what you need to know
- Meta’s roadmap extends to 2026, with a focus on developing AI models that can power all videos.
- Users who like something in Reels will see more of that content in their feed.
- The company is working to expand the model with suitable hardware and data support.
The artificial intelligence revolution is in full swing, and Meta, the parent company behind Facebook and Instagram, is investing more energy and resources into the technology giant.
As first reported by CNBC, the tech giant is ramping up its efforts in generating artificial intelligence. In a wide-ranging conversation at the Morgan Stanley San Francisco technology conference, Facebook head Tom Alison noted that Meta’s roadmap will extend into 2026 and focus on developing a platform that can serve all videos, including Reels and longer videos. release form) that provides content for supported AI models.
Currently, every Meta product relies on artificial intelligence models. In this case, Reels, Groups, and Home Feed all have their own unique suggestions. Alison said the goal moving forward is to move from a single product model to an entire ecosystem while making user recommendations more compelling.
“For example, we’re working on a project to power the entire video ecosystem with a single model, not just Reels,” Allison said, according to a transcript obtained by Android Central. “Then, Can we add our feed recommended products so that the model serves that as well?”
Alison said Reels accounts for a third of total watch time on Facebook, meaning two-thirds are based on the social media company’s existing profile. A more modern unified approach will pay dividends for users and companies.
Still, there’s still a lot of work to be done, both technically and on the user side. “Now, you see, all of this also requires a lot of hardware investment and planning,” Allison points out. “So, on top of that, frankly, we’re reconfiguring the data center to figure out how to connect more GPUs together.”
To that end, Allison said the company is working to scale the model with the right hardware and data support, in part because billions of user interactions must be analyzed. Currently, Facebook has over 3 billion active users, which means a lot of processing work. Alison said there are also advertising factors and general user behavior to consider.
This grand plan is just a small part of Meta’s overall goals when it comes to artificial intelligence. The company also wants to enhance its chatbot assistant so that it can work with increased suggestions and improve responsiveness.
In one of Allison’s use cases, users who were recommended content about Taylor Swift could leverage Meta Assistant to find out about her next tour dates, rather than waiting for commenters to answer questions or jumping out of the platform to search.
In another case involving groups, users can rely on Meta AI to answer baking-related questions faster than reviewers. Eventually, the AI can send previous posts related to the user’s question for the user to read, and the user can also interact with the AI.
All of this suggests that Meta hopes AI will continue to further engage users, making its platform a one-stop showcase for entertainment, connection and information.