Meta: AI speeds up development of new apps, with more on the way
According to CEO Mark Zuckerberg, Meta uses LLMs to develop new apps faster (Forum, Seller, Instagram Instants and others) and is preparing more products. The app Threads already has 500 million monthly users.
According to CEO Mark Zuckerberg, Meta uses large language models (LLMs) to develop and test new apps faster. On the second-quarter earnings call, Zuckerberg said that AI had made it significantly easier to bring new products to market, and the company therefore plans to test more ideas at a faster pace and use its own recommendation systems to scale them. In recent months, Meta has launched several new offerings: the Instagram Instants feature, a standalone app for groups called Forum, a standalone marketplace app called Seller, a so-called “vibe-coded" game, a new photo app for Instagram, and an experiment with AI-generated bedtime stories. According to Zuckerberg, more new consumer products are expected to follow in the foreseeable future.
Meta has repeatedly attempted similar experiments without lasting success – in the past, it ran the internal incubator Creative Labs (apps such as Slingshot, Rooms, Paper, Moments or Riff, shut down in 2015) and later NPE Team (apps including Bump, Aux, Move, Spark, CatchUp and others), but none of these apps gained traction, and they were gradually discontinued. The only exception so far is the app Threads, which now has 500 million monthly active users; Zuckerberg has repeatedly expressed his expectation that it will one day become a product with a billion users.
CFO Susan Li said that LLMs are delivering “significant gains" for the company in content ranking and recommendations – according to her, the models improve the accuracy of existing systems by understanding content better and generating higher-quality training data, while AI agents help development teams assess content quality, monitor trends and test changes to ranking. According to the company, every Reel or Feed post on Instagram now automatically undergoes LLM analysis of its topic and tone. The company is also developing recommendation systems built natively on LLMs, which are expected to help scale future new apps as well.
Investors on the call did not ask further questions about the planned new apps, focusing instead on AI spending and the development of enterprise activities at Meta.
Why it matters
The report provides a concrete example of a company that, after years of unsuccessful attempts with new consumer apps (Creative Labs, NPE Team), now claims that deploying LLMs enables it to develop and test products faster while also improving recommendation algorithms through automated content analysis. For the industry, it is one piece of evidence showing how large platforms are moving AI from the experimental phase into the core of product development and content personalization.
Relevant practical impact
What this means
For a business
The case of Meta illustrates a broader trend in which LLMs shorten the time needed to develop and test new products while also improving recommendation systems through automated content analysis – relevant context for companies considering deploying AI in product development or personalization.
DevelopmentCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.