Apple May Have Their Television After All...
Vision Pro's NBA Game • Thinking Machines Drama • Wikipedia Turns 25 • The Apple Card Fiasco • Meta Beyond Metaverse • Big Tech CapEx Perspectives • OpenAI/Cerebras Deal
I finally dusted off my Vision Pro and watched the Lakers/Bucks game in the 'Apple Immersive' format. Once I got past the laborious nature of doing so (including working past an error in the NBA app that didn’t alert me was an error, it just hung there until I figured it out), it was… pretty great actually. Obviously there are issues, but they’re largely around presentation choices Apple is making with the broadcast — and really, they should let you, the viewer have more options — but it’s really quite something to behold. And if Apple can figure out a way to get more people to behold such things… There’s a path to success here. But I’ve been saying that from the get go…
🏀 The Vision Pro Slam Dunk
Their immersive NBA stream highlights how close Apple is to making Vision Pro content work...
I Wrote…
🎭 OpenAI Transfers Their Drama IP to Thinking Machines Lab
AI startups are losing co-founders like they’re OpenAI...
I Note…
🌏 Wikipedia Turns 25
While it may feel like it launched alongside the internet itself and has always been with us, it actually launched on January 15, 2001 — well after Google and a number of other names you know. And what an interesting time to celebrate the birthday, with the site under assault from AI. At the same, the site is also obviously one of the main feeders of AI content — and they seemingly now have a way for those building out AI to pay for it, with Microsoft, Meta, Amazon, Perplexity, Mistral, and yes, Google now paying for special API access to the site’s content. How much? Undoubtedly not enough, certainly not to end the donation pop-ups. Of course, they’ve been under assault before — remember Google Knol? Oh, you don’t? It’s long since dead but you can read about it you-know-where. Here’s to another 25. [Verge]
💳 Behind the Scenes of the Apple Card Break-Up
A good look behind the scenes of this (finally resolved) shitshow — it’s pretty damning of both Goldman and Apple — I mean, did Apple really not understand how credit works? More likely, they thought they could disrupt the market, just as they had done with the iPhone, to the point where they’d work with one willing partner to do things that others wouldn’t dare. So in this scenario, Goldman was AT&T. Unfortunately, Apple’s notion of approving a credit card for almost any customer that asked — they were Apple customers with iPhones after all — ran head-first into the realities of sub-prime lending. The situation grew bad enough, even before the CFPB came calling, that it may have soured Goldman on the credit card business entirely. Their literal loss — selling the $20B book at a 7% discount — may be JPMorgan’s gain, as the deal hit last-minute hiccups in part because the business, after much refinement, may actually be profitable now. But the relationship damange between Apple and Goldman was beyond repair. Oof. It is a great product offering! But over 30% of the balances were sub-prime?! [WSJ 🔒]
🥽 A Rough Day for the Metaverse
Some good balance by Alex Heath I think to the Meta egg-on-face Metaverse situation. Look, they tried. But despite billions upon billions (and years upon years) spent, the VR market, for the 40th year in a row, just isn’t there (yet?). Meta controls the market for headsets, but it’s very much a king of the hill situation. Without new hardware from Meta to prop it up a bit, that market shrank to a seven-year low (!) in 2025. And that was with the Vision Pro being only two years old (and with an updated one launched at the end of that year). If Apple can’t make it a market… So it’s embarrassing narratively for Meta to pull back — namely due to the whole name change thing! — but the AI race gives them the perfect cover to do so. And the AI spend gives them the real reason to do so now. And they still have other Metaverse-y avenues to explore, namely in mobile, and AI can help fuel some of that. But come on, this is not the Metaverse Zuck promised, clearly. And yes, Roblox or Fortnite may get to the software one first. But someone eventually has to do the purpose-built hardware for such worlds, right? Right?! [Sources]
I Spy…
While the FT article on the Apple/Google Gemini deal gives us the Big Tech CapEx charts through 2024, I used my new friend Claude to plug in the 2025 numbers and even extrapolate out to 2026. Wild to see Apple in this context.
The article also got some chatter for the reporting that OpenAI may have turned down Apple when it came to powering the new Siri, which Mark Gurman pushed back against. But if you read the wording, both things can likely be true: OpenAI may not have wanted to do a white label model and Apple may have only partnered with someone willing to do that. That doesn’t mean OpenAI turned them down explicitly for this task, necessarily. Maybe it was discussed around the original partnership…
Asides…
Another day, another big new AI chip deal for OpenAI, this time with startup Cerebras, known for their literally big ass chips, which is said to help with speed for doing inference. [NYT]
Clearly not coincidentally, the news hits just a day after reports of the company raising $1B at a $22B valuation. Which itself seems set by the price of NVIDIA’s Groq "hackquisition". [Information 🔒]
While it seems weird that OpenAI would launch a stand-alone ChatGPT Translate site — what happened to the "Code Red" and focus?! — I guess it sort of makes sense as a lead-gen tool... [Android Authority]
I’ll reserve more thoughts on Google’s "Personal Intelligence" push with Gemini leverage Gmail, Google Photos, etc, for when I have access. But it’s sort of wild that the pitch sounds exactly like what Apple’s ill-fated pitch was back at WWDC 2024… Presumably Apple will make that push again at WWDC 2026, but now powered by… Google. [Verge]
The new Digg is now in open beta, and after several months in closed beta, is tweaked to even more fully go after Reddit, which is something given Alexis Ohanian’s involvement. [TechCrunch]
If OpenAI, Anthropic, or SpaceX go IPO in 2026 — I’m still skeptical that any of the Big AI folks will go out — will any of them be able to best the Saudi Aramco $1.7T opening valuation hit in 2019? OpenAI has already blown past the largest amount of money ever raised in an IPO in their last private round… [NYT]
Leaks (from a leaker with a good enough track record) suggest the iPhone 18 Pro models will have the same screen sizes as the 17 models (well, aside from the iPhone Fold, which will be new, of course), no surprise, but that Face ID under the screen is a go, and what will indeed shift the front camera to the top-left. What that means for Dynamic Island is a question mark… [MacRumors]
Speaking of that foldable, another rumor puts "liquid metal" as the key to the device. How very Terminator 2. [MacRumors]
Netflix is officially in the podcast business (meaning creating their own, not just partnering/licensing) with Pete Davidson and Michael Irving as the first two partners. We’re a long way from those red envelopes… [Verge]
The Las Vegas facility for UFC fights will now be named the Meta Apex — yes, after that Meta. I mean, it’s hardly new to do deals on the whim of the leader of a company, but announcing them at the same time you’re laying off thousands? Not the best optics, Zuck! [THR]
Disney is pivoting the 'Galaxy’s Edge' area of Disneyland to the original Star Wars trilogy. This is seemingly a sign that while financially successful, the latest trilogy hasn’t really had the staying power you’d expect for the franchise. Even Episodes I, II, and III, as bad as they are, may have more nostalgic value right now! The best news: the Cantina Band will be playing in the Cantina area. [THR]
I Quote…
“What is driving the market right now this extraordinary demand for fast compute.”
— Andrew Feldman, the CEO of Cerebras, talking about the $10B+ deal with OpenAI.
There’s a lot going on with this one. Cerebras filed to go public in 2024 but withdrew the filing a year later seemingly on concerns that nearly all of their revenue was from G42, the Abu Dhabi-based major investor in the company, which also meant a CFIUS review (there may have been other concerns as well). They have seemingly since diversified the business thanks to deals with IBM, Meta, and yes, now this one. Of course, this one has a wrinkle too, as Sam Altman is an investor in Cerebras as well. Because it’s just impossible to do a deal without some sort of conflict in AI.
One more thing: buried at the end of the WSJ report is the following:
In 2017, OpenAI discussed forming a partnership with Cerebras, according to court documents made public in pending litigation between Altman and Elon Musk, who co-founded OpenAI. Feldman said in the interview that he rejected an offer from Musk to buy the ChatGPT-maker around that time.
OpenAI is in the early stages of another massive fundraising to continue to fund its extensive growth plans, the Journal reported. The new investment, which is expected to precede an initial public offering, could value the company at $830 billion, before the new investment.
For the first part, I assume they mean that OpenAI tried to buy Cerebras in 2017 (as yes, the lawsuits have uncovered), not the other way around!
For the second part, they’re saying OpenAI’s funding is being done at $830 pre, which is not what the previous reports (also from WSJ) on the round said. This matters because if OpenAI raised enough — $170B to be exact — they could be the first $1T startup! To be clear, I assume this is a typo.