Desperate Times, Desperate Measures

13 hours ago 3

Next year is going to be big. Well, I personally don't think it'll be big, but if you ask the AI industry, here's the things that will happen by the end of 2026:

How much of this actually sounds plausible to you?

Jony Ive and "The Device"

I thought I couldn't be more disappointed in parts of the tech media, then OpenAI went and bought former Apple Chief Design Officer Jony Ive's "Io," a hardware startup that it initially invested in to create some sort of consumer tech device. As part of the ridiculous $6.5 billion all-stock deal to acquire Io, Jony Ive will take over all design at OpenAI, and also build a device of some sort.

At this point, no real information exists. Analyst Ming-Chi Kuo says it might have a "form factor as compact and elegant as an iPod shuffle," yet when you look at the tweet everybody is citing Kuo's quotes from, most of the "analysis" is guesswork outside of a statement about what the prototype might be like. 

Let's Talk About Ming-Chi Kuo!

It feels like everybody is quoting analyst Ming-Chi Kuo as a source as to what this device might be as a means of justifying writing endless fluff about Jony Ive and Sam Altman's at-this-point-theoretical device.

Kuo is respectable, but only when it comes to the minutiae of Apple — changes in strategy around the components and near-term launches. He has a solid reputation when it comes to finding out what’s selling, what isn’t, and what the company plans to launch. That’s because analysts work by speaking to people — often people working at companies in the less glamorous element of the iPhone and Mac supply chain, like those that manufacture specific components — and asking what orders they’ve received, for what, and when. If a company massively cuts down on production for, say, iPhone screens, you can infer that Apple’s struggling to shift the latest version of the iPhone. Similarly, if a company is having to work around the clock to manufacture an integrated circuit that goes into the newest MacBook, you can assume that sales are pretty brisk. 

Outside of that, Kuo is fucking guessing, and assuming much more than that allows reporters to

make ridiculous and fantastical guesses based on nothing other than vibes. If you are writing that Kuo "revealed details" about the device you have failed your readers, first by putting Kuo on a mythological pedestal (which he already has, to some extent), and secondly by failing to put into context what an analyst does, and what an analyst can’t do. 

And yeah, Kuo is guessing. Jony Ive may have worked at Apple, but he is not Apple. Ive was not a hardware guy — at least when it came to the realm beyond industrial and interface design —, nor did he handle operations at Apple. While Kuo's sources may indeed have some insight, it's highly doubtful he magically got his sources to talk after the announcement versu

Kuo also predicted in 2021 that Apple would release 15-20 million foldable iPhones in 2023, and predicted Apple would launch some sort of AR headset almost every year, claiming it would arrive in 2020, 2022 (with glasses in 2025!), second quarter 2022, "late 2022" (where he also said that Apple would somehow also launch a second-generation version in 2024 with a lighter design), or 2023, but then in mid-2022 decided the headset would be announced in January 2023, and become available "2-4 weeks after the event," and predicted that, in fact, Apple would ship 1.5 million units of said headset in 2023. Sadly, by the end of 2023, Kuo said that the headset would be delayed until the second half of 2023, before nearly getting it right, saying that the device would be announced at WWDC 2023 (correct!), but that it would ship "the second or third quarter of 2023."

Not content with being wrong this many times, Kuo doubled down (or quadrupled down, I’ve lost count)  in February 2023, saying that Apple would launch "

high-end and low-end versions of second-generation headset in 2025," at a point in time when Apple had yet to announce or ship the first generation. Then, finally, literally a day before the announcement of the Vision Pro, Kuo predicted it "could launch as late as 2024," the kind of thing you could've learned from a single source at Apple telling you what would be announced in 24 hours, or, I dunno, the press embargoOn December 25 2023, Kuo successfully predicted that the Vision Pro would launch "in late January or Early February 2024." It launched in the US February 2 2024. Mark Gurman of Bloomberg reported that Apple planned to launch the device "by February 2024" five days earlier on December 20 2023.

Kuo then went on to predict Apple would only produce "up to 80,000 Vision Pro headsets for launch"

on January 11 2024, only to say that Apple had sold "up to 180,000" of them 11 days later. On February 28 2024, after predicting no less than twice that Apple would make multiple models, said that Apple had not started working on a second-generation or lower-priced Vision Pro.

This was a very long-winded way to say that anybody taking tweets by Ming-Chi Kuo as even clues as to what Jony Ive and Sam Altman are making is taking the piss. He has a 72.5% track record for getting things right, according to Apple Track, which is decent, but far from perfect. Any journalist that regurgitates a Ming-Chi Kuo prediction without mentioning that is committing criminal levels of journalistic malpractice. 

So, now that we've got that out the way, here's what we actually know — and that’s a very load-bearing “know” — about this device, according to the Wall Street Journal:

OpenAI Chief Executive Sam Altman gave his staff a preview Wednesday of the devices he is developing to build with the former Apple designer Jony Ive, laying out plans to ship 100 million AI “companions” that he hopes will become a part of everyday life.

...

Altman and Ive offered a few hints at the secret project they have been working on. The product will be capable of being fully aware of a user’s surroundings and life, will be unobtrusive, able to rest in one’s pocket or on one’s desk, and will be a third core device a person would put on a desk after a MacBook Pro and an iPhone.

The Journal earlier reported that the device won’t be a phone, and that Ive and Altman’s intent is to help wean users from screens. Altman said that the device isn’t a pair of glasses, and that Ive had been skeptical about building something to wear on the body.

Let's break down what this actually means:

  • It will be "...capable of being fully aware of a user’s surroundings and life": Multimodal generative AI that can accept both visual and audio inputs are already a feature in basically every major Large Language Model.
  • "It will be unobtrusive, able to rest in one’s pocket or on one’s desk" (and it won't have a screen?): I cannot express how bad it is that this device, which will allegedly ship in a year, is so vague about how big it is. How big are your pockets? Is it smartphone sized? Smaller? If it's able to be "aware" that suggests that it'll have a bunch of sensors and maybe a camera inside it? If that’s the case, wouldn’t putting it in your pocket defeat the point? 
  • "...will be a third core device a person would put on a desk after a MacBook Pro and an iPhone": This means absolutely nothing. It's a statement made to the journalist or from marketing material intentionally shared with the journalist. A "third core device" has never really taken shape — Apple has sold, at best, a hundred million Apple Watches, and sales have begun to tumble. Products like Google’s Glass similarly failed — partially because it was expensive, partially because they became fatally uncool overnight, and partially because the battery life was dismal. The only "third core device" that's stuck is...tablets. And that's a computer! 
    • Also, calling a tablet a “core device” is, at best, a push. According to Canalys — a fairly reliable analyst firm that does the kind of supply-chain investigations I mentioned earlier — fewer than 40 million tablets were shipped worldwide in Q4 last year. That’s talking about shipments, not sales, and it also takes into account demand from educational and business customers, who likely represent a large proportion of global tablet demand. 

The Journal's story also has one of the most ludicrous things I've read in the newspaper: that "...Altman suggested the $6.5 billion acquisition has the potential to add $1 trillion in value to OpenAI," which would mean that OpenAI acquiring a washed former Apple designer who has designed basically nothing since 2019 to create a consumer AI device — a category that has categorically failed to catch on — would somehow nearly quadruple its valuation. Printing that statement is journalistic malpractice without a series of sentences about how silly it is.

But something about Jony Ive gives reporters, analysts and influencers a particular kind of madness. Reporters frame this acquisition as "the big bet that Jony Ive can make AI hardware work," that this is OpenAI "crashing Apple's party," that this is "a wake up call" for Apple, that this is OpenAI "breaking away from the pack" by making "a whole range of devices from the ground up for AI."

Based on this coverage, one might think that Jony Ive has been, I dunno, building something since he left Apple in 2019, which CNBC called "the end of the hardware era at Apple" about six months before Apple launched its M1 series processors and markedly improved its hardware as a result.Hell, much of Apple’s hardware improvement has been because it walked away from Ive’s dubious design choices. Ive’s obsession with thinness led to the creation of the Butterfly Keyboard — a keyboard design that was deeply unpleasant to type on, with very little travel (the distance a key moves when pressed), and a propensity to break at the first glimpse of a speck of dust.

Millions of angry customers — including famed film director Taika Waititi — and a class-action lawsuit later, Apple ditched it and returned to the original design. Similarly, since Ive’s exit, Apple has added HDMI ports, SD card readers, and MagSafe charging back to its laptops. Y’know, the things that people — especially creatives — wanted and liked, but had to be eliminated because they added negligible levels of heft to a laptop. 

What Has Jony Ive Been Up To?

On leaving Apple in 2019 — where he'd been part time since 2015 (though the Wall Street Journal says he returned as a day-to-day executive in 2017, just in time to promise and then never ship the AirPower charging pad) — Ive formed LoveFrom, a design studio with Apple as its first (and primary) client with a contract valued at more than $100 million, according to the New York Times, which reported the collapse of the relationship in 2022:

In recent weeks, with the contract coming up for renewal, the parties agreed not to extend it. Some Apple executives had questioned how much the company was paying Mr. Ive and had grown frustrated after several of its designers left to join Mr. Ive’s firm. And Mr. Ive wanted the freedom to take on clients without needing Apple’s clearance, these people said.

In 2020, LoveFrom signed a non-specific multi-year relationship to “design the future of Airbnb.” LoveFrom also worked on some sort of seal for King Charles to give away during the coronation to — and I quote — “recognize private sector companies that are leading the way in creating sustainable markets.” It also designed an entirely new font for the multi-million dollar event, which does not matter to me in the slightest but led to some reporters writing entire stories about it. The project involves King Charles encouraging space companies. I don’t know, man.

I cannot find a single thing that Jony Ive has done since leaving Apple other than "signing deals." He hasn't designed or released a tech product of any kind. He was a consultant at Apple until 2022, though it's not exactly obvious what it is he did there since the death of Steve Jobs. People lovingly ascribe Apple's every success to Ive, forgetting that (as mentioned) Ive oversaw the truly abominable butterfly keyboard, as well as numerous other wonky designs, including the trashcan-shaped Mac Pro, the PowerMac G4 Cube (a machine aimed at professionals, with a price to match, but limited upgradability thanks to its weird design), and the notorious “hockey puck” mouse.

In fact, since leaving Apple, all I can confirm is that Jony Ive redesigned Airbnb in a non-specific way, made a new font, made a new system for putting on clothing, made a medal for the King of England to give companies that recycle, and made some non-specific contribution to creating an electric car that has yet to be shown to the public.

Are You Kidding Me?

Anyway, this is the guy who's going to be building a product that will ship 100 million units "faster than any company has ever shipped 100 million of something new before."

It took 3.6 years for Apple to sell 100 million iPhones, and nearly six years for them to it 100 million Apple Watches. It took four years for Amazon to sell 100 million Echo devices, Former NFT scam Rabbit claims to have sold over 130,000 units of its "barely reviewable" "AI-powered" R1 device, but told FastCompany last year that the product had barely 5000 daily active users. The Humane Pin was so bad that their returns outpaced their sales, with 10,000 devices shipped but many returned due to, well, it sucking. I cannot find another comparison point, because absolutely nobody has succeded in making the next smartphone or "third device."

To give you another data point, Gartner — another reliable analyst firm, at least when it comes to historical sales trends, although its future-looking predictions about AI and the metaverse can be more ‘miss’ than ‘hit’ — says that the number of worldwide PC shipments (which includes desktops and laptops) hit 64.4 million in Q4 2024. OpenAI thinks that it’ll sell nearly twice as many devices in one year as PCs were sold during the 2024 holiday quarter. That’s insane. And that’s without mentioning things like… uh, I don’t know, who’ll actually build them? Where will you get your parts, Sam? Where will you get your chips? Most semiconductor manufacturers book orders months — if not years — in advance. And I doubt Qualcomm has a spare 100 million chipsets lying around that it’ll let you have for cheap. 

Yet people seem super ready to believe — much like they were with the Rabbit R1 — except they're asking even less of Jony Ive and Sam Altman, the Abbott and Costello of bullshit merchants. It's hard to tell exactly what it is that Ive did at Apple, but what we do know is that Ive designed the Apple Watch, a product that flopped until it refocused on fitness over fashion, and apparently wanted the watch to be a "high-end fashion accessory" rather than the "extension of the iPhone" that Apple executives wanted according to the Wall Street Journal, heavily suggesting that Ive was the reason the Apple Watch flopped far more than the great mind that made Apple a success.

Anyway, this is the guy who's going to build the first true successor to the smartphone, something Jony Ive already failed to do with the full backing of the entire executive team at Apple, a company he worked at for decades, and one that has literally tens of billions of cash sitting in its bank accounts.

Jony Ive hasn't overseen the design or launch of a consumer electronics product in — at my most charitable guess — three years, though I'd be very surprised if his two-or-three-year-long consultancy deal with Apple involved him leading design on any product, otherwise it would have extended it.

If I was feeling especially uncharitable — and I am — I’d guess that Ive’s relationship with Apple ended up looking like that between Alicia Keys and Research in Motion, which in 2013 appointed the singer its “Global Creative Director,” a nebulous job title that gives Prabhakar Raghavan’s “Chief Technologist” a run for its money. Ive acted as a thread of continuity between the Jobs and Cook eras of Apple, while also adding a degree of celebrity to the company that Apple’s other execs — like Phil Schiller and Craig Federighi — otherwise lacked. 

He's teamed up with Sam Altman, a guy who has categorically failed to build any new consumer-facing product outside of the launch of ChatGPT, a product that loses OpenAI billions of dollars a year, to do the only other thing that loses a bunch of money — building hardware.

No, really, hardware is hard. You don't just design something and then send it to a guy in China - you have to go through multiple prototypes, then find one that actually does something using, then work out how to mass-produce it, then actually build the industrial rails to do so, then build the infrastructure to run it, then ship it. At that point, even if the device is really good (it won't be, if it ever launches), you have to sell one hundred million of them, somehow.

I repeat myself - Hardware is hard, to the point where even Apple and Microsoft can cock-up in disastrous (and expensive) ways. Pretty much every 2011 year MacBook Pro — at least, those with their own discrete GPUs — is now e-waste, in part because the combination of shoddy cooling and lead-free solder led these machines to become expensive bricks. The same was true of the Xbox 360. Even if you think the design and manufacturing processes go swimmingly, there’s no guarantee that problems won’t creep up later down the line. 

I beg, plead, scream and yell to the tech media to take one fucking second to consider how ludicrious this is. Io raised $225 million in total funding (and OpenAI already owned 23% of the company from those rounds), a far cry from the billion dollars that The Information was claiming it wanted to raise in April 2024, heavily suggesting that whatever big, secret, sexy product was sitting there wasn't compelling enough to attract anyone other than Sutter Hill Ventures (which famously burned hundreds of millions of dollars investing in Lacework, a company that sold for $200 million and once gave away $30,000 of Lululemon gift cards in one night to anyone that would meet with the company’s sales representatives), Thrive (which has participated in or led multiple OpenAI funding rounds), Emerson Collective (run by Lauren Powell Jobs, a close friend of Jony Ive and Altman according to The Information) and, of course, OpenAI itself, which bought the company in its own stock after already owning 23% of its shares.

This deal reeks of desperation, and is, at best, a way for venture capitalists that feel bad about investing in Jony Ive's lack of productivity to get stock in OpenAI, a company that also doesn't build much product.

While OpenAI has succeeded in making multiple different models, what actual products have come out of GPT, Gemini or other Large Language Models? We're three joyless years into this crap, and there isn't a single consumer product of note other than ChatGPT, a product that gained its momentum through a hype campaign driven by press and markets that barely understood what they were hyping.

Despite all that media and investor attention — despite effectively the entirety of the tech industry focusing on this one specific thing — we're still yet to get any real consumer product. Somehow Sam Altman and Jony Ive are going to succeed where Google, Amazon, Meta, Apple, Samsung, LG, Huawei, Xiaomi, and every single other consumer electronics companies has failed, and they're going to do so in less than a year, and said device is going to sell 100 million units.

OpenAI didn't acquire Jony Ive's company to build anything — it did so that it could increase the valuation of OpenAI in the hopes that it can raise larger rounds of funding. It’s the equivalent of adding an extension to a decrepit, rotting house. 

OpenAI, as a company, is lost. It has no moat, its models are hitting the point of diminishing returns and have been for some time, and as popular as ChatGPT may be, it isn't a business and constantly loses money.

On top of that, it requires more money than has ever been invested in a startup. SoftBank had to take out a $15 billion bridge loan from 21 different banks just to fund the first $7.5 billion of the $30 billion it’s promised OpenAI in its last funding round.

At this point, it isn't obvious how SoftBank affords the next part of that funding, and OpenAI using stock rather than cash to buy Jony Ive's company suggests that it doesn’t have much to spare. OpenAI is allegedly also buying AI coding company Windsurf for $3 billion. The deal was announced on May 6 2025 by Bloomberg, but it's not clear if it closed, or whether the deal would be in cash or stock, or really anything, and I have to ask: how much money does OpenAI really have?

And how much can it afford to burn? OpenAI’s operating costs are insane, and the company has already committed to several grand projects, while also pushing deeper and deeper into the red. And if — when? — its funding rounds convert into loans, because it failed to convert into a for-profit, OpenAI will have even less money to splash on nebulous vanity projects. Then again, asking questions like that isn't really how the media is doing business with OpenAI — or, for that matter, has done with the likes of Mark Zuckerberg, Satya Nadella, or Sundar Pichai. Everything has to be blindly accepted, written down and published, for fear of...what, exactly? Not getting the embargo to a product launch everybody else got? Missing out on the chance to blindly cover the next big thing, even if it won't be big, and might not even be a thing?

Into The Bullshitverse

So, I kicked off this newsletter with a bunch of links tied to the year 2026, and I did so because I want — no, need — you to understand how silly all of this is.

Sam Altman's OpenAI is going to, in the next year, according to reports:

  • Design, prototype, manufacture and ship the next big consumer tech device, shipping 100 million units — or, as mentioned, nearly twice the number of PCs shipped in Q4 2024 — faster than any other company in history.
  • Bring both the barely-started Stargate Texas and the still-theoretical Stargate UAE data centers online.
    • As a note, SoftBank, who will have full financial responsibility for the project, is having trouble raising the supposed $100 billion to build it.
    • This project is also dependent on Crusoe, a company that has never built an AI data center, and it is being, to quote CEO Chase Lochmiller, forced to "...deliver on the fastest schedule that a 100-megawatt-or-greater data center has ever been built."
    • In fact, the entire Texas project is contingent on debt. Bloomberg reports that both OpenAI and SoftBank will put $19bn each "to start" (with what money?) and Abu Dhabi-based investment firm MGX and Oracle are putting in $7 billion each. Oracle has also signed a 15-year-long lease with Crusoe, and Stargate has one customer — OpenAI.
  • Launch a non-specific AI-specific chip with Broadcom.
    • I cannot express how unlikely it is that this happens. Silicon is even harder than hardware!

Even one of these projects would be considered a stretch. A few weeks ago, Bloomberg Businessweek put out a story called "Inside the First Stargate AI Data Center."  Just to be clear, this will be "fully operational" (or "constructed" depending on who you ask!) by the middle of 2026. The real title should've been "Outside the First Stargate AI Data Center," in part because Bloomberg didn't seem to be allowed into anything, and in part because it doesn't seem like there's an inside to visit.

Again, if I’m being uncharitable — which I am — this whole thing reminds me of that model town that North Korea built alongside the demilitarized zone to convince South Koreans about the beauty of the Juche system and the wisdom of the Dear Leader — except the beautiful, ornate houses are, in fact, empty shells. A modern-day Potemkin village. Bloomberg got to visit a Potemkin data center. 

Data centers do not just pop out of the ground like weeds. They require masses of permits, endless construction, physical service architecture, massive amounts of power, and even if you somehow get all of that together you still have to make everything inside it work. While analysts believe that NVIDIA has overcome the overheating issues with its Blackwell chips, Crusoe is brand fucking spanking new at this, and The Information described Stargate as "new terrain for Oracle...relying on scrappy but unproven startups...[and] more broadly, [Oracle] has less experience than its larger rivals in dealing with utilities to secure power and working with powerful and demanding customers whose plans change frequently."

In simpler terms, you have a company (Oracle) building something at a scale it’s never built at before, using a partner (Crusoe) which has never done this, for a company (OpenAI) that regularly underestimates the demands it puts on its servers. The project being built is also the largest of its kind, and is being built during the reign of an administration that births and kills a new tariff seemingly every day.

Anyway, all of this needs to happen while OpenAI also funds its consumer electronic product, as well as their main operations which will lose them $14 billion in 2026, according to The Information.

It also needs to become a non-profit by the end of 2025 or lose $10 billion of SoftBank's funding, a plan that SoftBank accepted but Microsoft is yet to approve, in part (according to the Information) because OpenAI wants to both give it a smaller cut of profits and stop Microsoft from accessing its technology past 2030.

This is an insane negotiation strategy — leaking to the press that you want to short-change your biggest investor both literally and figuratively — and however it resolves will be a big tell as to how stupid the C-suite at Microsoft really is. Microsoft shouldn't budge a fucking inch. OpenAI is a loser of a company run by a career liar that cannot ship product, only further iterations of an increasingly-commoditized series of Large Language Models.

At this point, things are so ridiculous that I feel like I'm huffing paint fumes every time I read Techmeme.

If you're a member of the media reading this, I implore you to look more critically on what's going on, to learn about the industries in question and begin asking yourselves why you continually and blandly write up whatever it is they say. If you think you're "not a financial journalist" or "not a data center journalist" and thus "can't understand this stuff," you're wrong. It isn't that complex, otherwise a part-time blogger and podcaster wouldn't be able to pry it apart.

That being said, there's no excuse for how everybody covered this Jony Ive fiasco. Even if you think this device ships, it took very little time and energy to establish how little Jony Ive has done since leaving Apple, and only a little more time to work out exactly how ridiculous everything about it is. I know you need stories about stuff — I know you have to cover an announcement like this — but god, would it fucking hurt to write something even a little critical? Is it too much to ask that you sit down and find out what Jony Ive actually does and then think about what that might mean for the future?

This story is ridiculous. The facts, the figures, the people involved, everything is stupid, and every time you write a story without acknowledging how unstable and untenable it is, you further misinform your readers. Even if I’m wrong — even if they somehow pull off all of this stuff — you still left out a valuable part of the story, refused to critique the powerful, and ultimately decided that marketing material and ephemera were more valuable than honest analysis. 

There is no reason to fill in the gaps or “give the benefit of the doubt” to billionaires, and every single time you do, you fail your audience. If that hurts to read, perhaps ask yourself why. 

Holding these people accountable isn’t just about asking tough questions, but questioning their narratives and actions and plans, and being willing to write that something is ridiculous, fantastical, or outlandish. Doing so — even if you end up being proven wrong — is how you actually write history, rather than simply existing as a vessel for Sam Altman or Jony Ive or Dario Amodei or any number of the world’s Sloppenheimers. 

Look, I am nobody special. I am not supernaturally intelligent, nor am I connected to vast swaths of data or suppliers that allow me to write this. I am a guy with a search engine who remembers when people said stuff, and the only thing you lack is my ability to write 5000 or more words in the space of five hours. If you need help, I am here to help you. If you need encouragement, I am here to provide it. If you need critiques, well, scroll up. Either way, I want to see a better tech media, because that’s what the world deserves.

You can do better.

Read Entire Article
  1. Homepage
  2. Technology
  3. Desperate Times, Desperate Measures

Related

Awesome Emacs on macOS

Awesome Emacs on macOS

24 minutes ago 0
'laughing gas' became a deadly – but legal – American addiction

'laughing gas' became a deadly – but legal – American addict...

28 minutes ago 0
Why Use a Monorepo Build Tool?

Why Use a Monorepo Build Tool?

29 minutes ago 0
Hostinger Web Hosting