We still believe that there will be a separation between the “application layer” companies and foundation model providers, with model companies specializing in scale and research and application layer companies specializing in product and UI. In reality, that separation hasn’t cleanly happened yet. In fact, the most successful user-facing applications out of the gate have been vertically integrated.

We predicted that the best generative AI companies could generate a sustainable competitive advantage through a data flywheel: more usage → more data → better model → more usage. While this is still somewhat true, especially in domains with very specialized and hard-to-get data, the “data moats” are on shaky ground: the data that application companies generate does not create an insurmountable moat, and the next generations of foundation models may very well obliterate any data moats that startups generate. Rather, workflows and user networks seem to be creating more durable sources of competitive advantage.

Some of the best consumer companies have 60-65% DAU/MAU; WhatsApp’s is 85%. By contrast, generative AI apps have a median of 14% (with the notable exception of Character and the “AI companionship” category). This means that users are not finding enough value in Generative AI products to use them every day yet.

generative AI’s biggest problem is not finding use cases or demand or distribution, it is proving value. As our colleague David Cahn writes, “the $200B question is: What are you going to use all this infrastructure to do? How is it going to change people’s lives?”

I think hardcore Twitter users have rose-colored glasses about the site’s coolness. The reason for its success, if you can argue that it was ever really successful, wasn’t that it was cooler than Facebook. It was because of its proximity to power. The reason it was so popular with activists, extremists, journalists, and shitposters was because what you posted there could actually affect culture.

The thing that ties together pretty much everything that’s happened on Twitter since it launched in 2006 was the possibility that those who were not in power (or wanted more) could influence those who were.

I subscribe to the belief that internet trends are defined by a ratio of laziness to social reward. Users will always do the laziest possible thing to achieve the maximum amount clout. So, if every platform becomes either a Twitter alternative or a short-form video feed, but all with their own unique requirements for virality, users won’t make individual posts for each. They will instead shotgun blast all of them with the same posts and bet on the odds that something will breakthrough eventually. Which means everything eventually just becomes a reuploaded video or a screenshot from somewhere else.

While trying to track down the actual hyperlink to a post I found a screenshot of on a closed social network I was struck by how on an internet full of closed platforms, broken embeds, and crumbling indexes, the last reliable way to share anything is a screenshot.

the camera roll is, at this point, the real content management system of the social web. This is something that TikTok realized faster than other platforms, with their downloadable watermarked videos that have now become ubiquitous on every platform that allows video.

My theory as to why New Yorkers were so allergic to independent content creators is because for all the tedious guffawing about being a city of hustlers, most of the people who live there crave, on some level, institutional legitimacy and influencers, by definition, don’t get it or really need it. It could also just be that New Yorkers hate tourists and content creators are, in some form, permanent tourists of their own lives.

I actually think the post-COVID New York TikTok boom is already cresting. I think once these trends become calcified enough to report on, they’re already on their way out. I also don’t think Gen Z TikTokers are driving rents up, but rather documenting its rise due to other factors, like landlords being able to blame TikTok hype to jack up their rents.

Meanwhile, whenever I talk to music people or book people, very quickly the conversation becomes a music industry conversation or a book industry conversation. What matters for music are artists and touring and labels and so on, and what matters for books are writers and publishers and rights and Amazon’s bargaining power in books and so on. These aren’t tech conversations.

Tech needed content to make their devices viable, but having got the content (by any means necessary), and with it of course completely resetting the dynamics of the industry, tech outgrew music and books and moved on to bigger opportunities.

All of this of course takes us to TV, the industry that’s next on the tech industry’s content journey. Just as new technology unlocked massive change in music and (rather less so) in books, it is now about to break apart the bundled, linear channel model of the TV industry (this is especially the case in the USA, which has a hugely over-served pay TV market). As this happens, there are all sorts of questions that follow on: what happens to channels that might be able to make more going direct to consumer (HBO, perhaps); what happens to channels that might benefit from being in a bundle and lose from having to go direct (ESPN, perhaps), where the syndication model goes, and so on, and so on.

Just as for music or books, though, these are all fundamentally TV industry questions. What viewing distribution, what rights structure, what exploitation chain, what relationship between creatives, financiers, aggregators and distributors - these are all southern California questions, not northern California questions. So, what are the northern California questions, and will this end up being any more strategic than books or music?

Amazon and Netflix have entered TV content creation and ownership in ways and on a scale that no-one from tech ever did for music or books. Amazon did try to get into book publishing and has a significant self-publishing arm, but it had little success recruiting existing mainstream authors; neither Apple nor Spotify created a record label.

Netflix, of course, is a TV company, in the context of this conversation - it isn’t using content for leverage for some other platform (Spotify is the same, without the commissioning). But Amazon clearly is using content for platform leverage - as something else to speed up the Prime flywheel. Prime has become a third pillar to Amazon’s business, next to logistics and the ecommerce platform, and Amazon is always looking for ways to add more perceived value to it, preferably with no marginal cost - TV content that it owns outright is exactly that.

You don't close your Facebook account - you just go there less. You might stop paying for the Youtube TV service, but that won’t cut off your access to any other part of Google - nor would anyone want it to - the purpose of these businesses is reach. Nor, really, will you fundamentally change your search behaviour if Google discovers the next Game of Thrones. That is, cancel Prime and you'd lose Amazon, but what do Google & FB have to cancel? Without some platform decision to lock you into, content is marketing, and revenue, but not a lever.

Apple has always preferred a very asset-light approach to things that are outside its core skills. It didn’t create a record label, or an MVNO, and it didn’t create a credit card for Apple Pay - it works with partners on the existing rails as much as possible (even the upcoming Apple Pay P2P service uses a partner bank).

Part of ‘content is king’ was the idea that (at least in theory) content companies can withhold access to their libraries entirely, and in the past one might have presumed that that meant they had the power to kill any new service at birth. In reality, rights-holders have always had too strong a need for short-term revenue to forgo broad distribution, and few of them individually had a strong enough brand to extract a fee that was high enough to justify exclusivity. They always have to take the cheques - individually to meet their bonus targets, and collectively to meet their earnings estimates.

This is a multi-sided market place with too many players on both sides for anyone to exert dominance: Apple dominated purchased music and Amazon dominates ebooks (thanks to the DoJ), but there is no such dominance on the buy or sell side for TV, for now.

The reason Apple TV, Chromecast, FireTV and everything else feel so anti-climactic is that getting onto the TV was a red herring - the device is the phone and the network is the internet. The smartphone is the sun and everything else orbits it. Internet advertising will be bigger than TV advertising this year, and Apple’s revenue is larger than the entire global pay TV industry. This is also why tech companies are even thinking about commissioning their own premium shows today - they are now so big that the budgets involved in buying or creating TV look a lot less daunting than they once did.