features. We might also see it able to look through your webcam or phone camera to provide real-world analysis.
We might see updates to Canvas, including features hinted at during its initial release, such as stacking different canvas elements within a chat. Inside ChatGPT, we might get other GPT-4o native features, such as vector graphics and video analysis, unlocked, making the chatbot more useful.
We might also get more announcements, such as a preview of o2 or GPT-5o. Outside language models, we might only get Sora-Turbo with a "Sora-full" coming soon. Then you've got Voice Engine, OpenAI's text-to-speech tool and 'operator', OpenAI's rumored but not confirmed agent framework due out next year.
Whatever gets announced over the 12 days of OpenAI, it will usher in a new era of artificial intelligence. It also comes as Google Gemini turns 1, so I suspect we will also get an announcement from Google in the same week — just as I'm preparing holiday features. Where's Skynet when you need it?
