Meta has quietly launched a new app called Pocket, an AI-focused tool that lets users create and share small interactive applications — referred to as "gizmos" — generated from natural language prompts. The Verge reported the development, noting that the name collision with the beloved read-it-later service that Mozilla shut down last year is, by all appearances, entirely coincidental and entirely awkward.
To understand why this moment is more interesting than it first appears, some background is worth establishing. The original Pocket, acquired by Mozilla in 2017, had spent years as one of the more quietly indispensable tools on the internet — a place to save articles for later reading, integrated directly into Firefox and used by millions of people who liked the idea of a slower, more deliberate relationship with online content. Mozilla's decision to shut it down was a blow to that particular corner of the web, and the service retained considerable goodwill even in death. Arriving into that name-shaped gap, Meta is now planting a flag with something that shares essentially nothing with its predecessor except four letters.
The new Pocket is best understood as part of a much broader strategic push Meta has been making into consumer-facing AI products. The company has been aggressive about embedding its Llama-based AI systems into its existing platforms — WhatsApp, Instagram, Facebook, the Ray-Ban smart glasses — and launching standalone experiments at a pace that suggests it is still searching for the form factor that makes AI feel genuinely useful to ordinary people rather than to developers and early adopters. The "gizmo" framing is notable here. Rather than positioning the app as a chatbot or an assistant, Meta appears to be betting that the more compelling pitch for mainstream users is creative output: build a small interactive thing, share it, have it spread. This suggests the company is looking at what has worked on TikTok and YouTube — where the loop of creation, sharing, and discovery drives engagement — and asking whether AI generation can slot into that loop in the way that video editing tools once did.
The competitive context matters too. OpenAI has made significant moves toward letting users build and share custom GPT configurations. Google has been pushing its Gemini integrations into Workspace and consumer products alike. The race is no longer simply about whose underlying model is most capable — that argument is becoming harder to win cleanly, as the frontier models converge — but about whose product layer is most sticky, most social, and most embedded in existing behavior. Meta's particular advantage, one that none of its rivals can easily replicate, is distribution. A product launched inside or alongside Meta's ecosystem has instant access to billions of potential users. Whether Pocket is positioned to use that access, or whether it is a more isolated standalone experiment, will significantly shape its trajectory.
The consequences of this launch will be felt differently depending on who is watching. For developers and independent creators who build small web tools and interactive experiences, an app that lets non-technical users generate comparable output from a text prompt represents a real shift in the competitive landscape, even if the first generation of AI-generated gizmos is likely to be modest in sophistication. For Mozilla, the name overlap is at minimum an uncomfortable footnote to an already painful chapter — the shutdown of the original Pocket was not received warmly, and seeing the name resurface under Meta's banner is the kind of irony that tends to circulate on technology forums for a while. For Meta itself, the launch is low-risk: a relatively small product experiment that, if it finds no traction, costs the company little, but that, if it catches on, could become another data point in the company's argument that it is a genuine AI innovator rather than a fast follower.
What to watch for next is whether Pocket develops any meaningful social dimension — whether the gizmos people build actually get shared and spread in the way Meta's framing implies, or whether the app becomes another entry in the long list of AI tools that demonstrate impressive generation capabilities but fail to find a reason for people to return. The other thing worth tracking is how Meta handles the inevitable comparisons to the original Pocket. The company is clearly not trading on any legacy goodwill from the Mozilla product, but naming choices in technology carry associations whether or not they are intended, and how the market receives the brand will be a small but telling signal about how much the previous Pocket's absence is still felt.
