It’s celebration time at OpenAI! Close on the heels of ChatGPT’s second birthday, the AI powerhouse has announced plans to release new models and features over the next twelve days. “12 days. 12 livestreams. A bunch of new things, big and small. 12 Days of OpenAI starts tomorrow,” the company shared on its official X account.
“
Starting tomorrow at 10 am Pacific, we are doing 12 Days of OpenAI. 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,” posted OpenAI chief Sam Altman on X.
With ChatGPT hitting over 300 million weekly active users, speculation about what to expect from OpenAI next has been rife. The company plans to grow this figure nearly fourfold, reaching a billion users over the next year. According to OpenAI, over one billion user messages are sent on ChatGPT every day and 1.3 million developers have built on OpenAI in the US.
OpenAI Christmas Announcements: What to Expect
1. o1 Full Release
According to a recent report, OpenAI is likely to release the complete o1 reasoning model, which will be multimodal. In a recent interview, Altman described the company’s latest AI model, o1, as being at the ‘GPT-2 stage’ of reasoning development. While the model is still early in its development, Altman said that significant improvements are expected in the coming months.
Altman also mentioned that users will notice rapid improvements in o1 as OpenAI transitions from the o1-preview model to the full release.
2. Sora
OpenAI announced Sora earlier this year but hasn’t made it publicly available yet. However, OpenAI gave early access to visual artists, designers, creative directors, and filmmakers, even engaging with Hollywood. Recently, the Sora API was leaked on Hugging Face and made available for early testing by some artists.
However, shortly after, the Hugging Face page experienced a 502 error due to high traffic. The company became aware of the incident and quickly shut down access.
3. New Voices (including Santa)
According to rumours, OpenAI is working on a new Christmas theme for Advanced Voice Mode. The voice mode icons will feature snowflake animations along with a new Santa voice. OpenAI is expected to add more voices and introduce a feature where users can generate voices as per their needs.
4. AI Agent
OpenAI is set to launch an AI agent called Operator in January 2025. The tool will perform tasks like writing code, booking travel, and automating routine activities in web browsers. OpenAI may preview it during their festive series live stream.
5. Web Browser
OpenAI could also give us a glimpse of its web browser. Darin Fisher, a key creator of Google Chrome and former Google VP of engineering, joined OpenAI as a technical staff member in November 2024.
The product is intended to allow websites to engage with visitors in new ways, enabling conversations similar to those users have with ChatGPT. For example, a user on a clothing retailer’s site could ask for coat recommendations for an upcoming trip, while someone on a cooking site like Bon Appétit could inquire about dishes that pair with tikka masala.
6. Desktop App
OpenAI recently updated its desktop app. This early beta update allows ChatGPT to examine coding apps and provide better answers for Plus and Team users. It assists with coding in apps like VS Code, Xcode, Terminal, and iTerm2, and also includes a voice assist feature to interact with users. The update lets users take screenshots, upload files, and search the web through SearchGPT.
7. New AI Device
Earlier this year, Jony Ive, Apple’s former design chief, confirmed that he’s working with OpenAI CEO Sam Altman on a secret AI hardware project. The two have been collaborating for about a year to create a device that uses generative AI to manage complex tasks more effectively than traditional software.
While the details of the device are still unknown, the goal is to develop an AI product that could change how users interact with AI. The design is being worked on in a newly acquired office in San Francisco, with plans for significant funding by the end of 2024.
8.Vision Fine-Tuning Upgrades
OpenAI is likely to improve GPT-4o’s ability to understand and analyse images. Developers may be able to customise the model for tasks like medical image analysis, satellite image interpretation, or art style recognition. Improvements could include better object detection, a stronger understanding of complex visuals, and better integration of images with text. This would make image-based AI applications more accurate and useful in different industries.
9. AI-Generated Music
OpenAI may unveil a feature to generate music directly from text prompts, similar to existing AI tools like Jukedeck. By analysing musical patterns and structures, this feature could allow users to create original compositions based on their specific preferences—whether for personal enjoyment or commercial use. This would open up new possibilities for content creators, filmmakers, and advertisers who need unique soundtracks.
10. Real-Time Collaboration Features
OpenAI could introduce collaborative features that allow users to work on projects in ChatGPT in real time. These features would enable users to collaborate easily, whether in writing, coding, or brainstorming sessions. This could be especially valuable for remote teams, as the AI would act as a real-time collaborator, suggesting ideas, making corrections, or solving problems as the team works.
Notably, OpenAI acquired Multi to develop the ChatGPT desktop app. Multi allows users to share applications across workspaces, enabling teams to code, edit, and collaborate as if they were side by side, regardless of their physical location.
11. AI Chips
OpenAI might unveil new AI chips or collaborate with chip manufacturers like TSMC or Broadcom. These chips would be specifically built to optimise AI model processing, providing faster, more efficient task execution.
12. GPT-5
Finally, the one we’re all waiting for. OpenAI could release GPT-5, trained using synthetic data from the new o1-series reasoning models. Earlier this year, in a podcast, Altman also spoke with Bill Gates about how GPT-5 would focus on customisation and personalisation. “The ability to know about you, your email, your calendar, how you like appointments booked, and connect to other data sources—all of that. These will be some of the most important areas of improvement,” Altman said.
Furthermore, Altman claimed that GPT-5 would have much better reasoning capabilities. “GPT-4 can reason in only extremely limited ways. Reliability is also a concern. If you ask GPT-4 most questions 10,000 times, one of those answers might be pretty good, but it doesn’t always know which one. You’d want the best response from those 10,000 each time,” said Altman.
The post 12 Days of OpenAI – The Countdown Begins appeared first on Analytics India Magazine.
