Summary of Everybody wants to rule the AI world

Overview of Everybody wants to rule the AI world

This Vergecast episode is a wide-ranging tour through the week’s AI chaos, led by a deep dive into the Elon Musk vs. OpenAI trial and the revealing text messages, journals, and emails coming out in discovery. David Pierce and Nilay Patel also explore why OpenAI may be building a phone, how Apple’s Siri/Apple Intelligence missteps are now turning into legal payouts, and why Google, Fitbit, and even a weird robot-companion startup are all chasing the same AI-driven hardware future.

The OpenAI v. Elon Musk Trial: “It’s all in the computer”

What the hosts focused on

• The trial’s second week was framed as even more chaotic than the first.
• A major theme: because executives now document everything in tools like email, journaling apps, Slack, and AI systems, discovery in future lawsuits will be explosive.
• The hosts argue that the whole AI era is generating a paper trail of corporate self-dealing, insecurity, and infighting.

Key trial takeaways

• OpenAI’s early structure and motives are being revisited:
  • The company began as a nonprofit with a safety-first story.
  • Evidence suggests Sam Altman and Greg Brockman were involved in complicated side deals and financial entanglements, while still presenting OpenAI as mission-driven.
• Elon Musk’s position looks conflicted:
  • He appears to have wanted control of OpenAI, including possibly folding it into Tesla.
  • When he lost that battle, he walked away and later became a fierce critic.
• The trial is full of memorable text messages, especially:
  • Mira Murati telling Sam Altman his situation was “directionally very bad.”
  • Sam asking if the board’s decision was to fire him “or some other thing?”
  • The now-famous line from Murati: “new guy is a rando Twitch guy” — referring to Emmett Shear as interim CEO.
• Satya Nadella emerges as the least chaotic adult in the room:
  • The hosts repeatedly joked that he avoids paper trails and handles the situation like a seasoned public-company CEO.
  • Microsoft’s lawyers repeatedly reminded the court that “Microsoft wasn’t there,” which became a running joke.

Big legal/business takeaway

• The case is less about one clean moral story and more about a messy collision of ambition, self-dealing, safety rhetoric, and power struggles.
• The hosts suggest Musk’s case may still uncover real misconduct, but also emphasize that OpenAI’s side is far from clean.

Why OpenAI building a phone actually makes sense

The core argument

The hosts argued that an OpenAI phone is both:

• Almost certainly doomed, and
• Probably the only move OpenAI can make

Why it might happen

• OpenAI wants a device that captures constant context: voice, camera, surroundings, behavior, meetings, and actions.
• An AI-first device needs to avoid Apple’s App Store and platform restrictions.
• If the future is agentic AI, the hardware has to be built around that future from scratch.

Why it likely won’t work

• Today’s AI models still can’t reliably do the things people want them to do.
• A phone can’t replace the iPhone unless it can match the iPhone’s app ecosystem, camera quality, and everyday utility.
• Without apps like Instagram, Uber, and messaging, people will keep carrying their old phones.

Bottom line

• The phone is a logical strategic move, but the hosts think it’s likely to fail because the underlying AI technology still can’t deliver the promise.

Apple, Siri, and the cost of fake AI promises

The lawsuit settlement

• Apple settled a class-action lawsuit for $250 million tied to Siri and Apple Intelligence marketing promises.
• The episode played a clip from Apple’s ad featuring a fake scenario where Siri could identify someone from a past meeting — a feature that didn’t exist.

The larger point

• Apple was pressured by the industry narrative that AI would replace the smartphone.
• Instead of shipping a truly capable AI assistant, Apple:
  • Overpromised,
  • Ran misleading ads,
  • And is now paying for it.

Broader significance

• The hosts argue Apple actually got lucky:
  • Its AI failures were so public that the company is now being forced into a better long-term strategy.
  • Apple’s dominant platform position still gives it a massive advantage over AI-first hardware startups.

Google’s Fitbit Air and the rebrand to Google Health

What it is

• Google launched a new Fitbit Air: a screenless, modular fitness tracker.
• It’s designed to be a cheap, approachable wearable that feeds data into Google’s AI health systems.

Why the hosts liked it

• A simple fitness tracker is a sensible product.
• The price point is accessible.
• The data analysis layer is where Google’s strengths can matter.

Why they disliked the branding

• Renaming Fitbit to Google Health was seen as a bad move.
• It carries too much “Google baggage” and raises privacy concerns.
• It makes the product sound more invasive and less trusted.
• It also signals that the real product is not the tracker, but the data pipeline behind it.

The larger concern

• Google likely wants to use health data as context across Gemini and other services.
• The hosts see this as part of a wider trend toward AI wellness coaching that often feels intrusive, obvious, and unhelpful.

Familiar Machines and Magic’s “Familiar”: the creepy AI pet robot

What it is

• A new robot called Familiar, made by the founder of iRobot, is intended to be a companion-like embodied AI device.
• It looks like a strange blend of:
  • robot,
  • dog,
  • polar bear,
  • and alien mascot.

The thesis

• The company’s idea is that the next wave of robotics is about human connection, not just utility.
• It’s meant to be pettable, cute, and emotionally engaging.

The hosts’ reaction

• They were fascinated, but largely creeped out.
• They thought the form factor looked unsettling and unnecessary.
• They were skeptical that “make it look like a pet” is the right lesson to take from Roomba’s success.

Their takeaway

• The product is interesting as a research and data-collection vehicle.
• But the emotional design feels off, and the “AI companion” angle is still deeply questionable.

Lightning round highlights

Nintendo’s new Star Fox game

• Nintendo announced a new version of Star Fox 64, effectively the fifth version of the same game.
• The most bizarre feature: a face-tracking mode that turns the player into a kind of VTuber/fursona version of Star Fox.
• The hosts were both horrified and delighted.

Zach Galifianakis’ gardening show

• Netflix has a short-format gardening series with Zach Galifianakis.
• It’s earnest, educational, and weirdly wholesome rather than sarcastic.
• The hosts enjoyed that it feels like a throwback to oddball 90s TV and web content.

Brendan Carr and the FCC: anti-broadband, pro-corporate

Main criticism

• Brendan Carr cheered a court ruling that struck down FCC broadband equity rules.
• The hosts framed this as classic Carr:
  • anti-regulation,
  • pro-telecom power,
  • and uninterested in actually helping underserved communities get internet access.

NFL antitrust exemption

• Carr also got involved in broader debates about whether the NFL should keep its antitrust exemption.
• The Green Bay Packers warned that changes to the current model could threaten their existence.
• One of the hosts disclosed that he owns shares in the Packers, which made the topic especially personal.

Dreamy: Chinese vaporware at maximum volume

What Dreamy is doing

• Dreamy, a Chinese company that started with robot vacuums, is now announcing:
  • smartphones,
  • smart rings,
  • smart glasses,
  • laundry robots,
  • lawnmowers,
  • and a rocket car.

Why it stood out

• The company is making spectacular, almost absurd claims with a straight face.
• The rocket car, in particular, was mocked as physically implausible and maybe one of the most extreme examples of vaporware ever attempted.

The hosts’ reaction

• They admired the ambition and the showmanship.
• They were also deeply skeptical that most of these products will ever ship in a meaningful form.

Elon’s latest empire move: SpaceX AI

What happened

• Elon Musk is further consolidating his companies.
• XAI appears to be getting folded into SpaceX, with the resulting structure being referred to as SpaceX AI.

Why it matters

• The hosts treated this as another example of Musk’s tendency to reorganize and rename corporate entities in ways that are confusing to outsiders and useful mainly to him.
• It fits the broader theme of the episode: giant AI entities with messy, overlapping structures and unclear accountability.

Closing note: Microsoft and Xbox are doing the opposite

Asha Sharma’s approach

• The episode ended with praise for Asha Sharma and the Xbox team for removing bad AI features rather than adding more of them.
• The hosts saw this as a good example of product leadership:
  • fewer gimmicks,
  • less clutter,
  • more obvious usefulness.

Big-picture takeaway

This episode argues that the AI era is being shaped less by clean technical breakthroughs and more by:

• power struggles,
• product overpromising,
• privacy tradeoffs,
• legal discovery,
• and companies trying to build around platform gatekeepers.

If there’s one recurring message, it’s this: everyone wants AI to be the future, but most of the industry still can’t make the future actually work.