Found 56 bookmarks
Newest
The narratives we build, build us — sindhu.live
The narratives we build, build us — sindhu.live
You see glimpses of it in how Epic Games evolved from game engines to virtual worlds to digital marketplaces, or how Stripe started as a payments processing platform but expanded into publishing books on technological progress, funding atmospheric carbon removal, and running an AI research lab.
Think about what an operating system is: the fundamental architecture that determines what's possible within a system. It manages resources, enables or constrains actions, and creates the environment in which everything else runs.
The dominant view looks at narrative as fundamentally extractive: something to be mined for short-term gain rather than built upon. Companies create compelling stories to sell something, manipulate perception for quick wins, package experiences into consumable soundbites. Oil companies, for example, like to run campaigns about being "energy companies" committed to sustainability, while their main game is still extracting fossil fuels. Vision and mission statements claim to be the DNA of a business, when in reality they're just bumper stickers.
When a narrative truly functions as an operating system, it creates the parameters of understanding, determines what questions can be asked, and what solutions are possible. Xerox PARC's focus on the architecture of information wasn't a fancy summary of their work. It was a narrative that shaped their entire approach to imagining and building things that didn't exist yet. The "how" became downstream of that deeper understanding. So if your narrative isn't generating new realities, you don't have a narrative. You have a tagline.
Most companies think they have an execution problem when, really, they have a meaning problem.
They optimise processes, streamline workflows, and measure outcomes, all while avoiding the harder work of truly understanding what unique value they're creating in the world. Execution becomes a convenient distraction from the more challenging philosophical work of asking what their business means.
A narrative operating system fundamentally shifts this dynamic from what a business does to how it thinks. The business itself becomes almost a vehicle or a social technology for manifesting that narrative, rather than the narrative being a thin veneer over a profit-making mechanism. The conversation shifts, excitingly, from “What does this business do?" to "What can this business mean?" The narrative becomes a reality-construction mechanism: not prescriptive, but generative.
When Stripe first articulated their mission to "increase the GDP of the internet" and “think at planetary scale”, it became a lens to see beyond just economic output. It revealed broader, more exciting questions about what makes the internet more generative: not just financially, but intellectually and culturally. Through this frame emerged problems worth solving that stretched far beyond payments:  What actually prevents more people from contributing to the internet's growth? Why has our civilisation's progress slowed? What creates the conditions for ambitious building? These questions led them down unexpected paths that seem obvious in retrospect. Stripe Atlas enables more participants in the internet economy by removing the complexity of incorporating a company anywhere in the world. Stripe Climate makes climate action as easy as processing a payment by embedding carbon removal into the financial infrastructure itself. Their research arm investigates why human progress has slowed, from the declining productivity of science to the bureaucratisation of building. And finally, Stripe Press—my favourite example—publishes new and evergreen ideas about technological progress.
The very metrics meant to help the organisation coordinate end up drawing boundaries around what it can imagine [1]. The problem here again, is that we’re looking at narratives as proclamations rather than living practices.
I don’t mean painted slogans on walls and meeting rooms—I mean in how teams are structured, how decisions get made, what gets celebrated, what questions are encouraged, and even in what feels possible to imagine.
The question to ask isn't always "What story are we telling?" but also "What reality are we generating?”
Patagonia is a great example of this. Their narrative is, quite simply: “We’re in business to save our home planet”. It shows up in their unconventional decision to use regenerative agriculture for their cotton, yes, but also in their famous "Don't Buy This Jacket" Black Friday campaign, and in their policy to bail out employees arrested for peaceful socio-environmental protests. When they eventually restructured their entire ownership model to "make Earth our only shareholder," it felt less like a radical move and more like the natural next step in their narrative's evolution. The most powerful proof of their narrative operating system was that these decisions felt obvious to insiders long before it made sense to the outside world.
Most narrative operating systems face their toughest test when they encounter market realities and competing incentives. There are players in the system—investors, board members, shareholders—who become active narrative controllers but often have fundamentally different ideas about what the company should be. The pressure to deliver quarterly results, to show predictable growth, to fit into recognisable business models: all of these forces push against maintaining a truly generative narrative.
The magic of "what could be" gets sacrificed for the certainty of "what already works." Initiatives that don't show immediate commercial potential get killed. Questions about meaning and possibility get replaced by questions about efficiency and optimisation.
a narrative operating system's true worth shows up in stranger, more interesting places than a balance sheet.
adaptability and interpretive range. How many different domains can the narrative be applied to? Can it generate unexpected connections? Does it create new questions more than provide answers? What kind of novel use cases or applications outside original context can it generate, while maintaining a clear through-line? Does it have what I call a ‘narrative surplus’: ideas and initiatives that might not fit current market conditions but expand the organisation's possibility space?
rate of internal idea generation. How many ideas come out of the lab? And how many of them don’t have immediate (or direct) commercial viability? A truly generative narrative creates a constant bubbling up of possibilities, not all of which will make sense in the current market or at all.
evolutionary resilience, or how well the narrative can incorporate new developments and contexts while maintaining its core integrity. Generative narratives should be able to evolve without fracturing at the core.
cross-pollination potential. How effectively does the narrative enable different groups to coordinate and build upon each other's work? The open source software movement shows this beautifully: its narrative about collaborative creation enables distributed innovation and actively generates new forms of cooperation we couldn't have imagined before.
There are, of course, other failure modes of narrative operating systems. What happens when narratives become dogmatic and self-referential? When they turn into mechanisms of exclusion rather than generation? When they become so focused on their own internal logic that they lose touch with the realities they're trying to change? Those are meaty questions that deserve their own essay.
·sindhu.live·
The narratives we build, build us — sindhu.live
American Disruption
American Disruption
manufacturing in Asia is fundamentally different than the manufacturing we remember in the United States decades ago: instead of firms with product-specific factories, China has flexible factories that accommodate all kinds of orders, delivering on that vector of speed, convenience, and customization that Christensen talked about.
Every decrease in node size comes at increasingly astronomical costs; the best way to afford those costs is to have one entity making chips for everyone, and that has turned out to be TSMC. Indeed, one way to understand Intel’s struggles is that it was actually one of the last massive integrated manufacturers: Intel made chips almost entirely for itself. However, once the company missed mobile, it had no choice but to switch to a foundry model; the company is trying now, but really should have started fifteen years ago. Now the company is stuck, and I think they will need government help.
companies that go up-market find it impossible to go back down, and I think this too applies to countries. Start with the theory: Christensen had a chapter in The Innovator’s Dilemma entitled “What Goes Up, Can’t Go Down”: Three factors — the promise of upmarket margins, the simultaneous upmarket movement of many of a company’s customers, and the difficulty of cutting costs to move downmarket profitably — together create powerful barriers to downward mobility. In the internal debates about resource allocation for new product development, therefore, proposals to pursue disruptive technologies generally lose out to proposals to move upmarket. In fact, cultivating a systematic approach to weeding out new product development initiatives that would likely lower profits is one of the most important achievements of any well-managed company.
So could Apple pay more to get U.S. workers? I suppose — leaving aside the questions of skills and whatnot — but there is also the question of desirability; the iPhone assembly work that is not automated is highly drudgerous, sitting in a factory for hours a day delicately assembling the same components over and over again. It’s a good job if the alternative is working in the fields or in a much more dangerous and uncomfortable factory, but it’s much worse than basically any sort of job that is available in the U.S. market.
First, blanket tariffs are a mistake. I understand the motivation: a big reason why Chinese imports to the U.S. have actually shrunk over the last few years is because a lot of final assembly moved to countries like Vietnam, Thailand, Mexico, etc. Blanket tariffs stop this from happening, at least in theory. The problem, however, is that those final assembly jobs are the least desirable jobs in the value chain, at least for the American worker; assuming the Trump administration doesn’t want to import millions of workers — that seems rather counter to the foundation of his candidacy! — the United States needs to find alternative trustworthy countries for final assembly. This can be accomplished through selective tariffs (which is exactly what happened in the first Trump administration).
Secondly, using trade flows to measure the health of the economic relationship with these countries — any country, really, but particularly final assembly countries — is legitimately stupid. Go back to the iPhone: the value-add of final assembly is in the single digit dollar range; the value-add of Apple’s software, marketing, distribution, etc. is in the hundreds of dollars. Simply looking at trade flows — where an imported iPhone is calculated as a trade deficit of several hundred dollars — completely obscures this reality. Moreover, the criteria for a final assembly country is that they have low wages, which by definition can’t pay for an equivalent amount of U.S. goods to said iPhone.
At the same time, the overall value of final assembly does exceed its economic value, for the reasons noted above: final assembly is gravity for higher value components, and it’s those components that are the biggest national security problem. This is where component tariffs might be a useful tool: the U.S. could use a scalpel instead of a sledgehammer to incentivize buying components from trusted allies, or from the U.S. itself, or to build new capacity in trusted locations. This does, admittedly, start to sound a lot like central planning, but that is why the gravity argument is an important one: simply moving final assembly somewhere other than China is a win — but not if there are blanket tariffs, at which point you might as well leave the supply chain where it is.
You can certainly make the case that things like castings and other machine components are of sufficient importance to the U.S. that they ought to be manufactured here, but you have to ramp up to that. What is much more problematic is that raw materials and components are now much cheaper for Haas’ foreign competitors; even if those competitors face tariffs in the United States, their cost of goods sold will be meaningfully lower than Haas, completely defeating the goal of encouraging the purchase of U.S. machine tools.
Fourth, there remains the problem of chips. Trump just declared economic war on China, which definitionally increases the possibility of kinetic war. A kinetic war, however, will mean the destruction of TSMC, leaving the U.S. bereft of chips at the very moment that A.I. is poised to create tremendous opportunities for growth and automation. And, even if A.I. didn’t exist, it’s enough to note that modern life would grind to a halt without chips. That’s why this is the area that most needs direct intervention from the federal government, particularly in terms of incentivizing demand for both leading and trailing edge U.S. chips.
my prevailing emotion over the past week — one I didn’t fully come to grips with until interrogating why Monday’s Article failed to live up to my standards — is sadness over the end of an era in technology, and frustration-bordering-on-disillusionment over the demise of what I thought was a uniquely American spirit.
Internet 1.0 was about technology. This was the early web, when technology was made for technology’s sake. This was when we got standards like TCP/IP, DNS, HTTP, etc. This was obviously the best era, but one that was impossible to maintain once there was big money to be made on the Internet. Internet 2.0 was about economics. This was the era of Aggregators — the era of Stratechery, in other words — when the Internet developed, for better or worse, in ways that made maximum economic sense. This was a massive boon for the U.S., which sits astride the world of technology; unfortunately none of the value that comes from that position is counted in the trade statistics, so the administration doesn’t seem to care. Internet 3.0 is about politics. This is the era when countries make economically sub-optimal choices for reasons that can’t be measured in dollars and cents. In that Article I thought that Big Tech exercising its power against the President might be a spur for other countries to seek to wean themselves away from American companies; instead it is the U.S. that may be leaving other countries little choice but to retaliate against U.S. tech.
There is, admittedly, a hint of that old school American can-do attitude embedded in these tariffs: the Trump administration seems to believe the U.S. can overcome all of the naysayers and skeptics through sheer force of will. That force of will, however, would be much better spent pursuing a vision of a new world order in 2050, not trying to return to 1950. That is possible to do, by the way, but only if you accept 1950’s living standards, which weren’t nearly as attractive as nostalgia-colored glasses paint them, and if we’re not careful, 1950’s technology as well. I think we can do better that that; I know we can do better than this.
·stratechery.com·
American Disruption
What kind of disruption? — Benedict Evans
What kind of disruption? — Benedict Evans
Where previous generations of tech companies sold software to hotels and taxi companies, Airbnb and Uber used software to create new businesses and to redefine markets. Uber changed what we mean when we say ‘taxi’ and Airbnb changed hotels.
But for all sorts of reasons, the actual effect of that on the taxi and hotel industries was very different. The regulation is different. The supply of people with a car and few hours to spare is very different from the supply of people with a spare room to rent out (indeed, there is adverse selection in that difference). The delta between waving your hand on a street corner and pressing a button on your phone is different to the delta between booking a hotel room and booking a stranger’s apartment.
Sometimes disruption is much more about new demand than challenging the existing market, or only affects a peripheral business, as happened with Skype.
it’s always easier to shout ‘disruption!’ or ‘AI!’ than to ask what kind.
·ben-evans.com·
What kind of disruption? — Benedict Evans
Apple innovation and execution — Benedict Evans
Apple innovation and execution — Benedict Evans
since the iPhone launched Apple has created three (!) more innovative and category-defining products - the iPad, Watch and AirPods. The iPad is a little polarising amongst tech people (and it remains unfinished business, as Apple concedes in how often it fiddles with the keyboard and the multitasking) but after a rocky start it’s stabilised as roughly the same size as the Mac. The Watch and the Airpods, again, have both become $10bn+ businesses, but also seem to have stabilised. (The ‘Wearables, Home and Accessories’ category also includes the Apple TV, HomePods and Apple’s sizeable cable, dongle & case business.)
Meanwhile, since both the Watch and AirPods on one side and the services on the other are all essentially about the attach rate to iPhone users, you could group them together as one big upsell, which suggests a different chart: half of Apple’s revenue is the iPhone and another third is iPhone upsells - 80% in total.
I think the car project was the classic Apple of Steve Jobs. Apple spent a lot of time and money trying to work out whether it could bring something new, and said no. . The shift to electric is destabilising the car industry and creating lots of questions about who builds cars and how they build them, and that’s a situation that should attract Apple. However, I also think that Apple concluded that while there was scope to make a great car, and perhaps one that did a few things better, there wasn’t really scope to do something fundamentally different, and solve some problem that no-one else was solving. Apple would only be making another EV, not redefining what ‘car’ means, because EVs are still basically cars - which is Tesla’s problem. It looks like the EV market will play out not like smartphones, where Apple had something unique, but like Android, where there was frenzied competition in a low-margin commodity market. So, Apple walked away - it said no.
People often suggest that Apple should buy anything from Netflix to telcos to banks, and I used to make fun of this by suggesting that Apple should buy an airline ‘because it could make the seats and the screens better’. Yes, Apple could maybe make better seats than Collins Aerospace, but that’s not what it means to run an airline. Where can Apple change the fundamental questions?
It ships MVPs that get better later, sure, and the original iPhone and Watch were MVPs, but the original iPhone also was the best phone I’d ever owned even with no 3G and no App Store. It wasn’t a concept. it wasn’t a vision of the future- it was the future. The Vision Pro is a concept, or a demo, and Apple doesn’t ship demos. Why did it ship the Vision Pro? What did it achieve? It didn’t sell in meaningful volume, because it couldn’t, and it didn’t lead to much developer activity ether, because no-one bought it. A lot of people even at Apple are puzzled.
The new Siri that’s been delayed this week is the mirror image of this. Last summer Apple told a very clear, coherent, compelling story of how it would combine the software frameworks it’s already built with the personal data in apps spread across your phones and the capabilities of LLMs to produce a new kind of personal assistant. This was the eats of Apple - taking a new primary technology and proposing way to make it useful for everyone else
·ben-evans.com·
Apple innovation and execution — Benedict Evans
Something Is Rotten in the State of Cupertino
Something Is Rotten in the State of Cupertino
Who decided these features should go in the WWDC keynote, with a promise they’d arrive in the coming year, when, at the time, they were in such an unfinished state they could not be demoed to the media even in a controlled environment? Three months later, who decided Apple should double down and advertise these features in a TV commercial, and promote them as a selling point of the iPhone 16 lineup — not just any products, but the very crown jewels of the company and the envy of the entire industry — when those features still remained in such an unfinished or perhaps even downright non-functional state that they still could not be demoed to the press? Not just couldn’t be shipped as beta software. Not just couldn’t be used by members of the press in a hands-on experience, but could not even be shown to work by Apple employees on Apple-controlled devices in an Apple-controlled environment? But yet they advertised them in a commercial for the iPhone 16, when it turns out they won’t ship, in the best case scenario, until months after the iPhone 17 lineup is unveiled?
“Can anyone tell me what MobileMe is supposed to do?” Having received a satisfactory answer, he continued, “So why the fuck doesn’t it do that?” For the next half-hour Jobs berated the group. “You’ve tarnished Apple’s reputation,” he told them. “You should hate each other for having let each other down.” The public humiliation particularly infuriated Jobs. Walt Mossberg, the influential Wall Street Journal gadget columnist, had panned MobileMe. “Mossberg, our friend, is no longer writing good things about us,” Jobs said. On the spot, Jobs named a new executive to run the group. Tim Cook should have already held a meeting like that to address and rectify this Siri and Apple Intelligence debacle. If such a meeting hasn’t yet occurred or doesn’t happen soon, then, I fear, that’s all she wrote. The ride is over. When mediocrity, excuses, and bullshit take root, they take over. A culture of excellence, accountability, and integrity cannot abide the acceptance of any of those things, and will quickly collapse upon itself with the acceptance of all three.
·daringfireball.net·
Something Is Rotten in the State of Cupertino
Your "Per-Seat" Margin is My Opportunity
Your "Per-Seat" Margin is My Opportunity

Traditional software is sold on a per seat subscription. More humans, more money. We are headed to a future where AI agents will replace the work humans do. But you can’t charge agents a per seat cost. So we’re headed to a world where software will be sold on a consumption model (think tasks) and then on an outcome model (think job completed) Incumbents will be forced to adapt but it’s classic innovators dilemma. How do you suddenly give up all that subscription revenue? This gives an opportunity for startups to win.

Per-seat pricing only works when your users are human. But when agents become the primary users of software, that model collapses.
Executives aren't evaluating software against software anymore. They're comparing the combined costs of software licenses plus labor against pure outcome-based solutions. Think customer support (per resolved ticket vs. per agent + seat), marketing (per campaign vs. headcount), sales (per qualified lead vs. rep). That's your pricing umbrella—the upper limit enterprises will pay before switching entirely to AI.
enterprises are used to deterministic outcomes and fixed annual costs. Usage-based pricing makes budgeting harder. But individual leaders seeing 10x efficiency gains won't wait for procurement to catch up. Savvy managers will find ways around traditional buying processes.
This feels like a generational reset of how businesses operate. Zero upfront costs, pay only for outcomes—that's not just a pricing model. That's the future of business.
The winning strategy in my books? Give the platform away for free. Let your agents read and write to existing systems through unstructured data—emails, calls, documents. Once you handle enough workflows, you become the new system of record.
·writing.nikunjk.com·
Your "Per-Seat" Margin is My Opportunity
Building LLMs is probably not going be a brilliant business
Building LLMs is probably not going be a brilliant business
In the 1960s, airlines were The Future. That is why old films have so many swish shots of airports in them. Airlines though, turned out to be an unavoidably rubbish business. I've flown on loads of airlines that have gone bust: Monarch, WOW Air, Thomas Cook, Flybmi, Zoom. And those are all busts from before coronavirus - times change but being an airline is always a bad idea.
That's odd, because other businesses, even ones which seem really stupid, are much more profitable. Selling fizzy drinks is, surprisingly, an amazing business. Perhaps the best. Coca-Cola's return on equity has rarely fallen below 30% in any given year. That seems very unfair because being an airline is hard work but making coke is pretty easy. It's even more galling because Coca-Cola don't actually make the coke themselves - that is outsourced to "bottling companies". They literally just sell it.
If you were to believe LinkedIn you would think a great business is made with efficiency, hard work, innovation or some other intrinsic reason to do with how hardworking, or clever, the people in the business are. That simply is not the case. What makes a good business is industry structure
Classically, there are five basic parts ("forces") to a company's position: The power of their suppliers to increase their prices The power of their buyers to reduce your prices The strength of direct competitors The threat of any new entrants The threat of substitutes It's industry structure that makes a business profitable or not. Not efficiency, not hard work and not innovation. If none of the forces are very much against you, your business will do ok. If they are all against you, you'll be in the position of the airlines. And if they're all in your favour: brill, you're Coca-Cola.
·calpaterson.com·
Building LLMs is probably not going be a brilliant business
The art of the pivot, part 2: How, why and when to pivot
The art of the pivot, part 2: How, why and when to pivot
people mix up two very different types of pivots and that it’s important to differentiate which path you’re on: Ideation pivots: This is when an early-stage startup changes its idea before having a fully formed product or meaningful traction. These pivots are easy to make, normally happen quickly after launch, and the new idea is often completely unrelated to the previous one. For example, Brex went from VR headsets to business banking, Retool went from Venmo for the U.K. to a no-code internal tools app, and Okta went from reliability monitoring to identity management all in under three months. YouTube changed direction from a dating site to a video streaming platform in less than a week. Hard pivots: This is when a company with a live product and real users/customers changes direction. In these cases, you are truly “pivoting”—keeping one element of the previous idea and doubling down on it. For example, Instagram stripped down its check-in app and went all in on its photo-sharing feature, Slack on its internal chat tool, and Loom on its screen recording feature. Occasionally a pivot is a mix of the two (i.e. you’re pivoting multiple times over 1+ years), but generally, when you’re following the advice below, make sure you’re clear on which category you’re in.
When looking at the data, a few interesting trends emerged: Ideation pivots generally happen within three months of launching your original idea. Note, a launch at this stage is typically just telling a bunch of your friends and colleagues about it. Hard pivots generally happen within two years after launch, and most around the one-year mark. I suspect the small number of companies that took longer regret not changing course earlier.
ou should have a hard conversation with your co-founder around the three-month mark, and depending on how it’s going (see below), either re-commit or change the idea. Then schedule a yearly check-in. If things are clicking, full speed ahead. If things feel meh, at least spend a few days talking about other potential directions.
Brex: “We applied to YC with this VR idea, which, looking back, it was pretty bad, but at the time we thought it was great. And within YC, we were like, ‘Yeah, we don’t even know where to start to build this.’” —Henrique Dubugras, co-founder and CEO
·lennysnewsletter.com·
The art of the pivot, part 2: How, why and when to pivot
The Right Kind of Stubborn
The Right Kind of Stubborn
Graham argues that persistence is more complex and effective in solving hard problems, while obstinacy is simpler and less likely to lead to success.
the obstinate don't want to hear you. When you point out problems, their eyes glaze over, and their replies sound like ideologues talking about matters of doctrine.
The reason the persistent and the obstinate seem similar is that they're both hard to stop. But they're hard to stop in different senses. The persistent are like boats whose engines can't be throttled back. The obstinate are like boats whose rudders can't be turned.
There will be some resistance to turning the rudder of a persistent person, because there's some cost to changing direction.
In the degenerate case they're indistinguishable: when there's only one way to solve a problem, your only choice is whether to give up or not, and persistence and obstinacy both say no. This is presumably why the two are so often conflated in popular culture. It assumes simple problems. But as problems get more complicated, we can see the difference between them. The persistent are much more attached to points high in the decision tree than to minor ones lower down, while the obstinate spray "don't give up" indiscriminately over the whole tree.
The persistent are attached to the goal. The obstinate are attached to their ideas about how to reach it.
the persistent must also be imaginative. To keep trying things, you have to keep thinking of things to try
persistence often requires that one change one's mind. That's where good judgement comes in. The persistent are quite rational. They focus on expected value. It's this, not recklessness, that lets them work on things that are unlikely to succeed.
in practice your energy and imagination and resilience and good judgement have to be directed toward some fairly specific goal. Not too specific, or you might miss a great discovery adjacent to what you're searching for, but not too general, or it won't work to motivate you.
When you look at the internal structure of persistence, it doesn't resemble obstinacy at all. It's so much more complex. Five distinct qualities — energy, imagination, resilience, good judgement, and focus on a goal — combine to produce a phenomenon that seems a bit like obstinacy in the sense that it causes you not to give up. But the way you don't give up is completely different. Instead of merely resisting change, you're driven toward a goal by energy and resilience, through paths discovered by imagination and optimized by judgement. You'll give way on any point low down in the decision tree, if its expected value drops sufficiently, but energy and resilience keep pushing you toward whatever you chose higher up.
·paulgraham.com·
The Right Kind of Stubborn
WWDC 2024: Apple Intelligence
WWDC 2024: Apple Intelligence
their models are almost entirely based on personal context, by way of an on-device semantic index. In broad strokes, this on-device semantic index can be thought of as a next-generation Spotlight. Apple is focusing on what it can do that no one else can on Apple devices, and not really even trying to compete against ChatGPT et al. for world-knowledge context. They’re focusing on unique differentiation, and eschewing commoditization.
Apple is doing what no one else can do: integrating generative AI into the frameworks in iOS and MacOS used by developers to create native apps. Apps built on the system APIs and frameworks will gain generative AI features for free, both in the sense that the features come automatically when the app is running on a device that meets the minimum specs to qualify for Apple Intelligence, and in the sense that Apple isn’t charging developers or users to utilize these features.
·daringfireball.net·
WWDC 2024: Apple Intelligence
Apple intelligence and AI maximalism — Benedict Evans
Apple intelligence and AI maximalism — Benedict Evans
The chatbot might replace all software with a prompt - ‘software is dead’. I’m skeptical about this, as I’ve written here, but Apple is proposing the opposite: that generative AI is a technology, not a product.
Apple is, I think, signalling a view that generative AI, and ChatGPT itself, is a commodity technology that is most useful when it is: Embedded in a system that gives it broader context about the user (which might be search, social, a device OS, or a vertical application) and Unbundled into individual features (ditto), which are inherently easier to run as small power-efficient models on small power-efficient devices on the edge (paid for by users, not your capex budget) - which is just as well, because… This stuff will never work for the mass-market if we have marginal cost every time the user presses ‘OK’ and we need a fleet of new nuclear power-stations to run it all.
Apple has built its own foundation models, which (on the benchmarks it published) are comparable to anything else on the market, but there’s nowhere that you can plug a raw prompt directly into the model and get a raw output back - there are always sets of buttons and options shaping what you ask, and that’s presented to the user in different ways for different features. In most of these features, there’s no visible bot at all. You don’t ask a question and get a response: instead, your emails are prioritised, or you press ‘summarise’ and a summary appears. You can type a request into Siri (and Siri itself is only one of the many features using Apple’s models), but even then you don’t get raw model output back: you get GUI. The LLM is abstracted away as an API call.
Apple is treating this as a technology to enable new classes of features and capabilities, where there is design and product management shaping what the technology does and what the user sees, not as an oracle that you ask for things.
Apple is drawing a split between a ‘context model’ and a ‘world model’. Apple’s models have access to all the context that your phone has about you, powering those features, and this is all private, both on device and in Apple’s ‘Private Cloud’. But if you ask for ideas for what to make with a photo of your grocery shopping, then this is no longer about your context, and Apple will offer to send that to a third-party world model - today, ChatGPT.
that’s clearly separated into a different experience where you should have different expectations, and it’s also, of course, OpenAI’s brand risk, not Apple’s. Meanwhile, that world model gets none of your context, only your one-off prompt.
Neither OpenAI nor any of the other cloud models from new companies (Anthropic, Mistral etc) have your emails, messages, locations, photos, files and so on.
Apple is letting OpenAI take the brand risk of creating pizza glue recipes, and making error rates and abuse someone else’s problem, while Apple watches from a safe distance.
The next step, probably, is to take bids from Bing and Google for the default slot, but meanwhile, more and more use-cases will be quietly shifted from the third party to Apple’s own models. It’s Apple’s own software that decides where the queries go, after all, and which ones need the third party at all.
A lot of the compute to run Apple Intelligence is in end-user devices paid for by the users, not Apple’s capex budget, and Apple Intelligence is free.
Commoditisation is often also integration. There was a time when ‘spell check’ was a separate product that you had to buy, for hundreds of dollars, and there were dozens of competing products on the market, but over time it was integrated first into the word processor and then the OS. The same thing happened with the last wave of machine learning - style transfer or image recognition were products for five minutes and then became features. Today ‘summarise this document’ is AI, and you need a cloud LLM that costs $20/month, but tomorrow the OS will do that for free. ‘AI is whatever doesn’t work yet.’
Apple is big enough to take its own path, just as it did moving the Mac to its own silicon: it controls the software and APIs on top of the silicon that are the basis of those developer network effects, and it has a world class chip team and privileged access to TSMC.
Apple is doing something slightly different - it’s proposing a single context model for everything you do on your phone, and powering features from that, rather than adding disconnected LLM-powered features at disconnected points across the company.
·ben-evans.com·
Apple intelligence and AI maximalism — Benedict Evans
What Apple's AI Tells Us: Experimental Models⁴
What Apple's AI Tells Us: Experimental Models⁴
Companies are exploring various approaches, from large, less constrained frontier models to smaller, more focused models that run on devices. Apple's AI focuses on narrow, practical use cases and strong privacy measures, while companies like OpenAI and Anthropic pursue the goal of AGI.
the most advanced generalist AI models often outperform specialized models, even in the specific domains those specialized models were designed for. That means that if you want a model that can do a lot - reason over massive amounts of text, help you generate ideas, write in a non-robotic way — you want to use one of the three frontier models: GPT-4o, Gemini 1.5, or Claude 3 Opus.
Working with advanced models is more like working with a human being, a smart one that makes mistakes and has weird moods sometimes. Frontier models are more likely to do extraordinary things but are also more frustrating and often unnerving to use. Contrast this with Apple’s narrow focus on making AI get stuff done for you.
Every major AI company argues the technology will evolve further and has teased mysterious future additions to their systems. In contrast, what we are seeing from Apple is a clear and practical vision of how AI can help most users, without a lot of effort, today. In doing so, they are hiding much of the power, and quirks, of LLMs from their users. Having companies take many approaches to AI is likely to lead to faster adoption in the long term. And, as companies experiment, we will learn more about which sets of models are correct.
·oneusefulthing.org·
What Apple's AI Tells Us: Experimental Models⁴
Apple Intelligence is Right On Time
Apple Intelligence is Right On Time

Summary

  • Apple remains primarily a hardware company, and an AI-mediated future will still require devices, playing to Apple's strengths in design and integration.
  • AI is a complement to Apple's business, not disruptive, as it makes high-performance hardware more relevant and could drive meaningful iPhone upgrade cycles.
  • The smartphone is the ideal device for most computing tasks and the platform on which the future happens, solidifying the relevance of Apple's App Store ecosystem.
  • Apple's partnership with OpenAI for chatbot functionality allows it to offer best-in-class capabilities without massive investments, while reducing the threat of OpenAI building a competing device.
  • Building out the infrastructure for API-level AI features is a challenge for Apple, but one that is solvable given its control over the interface and integration of on-device and cloud processing.
  • The only significant threat to Apple is Google, which could potentially develop differentiated AI capabilities for Android that drive switching from iPhone users, though this is uncertain.
  • Microsoft's missteps with its Recall feature demonstrate the risks of pushing AI features too aggressively, validating Apple's more cautious approach.
  • Apple's user-centric orientation and brand promise of privacy and security align well with the need to deliver AI features in an integrated, trustworthy manner.
·stratechery.com·
Apple Intelligence is Right On Time
Gemini 1.5 and Google’s Nature
Gemini 1.5 and Google’s Nature
Google is facing many of the same challenges after its decades long dominance of the open web: all of the products shown yesterday rely on a different business model than advertising, and to properly execute and deliver on them will require a cultural shift to supporting customers instead of tolerating them. What hasn’t changed — because it is the company’s nature, and thus cannot — is the reliance on scale and an overwhelming infrastructure advantage. That, more than anything, is what defines Google, and it was encouraging to see that so explicitly put forward as an advantage.
·stratechery.com·
Gemini 1.5 and Google’s Nature
AI Integration and Modularization
AI Integration and Modularization
Summary: The question of integration versus modularization in the context of AI, drawing on the work of economists Ronald Coase and Clayton Christensen. Google is pursuing a fully integrated approach similar to Apple, while AWS is betting on modularization, and Microsoft and Meta are somewhere in between. Integration may provide an advantage in the consumer market and for achieving AGI, but that for enterprise AI, a more modular approach leveraging data gravity and treating models as commodities may prevail. Ultimately, the biggest beneficiary of this dynamic could be Nvidia.
The left side of figure 5-1 indicates that when there is a performance gap — when product functionality and reliability are not yet good enough to address the needs of customers in a given tier of the market — companies must compete by making the best possible products. In the race to do this, firms that build their products around proprietary, interdependent architectures enjoy an important competitive advantage against competitors whose product architectures are modular, because the standardization inherent in modularity takes too many degrees of design freedom away from engineers, and they cannot not optimize performance.
The issue I have with this analysis of vertical integration — and this is exactly what I was taught at business school — is that the only considered costs are financial. But there are other, more difficult to quantify costs. Modularization incurs costs in the design and experience of using products that cannot be overcome, yet cannot be measured. Business buyers — and the analysts who study them — simply ignore them, but consumers don’t. Some consumers inherently know and value quality, look-and-feel, and attention to detail, and are willing to pay a premium that far exceeds the financial costs of being vertically integrated.
Google trains and runs its Gemini family of models on its own TPU processors, which are only available on Google’s cloud infrastructure. Developers can access Gemini through Vertex AI, Google’s fully-managed AI development platform; and, to the extent Vertex AI is similar to Google’s internal development environment, that is the platform on which Google is building its own consumer-facing AI apps. It’s all Google, from top-to-bottom, and there is evidence that this integration is paying off: Gemini 1.5’s industry leading 2 million token context window almost certainly required joint innovation between Google’s infrastructure team and its model-building team.
In AI, Google is pursuing an integrated strategy, building everything from chips to models to applications, similar to Apple's approach in smartphones.
On the other extreme is AWS, which doesn’t have any of its own models; instead its focus has been on its Bedrock managed development platform, which lets you use any model. Amazon’s other focus has been on developing its own chips, although the vast majority of its AI business runs on Nvidia GPUs.
Microsoft is in the middle, thanks to its close ties to OpenAI and its models. The company added Azure Models-as-a-Service last year, but its primary focus for both external customers and its own internal apps has been building on top of OpenAI’s GPT family of models; Microsoft has also launched its own chip for inference, but the vast majority of its workloads run on Nvidia.
Google is certainly building products for the consumer market, but those products are not devices; they are Internet services. And, as you might have noticed, the historical discussion didn’t really mention the Internet. Both Google and Meta, the two biggest winners of the Internet epoch, built their services on commodity hardware. Granted, those services scaled thanks to the deep infrastructure work undertaken by both companies, but even there Google’s more customized approach has been at least rivaled by Meta’s more open approach. What is notable is that both companies are integrating their models and their apps, as is OpenAI with ChatGPT.
Google's integrated AI strategy is unique but may not provide a sustainable advantage for Internet services in the way Apple's integration does for devices
It may be the case that selling hardware, which has to be perfect every year to justify a significant outlay of money by consumers, provides a much better incentive structure for maintaining excellence and execution than does being an Aggregator that users access for free.
Google’s collection of moonshots — from Waymo to Google Fiber to Nest to Project Wing to Verily to Project Loon (and the list goes on) — have mostly been science projects that have, for the most part, served to divert profits from Google Search away from shareholders. Waymo is probably the most interesting, but even if it succeeds, it is ultimately a car service rather far afield from Google’s mission statement “to organize the world’s information and make it universally accessible and useful.”
The only thing that drives meaningful shifts in platform marketshare are paradigm shifts, and while I doubt the v1 version of Pixie [Google’s rumored Pixel-only AI assistant] would be good enough to drive switching from iPhone users, there is at least a path to where it does exactly that.
the fact that Google is being mocked mercilessly for messed-up AI answers gets at why consumer-facing AI may be disruptive for the company: the reason why incumbents find it hard to respond to disruptive technologies is because they are, at least at the beginning, not good enough for the incumbent’s core offering. Time will tell if this gives more fuel to a shift in smartphone strategies, or makes the company more reticent.
while I was very impressed with Google’s enterprise pitch, which benefits from its integration with Google’s infrastructure without all of the overhead of potentially disrupting the company’s existing products, it’s going to be a heavy lift to overcome data gravity, i.e. the fact that many enterprise customers will simply find it easier to use AI services on the same clouds where they already store their data (Google does, of course, also support non-Gemini models and Nvidia GPUs for enterprise customers). To the extent Google wins in enterprise it may be by capturing the next generation of startups that are AI first and, by definition, data light; a new company has the freedom to base its decision on infrastructure and integration.
Amazon is certainly hoping that argument is correct: the company is operating as if everything in the AI value chain is modular and ultimately a commodity, which insinuates that it believes that data gravity will matter most. What is difficult to separate is to what extent this is the correct interpretation of the strategic landscape versus a convenient interpretation of the facts that happens to perfectly align with Amazon’s strengths and weaknesses, including infrastructure that is heavily optimized for commodity workloads.
Unclear if Amazon's strategy is based on true insight or motivated reasoning based on their existing strengths
Meta’s open source approach to Llama: the company is focused on products, which do benefit from integration, but there are also benefits that come from widespread usage, particularly in terms of optimization and complementary software. Open source accrues those benefits without imposing any incentives that detract from Meta’s product efforts (and don’t forget that Meta is receiving some portion of revenue from hyperscalers serving Llama models).
The iPhone maker, like Amazon, appears to be betting that AI will be a feature or an app; like Amazon, it’s not clear to what extent this is strategic foresight versus motivated reasoning.
achieving something approaching AGI, whatever that means, will require maximizing every efficiency and optimization, which rewards the integrated approach.
the most value will be derived from building platforms that treat models like processors, delivering performance improvements to developers who never need to know what is going on under the hood.
·stratechery.com·
AI Integration and Modularization
Companionship Content is King - by Anu Atluru
Companionship Content is King - by Anu Atluru

Long-form "companionship content" will outlast short-form video formats like TikTok, as the latter is more mentally draining and has a lower ceiling for user engagement over time.

  • In contrast, companionship content that feels more human and less algorithmically optimized will continue to thrive, as it better meets people's needs for social connection and low-effort entertainment.
  • YouTube as the dominant platform among teens, and notes that successful TikTok creators often funnel their audiences to longer-form YouTube content.
  • Platforms enabling deep, direct creator-fan relationships and higher creator payouts, like YouTube, are expected to be the long-term winners in the content landscape.
Companionship content is long-form content that can be consumed passively — allowing the consumer to be incompletely attentive, and providing a sense of relaxation, comfort, and community.
Interestingly, each individual “unit” of music is short-form (e.g. a 3-5 minute song), but how we consume it tends to be long-form and passive (i.e. via curated stations, lengthy playlists, or algorithms that adapt to our taste).
If you’re rewatching a show or movie, it’s likely to be companionship content. (Life-like conversational sitcoms can be consumed this way too.) As streaming matures, platforms are growing their passive-watch library.
content isn’t always prescriptively passive, rather it’s rooted in how consumers engage it.
That said, some content lends better to being companionship content: Long-form over short. Conversational over action. Simple plot versus complex.
Short-form video requires more attention & action in a few ways: Context switching, i.e. wrapping your head around a new piece of context every 30 seconds, especially if they’re on unrelated topics with different styles Judgment & decision-making, i.e. contemplating whether to keep watching or swipe to the next video effectively the entire time you’re watching a video Multi-sensory attention, i.e. default full-screen and requires visual and audio focus, especially since videos are so short that you can easily lose context Interactive components, e.g. liking, saving, bookmarking,
With how performative, edited, and algorithmically over-optimized it is, TikTok feels sub-human. TikTok has quickly become one of the most goal-seeking places on earth. I could easily describe TikTok as a global focus group for commercials. It’s the product personification of a means to an end, and the end is attention.
even TikTok creators are adapting the historically rigid format to appeal to more companionship-esque emotions and improve retention.
When we search for a YouTube video to watch, we often want the best companion for the next hour and not the most entertaining content.
While short-form content edits are meant to be spectacular and attention-grabbing, long-form content tends to be more subtle in its emotional journey Long-form engagement with any single character or narrative or genre lets you develop stronger understanding, affinity, and parasocial bonds Talk-based content (e.g. talk shows, podcasts, comedy, vlogs, life-like sitcoms) especially evokes a feeling of companionship and is less energy-draining The trends around loneliness and the acceleration of remote work has and will continue to make companionship content even more desirable As we move into new technology frontiers, we might unlock novel types of companionship content itself, but I’d expect this to take 5-10 years at least
TikTok is where you connect with an audience, YouTube is where you consolidate it.5 Long-form content also earns creators more, with YouTube a standout in revenue sharing.
YouTube paid out $16 billion to creators in 2022 (which is 55% of its annual $30 billion in revenue) and the other four social networks paid out about $1 billion each from their respective creator funds. In total, that yields $20 billion.”
Mr. Beast, YouTube’s top creator, says YouTube is now the final destination, not “traditional” hollywood stardom which is the dream of generations past. Creators also want to funnel audiences to apps & community platforms where they can own user relationships, rely less on algorithms, engage more directly and deeply with followers, and enable follower-to-follower engagement too
Interestingly of course, an increasing amount of short-form video, including formats like clips and edits, seems to be made from what originally was long-form content.8 And in return, these recycled short-form videos can drive tremendous traffic to long-form formats and platforms.
90% of people use a second screen while watching TV. We generally talk about “second screen” experiences in the context of multiple devices, but you can have complementary apps and content running on the same device — you can have the “second screen” on the same screen.
YouTube itself also cites a trend of people putting YouTube on their real TV screens: “There are more Americans gathering around the living room TV to watch YouTube than any other platform. Why? Put simply, people want choices and variety … It’s a one stop shop for video viewing. Think about something historically associated with linear TV: Sports. Now, with [our NFL partnership], people can not only watch the games, but watch post-game highlights and commentary in one place.”
If I were to build an on-demand streaming product or any kind of content product for that matter, I’d build for the companionship use case — not only because I think it has a higher ceiling of consumer attention, but also because it can support more authentic, natural, human engagement.
All the creators that are ‘made’ on TikTok are looking for a place to go to consolidate the attention they’ve amassed. TikTok is commercials. YouTube is TV. (Though yes, they’re both trying to become each other).
certainly AI and all the new creator tools enabled by it will help people mix and match and remix long and short formats all day, blurring the historically strict distinctions between them. It’ll take some time before we see a new physical product + content combo thrive, and meanwhile the iPhone and its comps will be competing hard to stay the default device.
The new default seems to be that we’re not lonely as long as we’re streaming. We can view this entirely in a negative light and talk about how much the internet and media is contributing to the loneliness epidemic. Or we could think about how to create media for good. Companionship content can be less the quick dopamine-hit-delivering clips and more of this, and perhaps even truly social.
Long-form wants to become the conversational third space for consumers too. The “comments” sections of TikTok, YouTube and all broadcast platforms are improving, but they still have a long way to go before they become even more community-oriented.
I’m not an “AI-head” but I am more curious about what it’s going to enable in long-form content than all the short-form clips it’s going to help generate and illustrate, etc.
The foreground tends to be utilities or low-cognitive / audio effort (text or silent video). Tiktok is a foreground app for now, YouTube is both (and I’d say trending towards being background).
·archive.is·
Companionship Content is King - by Anu Atluru
Welcome to the video bloat era
Welcome to the video bloat era
A Pivot To Video tends to arrive in stages, with each stage being more expensive and producing less interesting content as things progress. Usually it goes like this: The experimentation phase, the factory phase, and the bloat phase. A great editor I worked for during the second Pivot To Video, roughly 2013-2017, who, herself worked through the first, roughly 2003-2007, described it as a massive waste of resources that wastes more resources as it becomes clearer to everyone not directly involved how much of a waste of resources it is.
It’s a fundamental issue with video as a medium that online platforms haven’t fixed and, I suspect, never will because it makes user-generated content platforms feel more professional and consistent. Like TV. The cost to produce video content always balloons as you add more people, more tools, more structure to the workflow, pushing out smaller creators and teams. And even with the pandemic lowering the barrier of entry for making video online considerably, it’s still happening again. We’re in the bloat phase now.
MrBeast, the platform’s biggest star, is spending between $3-$5 million per video right now, up from around $200,000 a video just a few years ago. To put that absolutely outrageous number in perspective, a MrBeast video is roughly the same cost per video as any episode from the first five seasons of Game Of Thrones.
Guides last year were saying you had to capture viewers in the first three seconds. I’ve read a few guides from this year that are now saying hooking a TikTok user has to happen in the first 1.5 seconds. There’s an oft-quoted “shoeshine boy” theory of markets, usually attributed to Joe Kennedy in the late 1920s, who said that when the boy shining his shoes had stock tips, he knew the market was about to collapse. Well, here’s a similar rule for digital video: If you’re trying to optimize your video in microseconds, the video pivot is probably already over.
YouTube is laser-focused on capturing the world’s televisions. In fact, the platform’s CEO, Neal Mohan announced yesterday that the platform is adding even more features for YouTube’s TV app. And TikTok, if it’s not banned or whatever, is trying to use its massive inventory of short-form video content to prop up both a search engine and an e-commerce operation. And we haven’t even talked about Meta’s video products here. There is simply no incentive for these platforms to regress even though users seem to want them to.
Tastes are clearly changing. The Washington Post article pointed to Sam Sulek, a giant muscleman on YouTube who posts 30-minute workout vlogs with barely any editing as a possible direction this is all headed in. I tried watching one of his recent videos and I’m not even sure it has any cuts in it? It’s possible that’s what’s coming next, but it’s less certain if platforms will, or rather can, allow it. Time to find out if they know how to pivot.
·garbageday.email·
Welcome to the video bloat era
A design reset (part I) - Linear blog
A design reset (part I) - Linear blog
On advocating for a widespread product redesign at a company that resists it
The challenges start from the fact that it's never a good time to do a redesign. It's hard to make it a priority. It's difficult to calculate the ROI on it. And if you run your product with A/B testing, every global redesign will tank the metrics in the short term.
The real need for redesigns often comes when you have created a successful product and it has evolved with the market and users over time.
We ship small changes daily, and something major almost every week. Every year, it's almost like a new product. This incremental way of building the product is hugely beneficial, and often necessary — though it unbalances the overall design, and leads to design debt. Each new capability adds stress on the product's existing surfaces for which it was initially designed. Functionality no longer fits in a coherent way. It needs to be rebalanced and rethought.
If your product evolves fast, you should be paying this debt every 2-3 years. The longer you wait and the more successful your product becomes, the more you will have to untangle.
Slowly the user sentiment and perception might start turning negative and you might start looking like a dinosaur incumbent. This leaves an opportunity for some nimbler player to come along and compete in your market. Companies often try to address this with brand refreshes, but if you don’t refresh the product, nothing truly changes about the experience.
While the design debt often happens in small increments, it’s best to be paid in larger sweeps. This goes against the common wisdom in engineering where complete code rewrites are avoided. The difference is that on the engineering side, a modular or incremental way of working can work as the technical implementation is not really visible. Whereas the product experience is holistic and visual. You cannot predict which path the user takes. If you update just one module or view at a time, the overall experience becomes more disjointed. Secondly, if your goal is to reset and rebalance the whole product UI and experience, you have to consider all the needs simultaneously. An incremental approach doesn’t let you do that.
I’ve never seen redesigns successfully executed without the CEO behind it. While design might have a seat at the table generally, they are usually not able to convince everyone around that table. Only the CEO can push through all the excuses and give the latitude to a project touching all of the surfaces the product needs.
The way to get the CEO involved is to tie a design reset into a larger company shift or directional change. For example, if a company is looking at a new product, or major new feature, a redesign project can be a way to imagine how it might look or feel. This can be the justification for why you need to spin up the team (and at the same time, you can make a case for updating the rest of the product experience).
Organizations are often quite stuck in their views and ways of doing things, making them less enthusiastic about something new. When I was at Airbnb, the mobile redesign project was a way to shift the company to become mobile-first. It set the tone and got the message across to the whole company that mobile was happening and that it was happening now. While it looks like an obvious change in hindsight, there were many arguments against it at the time and it took a lot of convincing. Switching to think about mobile meant the design and features had to be rethought to work in that platform.
While Linear is a smaller and younger company, we’re also undergoing a shift. The product vision has widened from a simple issue tracker to a purpose-built system for product development. We are now moving into planning workflows that naturally come before the building or execution phase of building products. This product evolution creates new future needs from the product design, and we have to make space for it.
When you realize that a design reset is needed for your product, how do you actually get started with the project? You start with a standalone team to explore the new concept design and create something the company can rally around.The auto industry has a practice of building “concept cars”, where they explore the next version of the car freely and boldly without considering practicality. A concept car sets the direction, but usually is not expected to land in production because it’s too impractical or costly to manufacture.
A secret I've learned is that when you tell people a design is a "concept" or "conceptual" it makes it less likely that the idea is attacked from whatever perspective they hold or problems they see with it. The concept is not perceived as real, but something that can be entertained. By bringing leaders or even teams along with the concept iterations, it starts to solidify the new direction in their mind, eventually becoming more and more familiar. That's the power of visual design.
·linear.app·
A design reset (part I) - Linear blog
AI startups require new strategies
AI startups require new strategies

comment from Habitue on Hacker News: > These are some good points, but it doesn't seem to mention a big way in which startups disrupt incumbents, which is that they frame the problem a different way, and they don't need to protect existing revenue streams.

The “hard tech” in AI are the LLMs available for rent from OpenAI, Anthropic, Cohere, and others, or available as open source with Llama, Bloom, Mistral and others. The hard-tech is a level playing field; startups do not have an advantage over incumbents.
There can be differentiation in prompt engineering, problem break-down, use of vector databases, and more. However, this isn’t something where startups have an edge, such as being willing to take more risks or be more creative. At best, it is neutral; certainly not an advantage.
This doesn’t mean it’s impossible for a startup to succeed; surely many will. It means that you need a strategy that creates differentiation and distribution, even more quickly and dramatically than is normally required
Whether you’re training existing models, developing models from scratch, or simply testing theories, high-quality data is crucial. Incumbents have the data because they have the customers. They can immediately leverage customers’ data to train models and tune algorithms, so long as they maintain secrecy and privacy.
Intercom’s AI strategy is built on the foundation of hundreds of millions of customer interactions. This gives them an advantage over a newcomer developing a chatbot from scratch. Similarly, Google has an advantage in AI video because they own the entire YouTube library. GitHub has an advantage with Copilot because they trained their AI on their vast code repository (including changes, with human-written explanations of the changes).
While there will always be individuals preferring the startup environment, the allure of working on AI at an incumbent is equally strong for many, especially pure computer and data scientsts who, more than anything else, want to work on interesting AI projects. They get to work in the code, with a large budget, with all the data, with above-market compensation, and a built-in large customer base that will enjoy the fruits of their labor, all without having to do sales, marketing, tech support, accounting, raising money, or anything else that isn’t the pure joy of writing interesting code. This is heaven for many.
A chatbot is in the chatbot market, and an SEO tool is in the SEO market. Adding AI to those tools is obviously a good idea; indeed companies who fail to add AI will likely become irrelevant in the long run. Thus we see that “AI” is a new tool for developing within existing markets, not itself a new market (except for actual hard-tech AI companies).
AI is in the solution-space, not the problem-space, as we say in product management. The customer problem you’re solving is still the same as ever. The problem a chatbot is solving is the same as ever: Talk to customers 24/7 in any language. AI enables completely new solutions that none of us were imagining a few years ago; that’s what’s so exciting and truly transformative. However, the customer problems remain the same, even though the solutions are different
Companies will pay more for chatbots where the AI is excellent, more support contacts are deferred from reaching a human, more languages are supported, and more kinds of questions can be answered, so existing chatbot customers might pay more, which grows the market. Furthermore, some companies who previously (rightly) saw chatbots as a terrible customer experience, will change their mind with sufficiently good AI, and will enter the chatbot market, which again grows that market.
the right way to analyze this is not to say “the AI market is big and growing” but rather: “Here is how AI will transform this existing market.” And then: “Here’s how we fit into that growth.”
·longform.asmartbear.com·
AI startups require new strategies
Rethinking the startup MVP - Building a competitive product - Linear
Rethinking the startup MVP - Building a competitive product - Linear
Building something valuable is no longer about validating a novel idea as fast as possible. Instead, the modern MVP exercise is about building a version of an idea that is different from and better than what exists today. Most of us aren’t building for a net-new market. Rather, we’re finding opportunities to improve existing categories. We need an MVP concept that helps founders and product leaders iterate on their early ideas to compete in an existing market.
It’s not good enough to be first with an idea. You have to out-execute from day 1.
The MVP as a practice of building a hacky product as quickly and cheaply as possible to validate the product does no longer work. Many product categories are already saturated with a variety of alternatives, and to truly test the viability of any new idea you need to build something that is substantially better.
Airbnb wanted to build a service that relied on people being comfortable spending the night at a stranger’s house. When they started in 2009, it wasn’t obvious if people were ready for this. Today, it’s obvious that it works, so they wouldn’t need to validate the idea. A similar analogy works for Lyft when they started exploring ridesharing as a concept.
Today, the MVP is no longer about validating a novel idea as quickly as possible. Rather, its aim is to create a compelling product that draws in the early users in order to gather feedback that you then use to sharpen the product into the best version of many.
If you look at successful companies that have IPO'd in the recent years–Zoom, Slack, TikTok, Snowflake, Robinhood–you see examples not of novel ideas, but of these highly-refined ideas.Since many of us are building in a crowded market, the bar for a competitive, public-ready MVP is much higher than the MVP for a novel idea, since users have options. To get to this high bar, we have to spend more time refining the initial version.
The original MVP idea can still work if you’re in the fortunate position of creating a wholly new category of product or work with new technology platforms, but that becomes rarer and rarer as time goes on.
Let’s jump over the regular startup journey that you might take today when building a new product:You start with the idea on how you want to improve on existing products in a category.You build your first prototype.You iterate with your vision and based on feedback from early users.You get an inkling of product market fit and traction.Optional: You start fundraising (with demonstrable traction).Optional: You scale your team, improve the product, and go to market.
In today’s landscape, you’re likely competing against many other products. To win, you have to build a product that provides more value to your users than your competition does.To be able to do this with limited resources, you must scope down your audience (and thus your ambitions) as much as possible to make competing easier, and aim to solve the problems of specific people.
When we started Linear, our vision was to become the standard of how software is built. This is not really something you can expect to do during your early startup journey, let alone in an MVP. But you should demonstrate you have the ability to achieve your bigger vision via your early bets. We chose to do this by focusing on IC’s at small startups. We started with the smallest atomic unit of work they actually needed help with: issue tracking.
We knew we wanted our product to demonstrate three values:It should be as fast as possible (local data storage, no page reloads, available offline).It should be modern (keyboard shortcuts, command menu, contextual menus).It should be multiplayer (real-time sync and teammates presence).
Remember, you’re likely not building a revolutionary or novel product. You’re unlikely to go viral with your announcement, so you need a network of people who understand the “why” behind your product to help spread the word to drive people to sign up. Any product category has many people who are frustrated with the existing tools or ways of working. Ideally you find and are able to reach out to those people.
Once you have a bunch of people on your waitlist, you need to invite the right users at each stage of your iteration. You want to invite people who are likely to be happy with the limited set of features you’ve built so far. Otherwise, they’ll churn straight away and you’ll learn nothing.
To recap:Narrow down your initial audience and build for them: Figure out who you're building the product for and make the target audience as small as possible before expanding.Build and leverage your waitlist: The waitlist is the grinding stone with which you can sharpen your idea into something truly valuable that will succeed at market, so use it effectively.Trust your gut and validate demand with your users: Talk, talk, talk to your users and find out how invested in the product they are (and if they’d be willing to pay)
·linear.app·
Rethinking the startup MVP - Building a competitive product - Linear
Competition is overrated - cdixon
Competition is overrated - cdixon
That other people tried your idea without success could imply it’s a bad idea or simply that the timing or execution was wrong. Distinguishing between these cases is hard and where you should apply serious thought. If you think your competitors executed poorly, you should develop a theory of what they did wrong and how you’ll do better.
If you think your competitor’s timing was off, you should have a thesis about what’s changed to make now the right time. These changes could come in a variety of forms: for example, it could be that users have become more sophisticated, the prices of key inputs have dropped, or that prerequisite technologies have become widely adopted.
Startups are primarly competing against indifference, lack of awareness, and lack of understanding — not other startups.
There were probably 50 companies that tried to do viral video sharing before YouTube. Before 2005, when YouTube was founded, relatively few users had broadband and video cameras. YouTube also took advantage of the latest version of Flash that could play videos seamlessly.
Google and Facebook launched long after their competitors, but executed incredibly well and focused on the right things. When Google launched, other search engines like Yahoo, Excite, and Lycos were focused on becoming multipurpose “portals” and had de-prioritized search (Yahoo even outsourced their search technology).
·cdixon.org·
Competition is overrated - cdixon
Muse retrospective by Adam Wiggins
Muse retrospective by Adam Wiggins
  • Wiggins focused on storytelling and brand-building for Muse, achieving early success with an email newsletter, which helped engage potential users and refine the product's value proposition.
  • Muse aspired to a "small giants" business model, emphasizing quality, autonomy, and a healthy work environment over rapid growth. They sought to avoid additional funding rounds by charging a prosumer price early on.
  • Short demo videos on Twitter showcasing the app in action proved to be the most effective method for attracting new users.
Muse as a brand and a product represented something aspirational. People want to be deeper thinkers, to be more strategic, and to use cool, status-quo challenging software made by small passionate teams. These kinds of aspirations are easier to indulge in times of plenty. But once you're getting laid off from your high-paying tech job, or struggling to raise your next financing round, or scrambling to protect your kids' college fund from runaway inflation and uncertain markets... I guess you don't have time to be excited about cool demos on Twitter and thoughtful podcasts on product design.
I’d speculate that another factor is the half-life of cool new productivity software. Evernote, Slack, Notion, Roam, Craft, and many others seem to get pretty far on community excitement for their first few years. After that, I think you have to be left with software that serves a deep and hard-to-replace purpose in people’s lives. Muse got there for a few thousand people, but the economics of prosumer software means that just isn’t enough. You need tens of thousands, hundreds of thousands, to make the cost of development sustainable.
We envisioned Muse as the perfect combination of the freeform elements of a whiteboard, the structured text-heavy style of Notion or Google Docs, and the sense of place you get from a “virtual office” ala group chat. As a way to asynchronously trade ideas and inspiration, sketch out project ideas, and explore possibilities, the multiplayer Muse experience is, in my honest opinion, unparalleled for small creative teams working remotely.
But friction began almost immediately. The team lead or organizer was usually the one bringing Muse to the team, and they were already a fan of its approach. But the other team members are generally a little annoyed to have to learn any new tool, and Muse’s steeper learning curve only made that worse. Those team members would push the problem back to the team lead, treating them as customer support (rather than contacting us directly for help). The team lead often felt like too much of the burden of pushing Muse adoption was on their shoulders. This was in addition to the obvious product gaps, like: no support for the web or Windows; minimal or no integration with other key tools like Notion and Google Docs; and no permissions or support for multiple workspaces. Had we raised $10M back during the cash party of 2020–2021, we could have hired the 15+ person team that would have been necessary to build all of that. But with only seven people (we had added two more people to the team in 2021–2022), it just wasn’t feasible.
We neither focused on a particular vertical (academics, designers, authors...) or a narrow use case (PDF reading/annotation, collaborative whiteboarding, design sketching...). That meant we were always spread pretty thin in terms of feature development, and marketing was difficult even over and above the problem of explaining canvas software and digital thinking tools.
being general-purpose was in its blood from birth. Part of it was maker's hubris: don't we always dream of general-purpose tools that will be everything to everyone? And part of it was that it's truly the case that Muse excels at the ability to combine together so many different related knowledge tasks and media types into a single, minimal, powerful canvas. Not sure what I would do differently here, even with the benefit of hindsight.
Muse built a lot of its reputation on being principled, but we were maybe too cautious to do the mercenary things that help you succeed. A good example here is asking users for ratings; I felt like this was not to user benefit and distracting when the user is trying to use your app. Our App Store rating was on the low side (~3.9 stars) for most of our existence. When we finally added the standard prompt-for-rating dialog, it instantly shot up to ~4.7 stars. This was a small example of being too principled about doing good for the user, and not thinking about what would benefit our business.
Growing the team slowly was a delight. At several previous ventures, I've onboard people in the hiring-is-job-one environment of a growth startup. At Muse, we started with three founders and then hired roughly one person per year. This was absolutely fantastic for being able to really take our time to find the perfect person for the role, and then for that person to have tons of time to onboard and find their footing on the team before anyone new showed up. The resulting team was the best I've ever worked on, with minimal deadweight or emotional baggage.
ultimately your product does have to have some web presence. My biggest regret is not building a simple share-to-web function early on, which could have created some virality and a great deal of utility for users as well.
In terms of development speed, quality of the resulting product, hardware integration, and a million other things: native app development wins.
After decades working in product development, being on the marketing/brand/growth/storytelling side was a huge personal challenge for me. But I feel like I managed to grow into the role and find my own approach (podcasting, demo videos, etc) to create a beacon to attract potential customers to our product.
when it comes time for an individual or a team to sit down and sketch out the beginnings of a new business, a new book, a new piece of art—this almost never happens at a computer. Or if it does, it’s a cobbled-together collection of tools like Google Docs and Zoom which aren’t really made for this critical part of the creative lifecycle.
any given business will find a small number of highly-effective channels, and the rest don't matter. For Heroku, that was attending developer conferences and getting blog posts on Hacker News. For another business it might be YouTube influencer sponsorships and print ads in a niche magazine. So I set about systematically testing many channels.
·adamwiggins.com·
Muse retrospective by Adam Wiggins