Meta’s latest pivot, in one keynote
At Meta’s developer event in Menlo Park, Mark Zuckerberg didn’t spend much time romanticizing the company’s old obsessions. The stage was built for Muse, his new AI product, and he treated it that way. The message was plain enough even without a neon sign: Meta wants people to look at the company less as the group that spent years pushing virtual worlds, and more as a company trying to plant itself in AI.
That matters because this was not just another product reveal tucked into a long list of demos. It felt like a reset attempt. Zuckerberg told the room, in effect, that AI now matters more than the metaverse vision he once sold so hard. That’s a noticeable turn for a chief executive who spent years arguing that immersive digital spaces were the future of social life, work, and maybe even leisure time. Now the pitch has changed. The new center of gravity is Muse.
When a company stops talking about its old obsession in public, it is usually trying to bury it under something newer.
There’s a reason that shift drew attention well beyond the usual tech news crowd. Meta does not get to rewrite its own story without a little friction. The company’s public identity still carries a lot of baggage from its metaverse push, and the name change from Facebook to Meta did not exactly wipe the slate clean. Yet at this event, that history was mostly kept offstage. The familiar social apps, the old arguments about attention and addiction, the endless debates around the company’s role in digital culture, all of it stayed in the wings.
That choice looked deliberate. By centering Muse and keeping the rest of Meta’s empire in the background, Zuckerberg was trying to do more than announce a product. He was trying to change the frame. If the metaverse era was about betting that people would one day live inside more simulated spaces, the new pitch is that AI can sit closer to daily routines, actual habits, and the sort of tasks people may gladly outsource. It is a less theatrical story, and probably an easier one to sell.
The question, though, is whether a new AI product can really help Meta move on from the metaverse era, or whether it just gives the company a fresh label for another reinvention. That question hangs over the event because Meta has spent years asking audiences to accept a new identity, then asking again when the first version wore thin. In tech policy circles, that can look like adaptability. In power and politics terms, it can look like a company testing how much reinvention its scale will allow. In the ordinary world, it can look like a very expensive way to ask for a second act.
Muse was presented as that second act. The presentation style mattered almost as much as the product itself. Zuckerberg spoke as if the company had already moved on, even though the rest of the business is still very much there, doing what it has always done, only quieter for the moment. The real tell was what didn’t get stage time. No grand tour through Facebook’s aging reputation. No long victory lap around Instagram. No attempt to relitigate the metaverse in public.
So the announcement landed less like a routine launch and more like a declaration of intent. Meta wants Muse to carry the company’s next chapter, or at least begin the job. Whether people buy that story is another matter. The next sections make clear why skepticism has plenty of historical support.
The company has tried this kind of reinvention before
That new Muse pitch did not appear out of thin air. Meta has spent years trying to convince the public that the next big thing is always just around the corner, and its track record explains why people might keep one eyebrow raised.
Back in 2016, the company got unusually excited about Messenger bots. The idea was simple enough: if people already spent time chatting inside Messenger, why not let brands build little automated helpers that could answer questions, take orders, and send updates? It sounded tidy, almost elegant. 1-800-FLOWERS built one. Media outlets built them too. A user could, in theory, get a weather update, read headlines, or place an order without leaving the app. On paper, that looked like the future of conversational commerce. In practice, many of those bots were clunky, limited, and a bit too eager to prove they were smarter than they actually were.
Meta has a habit of presenting a new interface as the next default way people will use the internet. The harder part is getting people to keep using it after the first demo.
The bot rush did not last long. Within a couple of years, Messenger’s bot ambitions had been trimmed back, and the platform was wound down. Some bots disappeared quietly. Others lingered, half-functional, like digital vending machines that had run out of snacks. The bigger lesson was hard to miss: a platform can announce a new behavior all it wants, but that does not mean people will make it part of their routine.
Then came the next left turn. David Marcus, who had been leading Messenger, left to chase Meta’s cryptocurrency dream. That project went through a couple of names and a mountain of scrutiny, but the basic promise stayed the same: make payments work inside the company’s apps, then somehow expand from there. Libra became Diem, and Diem never became the clean, consumer-friendly product the company seemed to imagine. Regulators took a close look. Partners drifted. The whole thing ended up as another reminder that Meta can move fast at the announcement stage and slowly, if at all, when the real-world complications show up.
By the time the metaverse arrived, the pattern was already familiar. Mark Zuckerberg sold virtual worlds as the next mass-market internet shift, complete with avatars, headsets, and a future where work, play, and social life would all blend together inside digital spaces. The company even changed its name to Meta, which was a serious act of branding confidence, the sort of move that suggests the old story is over and the new one has already won. Reality Labs absorbed enormous sums. The hardware got better. The demos got slicker. Adoption, though, stayed stubbornly narrower than the company hoped. Most people did not wake up and decide they needed a virtual office, a virtual concert, and a virtual living room before lunch.
That history matters because Muse is not just another product launch. It is another attempt to tell investors, developers, and users that the old obsession no longer needs to dominate the conversation. First bots. Then crypto. Then the metaverse. Now an AI agent. The line keeps changing, but the underlying move is the same: find a new story that sounds useful enough, modern enough, and sticky enough to justify the scale of the company’s ambitions.
The money behind the reboot is hard to ignore. Meta is preparing for a huge spending year, and Reality Labs has already burned through well over $80 billion since 2021. That kind of number does not merely describe product development. It creates pressure. When a company has spent that much on one future and the public still isn’t fully living in it, the next future has to arrive quickly and behave itself.
That is why the current AI push feels so loaded. Meta is not pitching Muse in a vacuum, and it is not exactly starting from a position of innocence. The company now has to persuade people that this time the new thing is not just a glossy demo with a short shelf life. It has to sound practical. It has to sound durable. It has to look like something more than a fresh coat of paint on a very expensive habit.
The company has already laid out the public paperwork for the reboot, with a Muse personal AI agent announcement, a post on Muse image and video tools, and its security and safety approach for AI agents. That kind of documentation reads differently from the old bot hype, at least on first glance. It suggests a more careful rollout, or at minimum a company aware that it has burned through a fair amount of goodwill on grand promises.
Still, the audience has reasons to be wary. Meta has already asked people to imagine a social app revolution, a payments revolution, and a virtual-world revolution. Now it wants them to picture an AI agent as the thing that finally sticks. Maybe this time the pitch lands. Maybe not. But the company has earned the skepticism.
What Muse actually is — and why Zuckerberg is betting on it
At Meta’s developer conference, Muse was presented as the thing the company wants people to talk about when they think of its future, not the metaverse, not another headset demo, and not a rerun of old social-app drama. In the company’s own framing, Muse is a personal agent that sits on top of your digital life and does the tedious bits you’d rather skip. Meta’s Connect 2026 recap describes it as a tool that can connect to services you already use for work, banking, and other accounts, then take action across them. The pitch is less “chatbot with opinions” and more “digital assistant with a clipboard.”
That clipboard matters. Muse is designed to handle the stuff people actually procrastinate on: filling out forms, making bookings, paying bills, and digging through the little administrative jobs that pile up in a tab graveyard. A dinner reservation, an expense form, a utility payment, a travel booking, a boring work request that needs three logins and a code sent to your phone. Meta clearly wants Muse to feel less like a novelty and more like a practical layer between the user and a mess of services that were never designed to be pleasant.
Muse’s most obvious appeal is also the least glamorous one: it saves time on chores that were already annoying before AI got a marketing budget.
There’s a catch, of course. The product still sounds more like a manager than a magician. You hand it a task, it goes off to do the paperwork, and then you check whether it got the details right. That makes Muse feel useful in a very ordinary way, which may actually be the point. A lot of consumer AI has spent the last two years trying to sound mystical. Meta seems to be betting that people would rather have a tool that quietly does the job than one that writes a poem about doing it.
The company also has bigger plans for how people interact with it. Voice input is on the roadmap, and Meta has already pointed to work on Muse Voice Transcribe as part of that push. In plain English, the idea is that you’ll be able to talk to it more naturally instead of typing every command like you’re filing a complaint with a very cheerful receptionist. That could matter a lot if Muse is meant to live inside Meta’s broader hardware and software stack, including the virtual-reality operating system the company keeps building around its headsets and upcoming devices. Meta has been explicit about pushing the assistant deeper into its platform work on its technologies page, which is where the long game starts to come into focus.
And then there’s the hardware. Muse Charm, the new device tied to the product, is supposed to arrive later this year. It’s small, meant to sit near the wrist rather than take over your whole desk, and it has a screen. Meta hasn’t said what it will cost yet, which is usually either a sign of pricing flexibility or a sign that the company wants everyone to keep staring at the shiny bit first. The charm-sized form factor gives Muse a very specific role. It isn’t trying to be a phone replacement. It looks more like a companion interface, something you glance at, tap, and speak to while the heavier lifting happens elsewhere.
That setup fits Meta’s broader hardware push better than a lot of its earlier ideas did. If the company wants Muse to feel normal, the assistant can’t live only inside a desktop window or a web page. It needs to move with the user. That is where the wrist-adjacent device, the headsets, and the work on Meta VR glasses start to connect. The pitch is that the assistant can follow you from a phone to a headset to whatever comes next without forcing you to relearn the whole setup every six months. Whether people want that much Meta in their day is a separate question, and not a small one.
Zuckerberg talked up broad interest in the product, which is exactly what a CEO is supposed to do when introducing a new toy. The early numbers were more modest. In the first week, usage was only in the low hundreds of thousands. That is not nothing, but it is also not the sort of figure that makes anyone reach for confetti. For a company with Meta’s reach, the number suggests curiosity more than habit. People tried it. They poked around. They probably asked it to do a few errands and then checked whether it got the confirmation email.
That gap between the pitch and the early uptake tells you plenty about where Muse sits right now. It has a clear use case, which is rarer than it should be in the AI market. It also has a familiar Meta problem: the company can get attention fast, but turning attention into daily behavior is harder. A flashy demo at a developer conference can pull people in for a week. Paying your electricity bill through a wrist gadget on a Thursday night is a different kind of commitment.
Still, Muse is not trying to be a moonshot in the old metaverse sense. It sounds more grounded than that. If Meta gets this right, the assistant becomes a habit, tucked into the boring seams of everyday life. If it doesn’t, Muse may end up as one more thing people admired for ten minutes before going back to old-fashioned tabs, passwords, and frustration.
The real test: better hardware, or just a better story?
If Muse is the pitch, the VR Glasses demo was the evidence that Meta still knows how to build hardware people might actually want to wear. That matters, because the company has spent years asking the public to care about a virtual future that often felt more like a corporate mood board than a product plan. This time, the device itself did some of the talking.
The headset Meta showed off was impressively light. The frame came in at roughly 100 grams, which puts it in the realm of “wait, that’s it?” rather than “how long before my face files a complaint?” Most of the computing sat in a separate puck, so the weight on the head stayed low and the device ran cooler than the bulkier headsets people have come to expect. Battery life landed at about three hours, and for longer sessions it could stay plugged in without turning into a forehead space heater. That kind of practical detail sounds boring until you remember how many headset demos fall apart the moment a user wants to wear them for more than ten minutes.
A lighter headset changes the conversation. It does not settle it.
In the demo, the display was also better than the usual VR party trick. It looked surprisingly close to Apple’s Vision Pro, at least from the outside looking in, even though Meta’s version is lighter and, by all accounts, much cheaper. That price gap matters. Apple’s headset has the polish, but it also has the sort of price tag that makes people pause, blink, and maybe close the shopping tab. Meta seems to be betting that a more affordable device with strong visuals will do the harder job: getting people to wear it often enough to form a habit.
The demo scenes were chosen carefully, which says a lot about where Meta wants this to go. Concert clips gave the headset a social angle. Basketball highlights made it feel like something you might use for sports without dragging out a giant TV. Games showed the obvious entertainment case. A surface keyboard, meanwhile, hinted at work, or at least the promise of doing more than zoning out inside a floating app grid. The message was pretty plain. Meta does not want this to be filed away under “fun novelty I forgot I owned.” It wants the device to feel useful enough that people reach for it without a speech from Zuckerberg first.
That’s where the tech news gets interesting, because the company’s best argument is no longer just a metaverse sermon. It’s showing hardware that appears more comfortable, more portable, and less ridiculous on paper than the previous wave. Meta can point to real engineering progress here. It can point to a device that looks less like a lab demo and more like something a normal person might wear for a commute, a flight, or an evening on the couch.
Still, the bigger problem hasn’t gone away. Better hardware can make the product easier to buy. It can make demos look cleaner. It can even make skeptical people stop rolling their eyes for a minute. What it can’t do on its own is force new habits into people’s lives. That part is much messier. AI agents have to prove they’re useful beyond a tidy keynote demo. Headsets have to earn repeat use after the first burst of curiosity fades. Meta has been through enough grand pivots by now that another polished reveal won’t be enough on its own.
So yes, this looks like a real step forward in device design. It also looks like the start of a harder question: can Meta turn these gadgets into routines, or will they join the long list of launches that made a splash and then drifted into the archive?
