I've been impressed but not surprised by any of OpenAI's announcements so far during its 12-day extravaganza. But we're only a third of the dozen announcements expected between now and December 20.
There are likely still surprises in store and we've yet to see any updates to the GPT-family of models, Advanced Voice or image creation in ChatGPT.
In CEO Sam Altman's own words, the company will "have a livestream with a launch or demo, some big ones and some stocking stuffers."
Sora dropping on day three is probably the biggest of all the announcements so far. It makes OpenAI a new major player in the AI video space. Then, yesterday we got ChatGPT Canvas for all users and part of the main model. This adds a code editor and text editor to OpenAI's already extensive product toolkit.
There are still eight announcements to go and with so much to cover — we're doing it live.
12 Days of OpenAI: Biggest News So Far
- ChatGPT Canvas launch: OpenAI has finally unleashed ChatGPT Canvas, its text and code editor, to all users. It is also being made available for use with Custom GPTs and gets the ability to run Python code.
- OpenAI launches Sora: OpenAI's artificial intelligence video generation tool, Sora, is official and enables you to generate videos and images in nearly any style from realistic to abstract. It's a whole new product for the company on a separate page from ChatGPT.
- Fine-tuning AI models: In a roundtable, OpenAI devs focused on the power behind OpenAI's models and reinforcement fine-tuning for AI models tailored for complex, domain-specific tasks. to make them work in specific fields like science, finance and medicine.
- ChatGPT Pro Tier: Sam Altman and his roundtable continued the 12 Days by announcing a Pro tier of ChatGPT meant for scientific research and complex mathematical problem solving that you can get for $200 a month (this also comes with unlimited o1 use and unlimited Advanced Voice).
- ChatGPT o1 model: OpenAI's 12 Days of AI kicked off with a rather awkward roundtable live session where Altman and his team announced that the o1 reasoning model is now fully released and no longer in public preview.
English (US) ·