Overview of Apple’s HomeView has an app problem
This episode of The Vergecast opens with a deep dive into Apple’s new HomeView smart-home display and why its app strategy feels oddly constrained. From there, the hosts move through a wide-ranging tour of tech’s biggest topics: the renewed race in AI models, the pull of X as a source for tech reporting, and a lightning round that hits Brendan Carr, carrier competition, agentic commerce, smartphone market softness, and software decompilation.
Apple HomeView: powerful hardware, confusing software
The core criticism
The hosts argue that HomeView’s app situation is basically “CarPlay for your kitchen,” not a true standalone platform.
- It appears to support only a limited set of Apple apps and CarPlay-style apps.
- There is no real app store experience.
- Developers are effectively being told: if you build for CarPlay, HomeView will support it too.
Why that feels wrong
The discussion centers on a simple complaint: this is a new device category that should have been treated like a fresh software platform.
- Apple is asking developers to support yet another screen size and interaction model.
- The product is plugged in and stationary, so the usual CarPlay constraints don’t make much sense.
- The hosts compare it to the early Apple Watch: initially a remote for the iPhone, but eventually a more independent computer.
The larger concern
The big worry is that Apple may have built a new hardware category without giving it enough software freedom.
- If HomeView stays tied to the iPhone/CarPlay model, it may never feel like a real shared home computer.
- The hosts wonder whether Apple would have been better off making something closer to a MagSafe-mounted iPad.
AI race: Google, Meta, and the question of product strategy
Google’s Gemini 4 Argon
The episode frames Google’s latest model release as a return to the front of the frontier-model race.
- Google says Gemini 4 Argon is competitive with leading models.
- A major selling point is cost: it’s much cheaper than comparable offerings.
- But the hosts remain skeptical that Google understands the product layer well enough.
Meta’s approach
Meta gets credit for doing something others haven’t: launching a consumer-facing AI product first.
- The hosts see Meta as unusually strong at consumer product design and engagement.
- They point to Meta’s willingness to ship first, then improve models over time.
- Meta’s “flywheel” is clearer: consumer product → model investment → better consumer product.
Google’s problem
Google is still seen as trying to bolt AI onto existing products instead of inventing a new consumer experience.
- Gemini feels like a model more than a product.
- Google is good at distribution but weaker at creating a compelling new AI interaction pattern.
- The hosts argue Google should probably be more aggressive about turning Gmail and other apps into actual AI workflows.
Side stories: OpenAI and Character.AI
Two cautionary tales come up:
- An OpenAI employee allegedly hid a recording device in a conference room, underscoring how intense trade-secret paranoia has become.
- Character.AI’s bankruptcy highlights the danger of these startups accumulating huge amounts of sensitive user data without long-term stewardship plans.
X, Threads, and the return of tech people to Twitter
Why some are going back
The hosts discuss the recent wave of tech journalists and insiders returning to X.
- For AI and startup coverage, X remains the place where many industry sources still post.
- The “important people” in AI are still there, even if the broader audience is elsewhere.
- For reporters, that makes X hard to ignore as a primary source.
Why that doesn’t mean X is “back”
The hosts push back on the idea that X has regained broad cultural importance.
- Outside tech and AI, X has much less influence than it once did.
- Threads, Bluesky, TikTok, Instagram, and other platforms have absorbed much of the old Twitter attention economy.
- X may still matter for a few verticals, but it is no longer the all-purpose public square.
The emotional split
There’s a clear distinction between:
- using X as a reporting tool, and
- participating in X as a user.
The hosts make clear that they can justify reading it for work while still feeling bad about actively posting there.
Lightning Round
Brendan Carr and the execution livestream controversy
One segment skewers FCC Chair Brendan Carr for dodging responsibility when asked whether broadcasters should carry a livestreamed military execution.
- Carr defers to Pete Hegseth and says it’s not an FCC issue if it’s not on broadcast.
- The hosts call out the contradiction: Carr wants power when it suits him, but defers when it gets uncomfortable.
- The whole exchange is framed as cowardly and politically opportunistic.
SpaceX, spectrum, and wireless competition
Another big item is SpaceX acquiring spectrum and moving toward a wireless-carrier play.
- This ties back to the T-Mobile/Sprint merger and the failed Dish “replacement competition” strategy.
- The hosts argue the FCC’s past choices helped create the conditions for SpaceX to step in.
- Even if SpaceX doesn’t become a true national carrier, the mere threat is already rattling incumbents.
Smartphone sales are down
A market report shows a major slowdown in smartphone sales.
- Global smartphone sales declined year over year.
- Samsung and Apple are taking a larger share of a shrinking market.
- Cheap phones are being squeezed out as manufacturers focus on margins and premium devices.
Agentic commerce runs into real-world limits
The hosts discuss banks and major platforms resisting AI agents that try to make purchases or interact with services autonomously.
- JPMorgan Chase and Bank of America are especially wary.
- The central issue is accountability: if an agent makes a bad transaction, who is responsible?
- The hosts argue that websites and services are private businesses, not free-for-all environments for bots.
“Super intelligence” as political theater
They also react to the rebranding of AI as “super intelligence” in Trump-era Silicon Valley discourse.
- The hosts view it as a mix of corruption, cowardice, and dumb branding.
- The term is criticized as empty and unserious compared with “AI,” which at least has some established meaning.
Decompiling software and the future of game modding
Finally, there’s a legal and technical discussion about AI-assisted decompilation of software, especially video games.
- The hosts think this will trigger major copyright battles.
- Decompiling binaries and reusing code will raise serious legal questions.
- Xbox’s willingness to embrace some of this is seen as both strategic and risky.
Key takeaways
- HomeView’s software model is the biggest question mark. Apple may have built a compelling device with an overly limited app strategy.
- Meta seems strongest at consumer AI products. Google may have the better model race story, but Meta has the clearer product vision.
- X still matters for tech and AI, but mostly as a source network. Its broader cultural power is much weaker than before.
- The next wave of tech fights will be about control, not just capability. That includes payments, agents, copyright, and who gets to define the interface.
Notable themes
- Apple’s hardware innovation is increasingly judged by software decisions.
- AI competition is shifting from “who has the best model?” to “who has the best product?”
- Social media relevance is fragmenting by audience and use case.
- Many of the next legal battles in tech will hinge on ownership, access, and platform control.