Recently, Paramount said they’ll be testing micro-dramas, the short serialized soap operas that have become so popular in China that some describe them as evidence of the nation’s soft power. In the back of the classroom, hiding under the chair, are the less prestigious serialized vertical videos that have found success, like Fruit Love Island, an AI-generated reality TV parody about a bunch of fruit looking for love.

I’m reminded of the drama around Quibi, the short-lived 2020 subscription streaming service that was supposed to bring short-form content to the masses. Founded by Hollywood legend Jeffrey Katzenberg (the K in “Dreamworks SKG”), it had a few obvious missteps (at least in retrospect). An emphasis on high Hollywood talent wasn’t really necessary. Some tech bells and whistles, like vertical and horizontal playability, were a waste of money. And the subscription fee itself created a paywall for a product that people didn’t know they wanted.

Yet the biggest bet — that people would be gobbling up dramatic short form content in every grocery line, at every stop light, and between all the other waking moments of their day — was miraculously right. Katzenberg was widely mocked as being the oracle that wasn’t, but now he just looks like he was early and poorly priced.

Of course, this is all a measure of what being “right” about a business really means, and how much you think timing has to be part of the equation. Though launching during a worldwide pandemic was actually fortuitous for reaching content-starved quarantined customers, Quibi was just too early on to be a hit. Too early for premium on phones, too early for premium short form, and perhaps too early for sub-premium attention spans. A right idea at the wrong time isn’t really a right idea.

So what’s the broader take on this for the business of tech? I wonder if the leading AI labs are producing a premium product that people will later get at a much reduced cost. Thus far, this newsletter must have seemed like a blessed respite from AI discourse, but I can’t help but think of the massive capital expenditures of OpenAI, which has sunk a lot of money into a partly proven business model. Established players like Google are doing the same, but it seems fair to say they’re lurking in the background and trying to make a profit.

That’s to say nothing of the open-weight models (AI you can download and run on your own) that are increasingly competitive with leading AI companies. They represent a potential disruption of AI in a highly uncertain marketplace, where we have yet to learn what customers really want, how much they’re willing to pay for it, and how much costs will balloon or deflate in the future.

In hindsight, Quibi was an incredible idea, but at the same time, it was right to mock it. It failed. Wondering if these labs are the next Quibi doesn’t express skepticism about AI, just about its most likely path toward profitability. Next time I read about a hundred billion AI budget, I’ll be wondering: is this the right bet? Or should I just wait for Fruit Love Island to come along?

The Codex Micro is a “drop” akin to a YouTuber releasing a new t-shirt. It’s a “collab” with keyboard maker Work Louder that allows for some customization and integration between a branded (and IMHO under-thought) hardware device and OpenAI’s Codex. It’s more than a stunt (the ChatGPT basketball is a stunt). But it’s also less than a product, especially with that $230 price tag.

It makes more sense as a hyper-targeted play for the type of power-users — or power employees — influencing the perception and discussion of AI in the larger world. This extremely niche form of marketing is kind of the norm when it comes to the frontier AI labs. In the absence of big marketing campaigns, these attempts at subculture virality have become the main marketing message. AI sucks up so much oxygen that messages intended to appeal to a small sliver of the public end up being mass marketing.

This is most obvious in the “There’s hope in hard questions” ad that Anthropic put out. If you haven’t seen it, it includes a burning house and a shot of Arlington National Cemetery. Bizarrely, it aired during the World Cup. There’s an argument that Anthropic is following an Exxon strategy to get in front of the disaster (we’re the only people who can stop ourselves!) but I think it’s as likely that messaging conceived for a small audience — AI engineers worried about AI — inadvertently went from narrowcast to broadcast without a lot of thought.

And that’s the basis for the argument that we haven’t really seen a mass-AI ad campaign yet. It seems almost ironic considering that every company blathers about AI whenever the opportunity arises. But a coherent frontier lab campaign hasn’t happened, likely due to the prioritization of spending that money on computation power rather than brandshare. The messaging that does come out is intended for an audience of regulators, influencers, and potential employees.

Google and Apple might be the only adults in the room, because they already have big businesses that target those outside the Silicon Valley bubble. Frontier AI products have increasingly baroque and mythologically complex product naming schema (apparently GPT is cosmological now), but Google has helpfully simplified Notebook LM to Gemini Notebook. Apple’s Siri AI does use the dreaded “AI” branding (though it putatively stands for Apple Intelligence), but the name is irrefutably simple.

All this means we are still in the early innings of marketing AI to ordinary people. But companies might do well to exhibit some caution in how they sell to early adopters, because everyone is watching.

This has the risk of sounding like one of those AI-written koans that show up on LinkedIn (“It’s not a moat — it’s a sand castle”). But I think the analogy aptly describes some of the weird stuff going on. Media writer John Hermann has a good piece in this week’s New York Magazine about how AI companies seem to be unable to forecast their own future, and I’d argue that extends to their own business defensibility.

Moats can have a few grooves (this SEO-optimized post by Charles River Ventures digs them all out). Let’s get sand between our toes and see what moats AI really has:

1. Network Effects: The 2010s were defined by these moats, in which friends on a service beget more friends on a service.

   These strike me as being really low for AI. There are attempts at getting network effects — the first dot biz video was about forward deployed engineers, which sends a hive of pseudo-consultants into a company to try to create some lock in. OpenAI attempted group chats at one point, but the effort seemed to fizzle. Their video social network, Sora, was shut down. For AI, the Network Effects piece hasn’t fallen in place.

2. Switching Costs: How hard is it to switch your AI?

   This strikes me as the ultimate AI paradox. The goal is to build a series of increasingly intelligent computers that can understand anything. The more they can understand, the easier it becomes to transfer data between domains. To add to this, much of the knowledge that an AI accumulates for individual power users is stored in laughably simple personal “markdown” files that describe, in plain English, how to replicate an AI’s function. Another AI can just…read those notes.

   This is especially acute when IDEs — the “integrated development environments” programmers use — can easily switch not only between AI models, but also between the companies providing those models.

   It seems unfair to claim there’s no lock-in at all — I know that as a personal user of a $100 plan, there’s a certain amount of laziness that keeps me from switching, especially when I presume that the next model drop will catch me up with the competition. But if that upgrade schedule weren’t predictable within the week? I’d jump ship immediately.

   Google has tried to increase switching costs by bundling their plans with storage, YouTube Premium, and a full suite of AI services. That heft may work in the long term, but it’s currently unclear.

3. Intangible Assets: This strikes me as the strongest moat for any company which would, in turn, worry me. It’s the squishiest. Looking at branding, each company has some intangible assets: OpenAI has the best consumer awareness and is synonymous with “AI” for many; Anthropic has safety vibes and gorgeous brand direction; Google is Google; and XAI has the edgelords on lock.

   These intangible assets can also extend to relationships with the government. Sam Altman and OpenAI have close ties to the Trump Administration, while Anthropic has a feud. Jensen Huang of NVidia and Trump are close as well. The list goes on. This type of crony-capitalist relationship is an untraditional measure of intangible assets, but that’s 2026 for you.

4. Cost Advantage: Daily updates about inference optimizations — what actually helps the AI answer questions — and chip crunches are too frequent to catalog here. To me, that says that no company has created an enduring cost advantage. Like the AI models themselves, companies can easily catch up to competitors.

   This cost advantage includes the constant business threat of open source models, which can be run locally or in the cloud for a fraction of the cost of models from the big players. Without judging the “hacking threat” new open source models may pose, open source is constantly nipping at the heels of the big companies, especially as customers become more budget-conscious.

5. Efficient Scale: In his interview with Dwarkesh Patel, Anthropic CEO Dario Amodei mentioned Cournot competition, which basically means markets can be stable if they’re super pricey to get into. Big training runs may have those characteristics, because they can cost hundreds of millions of dollars of computing power. But open source may show that these big expenditures aren’t necessary.
   It’s another possible case of an AI paradox: the more the creation of models is optimized, the easier it gets for competitors to make new ones.

This analysis is highly subjective, and the billions of dollars flowing into AI indicate that it’s not shared by a lot of big money players. But a moat is supposed to be hard to cross, and these moats seem pretty shallow, narrow, or dry. A sandcastle? It can be beautiful, and it can be amazing for a day. But you can also copy it a few feet over on the sand. And it can be washed away.

You don’t need to know what OFC is, or peak attendance, to guess that the blue line’s crest is the puffing of a truly epic bubble before an equally epic pop. And as your eyes descend to the timeline, it becomes clear that the bubble is the dot com bubble.

Except, it isn’t.

Well, not exactly.

“OFC peak attendance” is the label of this chart, and the chart is paired with a chart of “OFC exhibit space,” which follows a nearly identical arc.

OFC, in both cases, represents the Optical Fiber Communication conference, and it comes from their retrospective on a narrow slice of the dot com bubble: the fiber bubble. These are the people who covered the United States with fiber optic cable in the late 90s.

Discussions of the current AI bubble mistily recall the dot com bubble of the late ‘90s and early 2000s, and AI-skeptics are likely to invoke famous dot com failures like Webvan, an early online grocery delivery company, or Pets.com, the other iconic example of early internet enthusiasms that failed to survive the dot com crash.

That analogy isn’t precise. There is an AI bubble, but it’s probably better represented by a large crop of AI wrappers: companies that plug into leading AI companies, like Anthropic or OpenAI, and sell their own integration of those leaders’ software. Imagine an “AI workout app” that has some nice branding, but basically sends preprogrammed instructions off to OpenAI for one of their GPT models to do the real work. The recent crop of small startups is littered with companies like this, and these provide a neat analog to the 2000s web bubble, since both groups include consumer-facing enterprises with unproven models.

But just because some consumer AI startups might be bubbly in their lack of revenue, that doesn’t mean AI as a category is a bubble. Anthropic has a $47 billion run rate revenue, and while that doesn’t indicate anything about their profitability, it does indicate massive real interest in their product. This is a big deal! The specific businesses in AI are deeply uncertain, but people want AI in the same way they really wanted fiber optic cable.

This dynamic is clear if you look at the stock price of Corning and Nvidia. The famous glassmaker makes optical fiber, and their stock price tracks the waxing and waning of fiber enthusiasm even better than attendance charts for a fiber-optic convention. They also benefit from a boom in AI, because their fiber supplies new data centers. NVidia, meanwhile, is mostly a modern phenomenon, since it was traded as a gaming company before its GPUs were used in the crypto boom and then AI. See how Corning has a pop in 1999, before roughly tracking NVidia’s rise (in price, not market cap)?

The fiber optic bubble was a real bubble. Companies over-invested in fiber and bankruptcies followed. But the interest and enduring value was real value that we still benefit from today. Infrastructure buildout done in the late 90s has meaningful implications today, but there were still big losers when the bubble burst.

The AI bubble is not a bubble like the consumer companies of the initial dot com frenzy — it’s a massive infrastructure boom, like fiber, with lots of possible bubbles that could pop. Like the fiber boom, or even the railroad boom a century before that, this might be frustrating because it makes it harder to invest around. But it does have the takeaway that AI shouldn’t be ignored, because the demand is real.

We’re in the middle of an ongoing takeover—tech is eating business, and it tastes like MANGOS.

This is the juicy new way to refer to Meta, Anthropic, Nvidia, Google, OpenAI, and SpaceX (we say Google rather than its parent company, Alphabet, for the fruitiness). These are the AI-centered companies that might define the next decade in tech, and they’re climbing close to a combined $15 trillion market cap.

Imagine you’re a doctor in Cleveland.

You’re smart, you’re curious, and you’re interested in what’s happening to your investments. But you’re also not listening to 4-hour podcasts with luminaries in artificial intelligence, and you’re not subscribing to paywalled newsletters about software-as-a-service. You also aren’t part of a San Francisco and Silicon Valley bubble, where chatter about tech is constantly in the air. You might not hear about tech news in casual conversation at all.

Where are the publications serving you? Business news is swiftly becoming just tech news, and yet it’s surprisingly hard to find explanations. This is the lens for dot biz, and it’s a necessary one.

This is a retread of the About page, but there’s a big problem here. Publications are highly polarized: they either hate technology companies and their social consequences, or they’re extreme boosters funded and/or owned by the very companies they cover. Both often come with insufficient disclosures (we’ll always have a disclosures page you can look at). If you’re a doctor in Cleveland, you’re able to formulate your own opinions about how these companies affect your work, life, and communities for good or for ill—you need a publication that provides basic explanation of the business decisions, not a sermon.

The other major problem? The information is polarized in a couple ways. It’s either super niche and technical focused, usually behind a paywall. Or it’s overly broad and hypey, playing a volume game that means there’s not a lot of editorial oversight. It’s often incredibly granular (which artificial intelligence model launched on which particular day, without any context to understand it) or incredibly broad (how humanity might be saved and/or doomed by some new breakthrough). Dot biz will sit in the middle, providing you carefully curated news and, crucially, the context to understand it.

Finally, our focus is going to be on the business. We’ll always let you know the business context behind decisions. Sometimes, that will necessarily get into technical details or history, but the focus is always the business dynamics at play. Who’s winning? Who’s losing? What might happen next? Many other angles on tech are fascinating, from the gadgets to the AI, but we’ll stick to the implications for business.

Right now, dot biz is a laptop sticker on Phil Edwards, a journalist who’s covered just about everything over the past decade. I’ll shift from the completely fictitious “we” for the rest of this paragraph. I’ll be finding my footing when it comes to publication cadence, formats, and more. But that doctor in Cleveland is going to remain my reader through all the short videos, long-form videos, newsletters, and any other ways I try to do the work and explain the businesses that haven’t been explained yet. I’ll also be writing this myself. I’ll use lots of AI in the production process, but the words and insights will be human-generated (and damn it I’ll use em dashes if I want to).

Of course, this publication is for all people who want to understand tech. But we’re thinking of that doctor in Cleveland with every story. Your feet are tired from standing. You can’t plug in the podcast. But you’re smart, you’re curious, and you deserve news about the mangos taking over the world.