Meta says AI is speeding up new app launches, more coming

Meta says artificial intelligence is helping its teams ship new consumer apps at a faster pace, and more are on the way. On the company’s second-quarter earnings call this week, CEO Mark Zuckerberg pointed to a recent run of launches — a Marketplace-seller app called Seller, a standalone Facebook Groups app called Forum, a “vibe-coded” gaming app, a new Instagram photos app, and an experiment with AI-generated bedtime stories — as evidence that large language models are changing how quickly the company can test new ideas.

Meta has tried and failed for years to build standalone social apps beyond its core platforms. In its early days as Facebook, the company ran an internal incubator called Creative Labs, which shipped apps including the photo-sharing app Slingshot, the anonymous chat app Rooms, a Flipboard rival called Paper, the Moments photo app, and a video app called Riff — all of which shut down after the effort ended in 2015. In the early 2020s, an internal R&D group called NPE Team tried again with apps such as Bump, Aux, Move, Spark, CatchUp, E.gg, Venue, Hotline, Super, Tuned and BARS. None became a breakout hit either.

This time, Zuckerberg argues LLMs make the process easier. “I expect it to become a lot easier to ship new apps,” he told investors, adding that Meta is planning to build out more ideas and lean on its recommendation systems to help them scale. He also said new consumer products are coming soon, though he did not detail what they are.

Threads is the clearest example so far of that dynamic: Meta calls it a modest hit, with 500 million monthly active users, and Zuckerberg likes to say it will one day become the company’s next app to reach a billion users. Meta seeded Threads’ early growth by pulling in users from Facebook and Instagram and has kept promoting it heavily across both platforms since — but CFO Susan Li said AI-powered recommendations are also a factor, since large language models are increasingly able to improve ranking and recommendation systems by better understanding content and generating training data. She added that LLM-powered agents now help engineers evaluate content quality, spot trends and test ranking changes, and that Meta recently reached a milestone in which every Reel and Feed post on Instagram is automatically processed through an LLM for topic and tone. The company is also building LLM-native recommendation systems that could help future apps scale faster.

Investors on the call were more focused on AI spending and Meta’s enterprise ambitions than on the new-app roadmap, and did not press executives for further detail.

Why it matters

After years of failed standalone-app bets, Meta is framing AI-driven development speed — not a single hit product — as its edge for trying, and discarding, new consumer ideas faster than before.