marvinvista.com

“...when it comes to making something wonderful, the magic is in the mundane. Greatness is created through the collision of little sparks, ignited by people at the peak of the craft who care a lot and won’t stop working until it gets there. There’s not much talk or strategy, there’s just feeling your way through it—fingers are on instruments the whole time, playing until it works.”

From CEO Karri Saarinen:

What’s the secret of his health? “To work...

I can no longer come to the restaurant every day ... but even at 100, I try to work if possible. I believe the best medicine is to work.”

Just beautiful.

How the Collaboration Started

SA: So I thought we would maybe start off by talking about how this whole collaboration started. Do you want to sort of tell the story?

JI: Well, I was at Apple for nearly 30 years, and I left about six years ago and started to build an interdisciplinary creative team that ranged from industrial designers, architects, graphic designers, user interface designers. And my goal and the goal of the group as we grew was to try to build the most powerful and experienced creative team that we could. The problem that we had was we weren't really quite sure why. It felt slightly like a moral crusade that we were on. But it wasn't until about three years ago when ChatGPT launched that we felt, and this sounds almost corny, but unfortunately it's true, that with the launch of ChatGPT, it felt like our purpose of the last six years became clear. And I reached out to Sam because we were starting to develop some ideas for Interface based on the capability of the technology these guys were developing, technology you guys were developing. And so we first met, yeah, it was about three years ago, wasn't it?

The “Mysterious Middle” of Ideation

SA: So the first time we met and since then, I've had this experience. I've always wanted to ask this question. But I've had this experience of Jony and I will talk, Jony and the team will talk. We'll look at the way that technology is developing and sort of say it feels like there's something important, something is going to change about how people use computers, how people interface with this kind of technology. It feels like as great as phones and computers are, there's something new to do, and we'll talk about the problem space and sort of what the technology allows and what people want to use. And then Jony and his team will go off for a while, and a few weeks later we'll come back together and they will have invented some entirely new different idea that seems obviously right and simple and beautiful and sort of always the thing that should have been the case but impossible to have come up with before. And then we talk about it and we can figure out ways to evolve it and use it. But the thing I would love to understand is how do you do that part in the middle? How do you do the part from exploring the problem in understanding how people interact with each other and how different materials work and how people think about what feels natural and what feels uncanny. You do that in-depth exploration, which I can kind of understand, and then you refine a product, which I can also kind of understand. But this thing in the middle of coming up with a totally new idea, how does that work?

JI: Well, I think, first of all, our motivation and our fuel is really important. And we're very clear the reason we're doing this is we love our species and we want to be useful. And we think that humanity deserves much better than humanity generally is given. And motivation and fuel, I think, are If those aren't quite right, it doesn't matter how hard you work in the subsequent part, you're going to end up somewhere unpleasant, I think. So I think, first of all, the motivation and the fuel is very much centered on people. I mean, as you know, we've always, I pay so much attention in desperately trying to understand the future. I pay a lot of attention to history and to the past. I mean, people are so clever. You know, we didn't just start to be clever. And to me, it's the most absurd arrogance to not carefully study history. So we do lots of research, but in our learning, I think this is one of the big struggles we have, is we just don't... we don't easily accept received wisdom. It's constantly why, you know... So I think it's the tenacity of the questions we ask... I think is very important. But at the end of the day, it's an idea. And, you know, I've talked about this before, even though they can become so powerful, ideas always start in a tentative, quiet, and essentially fragile way. And I think that the team, there's a number of things that we benefit from having. I mean, some of us have worked together for 25 years. And we can... And the trust and the biography we have together, I think, creates an atmosphere that is really supportive of trying to gently explore an idea that you can barely give words to. You know, when you're thinking of something and you have a sense of it, but the next step from what's in your head, you try to give a thought body which tends to be language and how you do that and who you do that with is everything and I think it is accepting your vulnerability because the creative process is so unpredictable but doing the rigorous hard work to try to increase the probability that you might have a good idea and we don't show you any of the terrible ideas

Craft, Care, and Our Relationship with Technology

SA: They're all very impressive. How do you think about craft in this process? I've never seen anyone sweat fractions of a millimeter so early in the design process and think about the subtlest differences in materials and how that's going to change the relationship with a piece of technology or even a thing that particularly... Well, let me ask that first, actually. How do you think about sort of the role of craft in exploring a new idea?

JI: Yeah, I've always thought that it's the mark of someone's, you know, it tells you a lot about someone. If they will care about something that is not seen, how they behave when it's not in public tells you an awful lot about who they are. And I've always thought that, to me, craft and care are almost interchangeable. But that obsession with... I mean, it can be a pleasurable thing to do. Very often, it's inconvenient. But if you only care when it's convenient... I mean, I don't know what that says about you. But I do think that over the years, I've come to believe passionately that we sense when people have cared. It doesn't mean that it's always right. Of course not. But in the same way we sense care, I think you might understand, it's easier almost to understand the concept if I say you sense carelessness. You sense when somebody does not care about you, they care about money and schedule. And you sense that, I think, a mile away. And so I don't know. I think that practice and the rigor around caring, whether it's convenient or not, whether it's going to be manifest in a way that's seen or not, that doesn't matter. The practice is an extension of our motivation, which is why I always come back to what is motivating, what's driving us. What is motivating and driving you? Well, one, I think it's the opportunity of... I've never in my career come across anything vaguely like the affordance, like the capability that we're now starting to sense. And that's, at one level, that could be it. But to me, that's part of the story, part of the drive. I think the other huge part of it is I don't think we have an easy relationship with our technology at the moment. And rather than see AI as an extension of those challenges, I see it very differently. I see it as a chance to use this most remarkable capability to full-on address a lot of the overwhelm and despair that people feel right now, and some of that is connected to the tools that they are using.

Why Build Something New Now?

SA: Speaking of that, a thing that you said to me very early on when we started talking about this that really stuck with me is, you know, the default answer, the most likely answer is we shouldn't do anything at all. The current devices are great. It's very, the general purposeness of a phone is an incredible thing, which I strongly agree with. And it's, there's got to be a really compelling reason for something new. What convinced you eventually that there was something here worth

JI: doing? Well, I think, I mean, as Sam said, I think what was amazing about multi-touch was that it supported a user interface that was not compromised by app A over app B over C. So it was the most remarkable interface that meant you could have a general purpose device, whether it was a phone, whether it was a tablet. And I think that general purposeness combined with its connectivity and its physical size was incredibly liberating. But that was in 2007. And I think you could frame the question the other way, which is, I think it would be absurd to assume that you could have technology that is just breathtaking delivered to us through legacy products, products that are decades old. And as I said, I mean, this is the most compelling, most exciting technology and capability I've ever experienced. And the nature of how we connect to it, the character of how we connect to this capability is going to be so important.

What Success Looks and Feels Like (Joy Included)

SA: And how will you know if we've gotten that right? Like, what do you hope people will say about it when they use this family of devices? What do you hope will feel different than previous generations of decades-old technology? And how do you think about what it means to build this in a way we can be really proud of?

JI: That's a great question. Well, I've always loved, you know, if you're trying to solve a complicated problem as a designer, I sort of think that's my job and the team's job. And I don't want to, like, be wagging my tail in your face, saying, oh, look what we solved. It was very hard, but look what we did. I think that it should be, you know, if the solution is clever, it should... I mean, this has been said many times, but that it should just work. It should seem inevitable. It should seem obvious, as if there wasn't possibly another rational solution to the problem. So I think a sense of, oh, well, of course, I've always found that, I'm always nervous of that because I wish there would be some, you know, intrinsic fanfare and trumpeting that I'll look at. But my experience consistently suggests to me that you'll look at something and think, well, yeah, of course you'd do it that way. Why did it take so long? I think it's partly that, but I also think, and we've talked about this a bit before, I do think having, you know, I've lived in San Francisco since 92, and I've seen so many changes in the Valley, and I think one of the things I miss is just, humor's probably overstating it a bit but I am just a little crushed by how seriously we will take ourselves now this happens to be um this is serious stuff um I mean truly and the ramifications and consequences of not caring and not being careful are truly horrendous but in terms of the devices we design, in terms of the interfaces we design, if we can't smile, honestly, if it's just another deeply serious sort of exclusive thing, I think that would do us all a huge

SA: disservice. That's been a thing I found very wonderful about working with you, is you can take the most serious things and bring some whimsy and joy to them in a way that makes the whole thing

JI: approachable and feel very different. I mean, to me, that's the, you know, one of the first things Sam said to me when we very first met, and when I meet someone, I don't take notes. I mean, it's a bit creepy, isn't it? And I was taking notes, and one of the things he said was, you know, this is computers are starting to think. And we went down a path of, you know, what we were describing clearly was an extraordinary level of intelligence. And of course, my references are going to be what my interface is to other members of our species. And when that is humorless, it's also not so efficient. because it's not enjoyable and you tend to do stuff with a bit more vim if you're enjoying it, I think.

Unexpected Challenges & the Need to Focus

SA: Any unexpected challenges so far?

JI: Yes. I mean, I think I was aware that the rate of change and the rate of progress, I've never known anything ever like that. And I do wish I was about 20 years younger. But I've spent a huge amount of my time as a designer trying to create momentum. You know, it's part of my job, I think, is to draw on my experience and try to figure out how we can create a momentum to help generate ideas. Again, this is a first for me, where my challenge is the momentum is so extraordinary. The technology momentum, the momentum of the OpenAI team, the momentum of LoveFrom and IO is almost overwhelming. and has led us to generate 15, 20 really compelling product ideas. And the challenge is trying to focus. And I used to be good at that, and I've lost some confidence because the choices are... It'll be easy if you really knew there were three good ones and everything else, you know, it's just not like that. And so, you know, obviously we're designing a family of products. We're not doing... But it's really trying to make sure we choose, you know, we're judicious and thoughtful in what we focus on and to then not be distracted. But I've never known that this sort of momentum and these guys they have so many distractions because it's just so from one week to the next there's something else and then something else so it is it's you know to try to be a little blinkered is I think is going to be necessary so that we can finish the first thing so we can get

Advice for Builders in a Vertigo‑Inducing Moment

SA: to the second to the third and to the fourth etc. Speaking of that we've got a room full of builders that I assume are facing the same sort of vertigo-inducing challenge. Do you have advice for people who are trying to build in a world like this?

JI: Well, one thing that really does strike me is, I don't know if you feel this way, but I feel there's something very equalizing about what's going on, because it's new to all You know, if anybody says, oh, I'm extraordinarily confident, I've got decades of experience with this, obviously that's a silly thing to say. So one thing I love is that there is something egalitarian and very somehow oddly inclusive about what's going on. And one of the things I've loved is I think success is going to be, you have to be curious and inquisitive and desperate to learn. And I think these are such healthy ways of being personally, but I know that they're critical ways of innovating. I think the advice, I mean, I can only speak, you know, to what I'm personally wrestling with, is I find myself every day, I think, having to challenge myself and think, where is my experience relevant? Where's my experience of sort of 30, 40 years of doing what I've been doing useful? But also be really clear where it's just, you know, is not relevant, in fact, can be an obstacle or an impediment. And so I do think that there's tremendous sort of solace and encouragement to think we're all in this, pretty much sort of level set, and to be really curious about why you have the beliefs that you have. And my sense is dogma, and I'm not talking about values, so separate from values, but sort of dogma associated with creating and engineering would seem to me out of place. I think it's always out of place, but I think it would be highly unsuccessful here.

Rethinking OS & UI in an AI World

SA: Along those lines, one thing you said to me a long time ago, but that has really stuck with me, is you're not even sure that things like an operating system or a UI in the way we traditionally think of one, what those things will mean in the future. So any new thoughts on where that's all going to go? But more generally, I agree it is kind of wonderful we're all novices in this together, but how do you think for this very different Brave New World?

JI: Well, I think we've spent as a team in our research, And some of the research we did was fairly... Well, it was extraordinarily broad and some of it really quite deep. But in terms of the nature of our... You know, I was reading something the other day about how important the character of our connection to an activity is. And what might be seen as a tedious, repetitive activity like gardening for example if the nature of your connection to that that activity is such it can be the most therapeutic um edifying um and I'm I love gardening so this I mean this is a great this is a I think this has really resonated with me and so I it's really we've gone very deep on trying to understand our interface with sort of person-to-person interface, how incredibly sophisticated and nuanced that is, but how very often we, you know, we're so entwined, the way we think is so entwined, you know, with other devices and tools, so how we often reference in our interface together when we're trying to connect, the role of other tools in that. And then there are just those very task-specific, you know, using scissors, you know, something so beautifully simple and focused and appliance-like. And so I think the clues and the pointers all exist, and it's just trying to sort of put them together, but it will be ideas. It will be a vision for what makes sense. And I think we're only going to arrive at that if we are very curious and light on our feet. And we both know humility is critical to be open to realizing, oh, this isn't quite the right direction we need to change.

Will AI Change Design Tools and Workflows?

SA: One thing that I've observed is that AI has really changed what it means to write code in a huge way. This year, the way people wrote code a year ago and the way people write code now are pretty different, surprisingly different for one year. Is that happening in the design process, or do you expect that to happen in the design process? Will these tools really change it?

JI: Yeah, it's funny. I think I want it to be. I really do. I think, I don't know if you guys feel the same way, but what we've struggled with as a team is we're so consumed with the work, is trying to make time to sort of incorporate and seek the help of new tools. I mean, one of the things that we've found, I've always found this generally an affirmation that what we're doing is indeed new when your tools are just irrelevant. And there's a sort of a tenacity that I sense in the design team where the things we want to explore, right now we're struggling to explore. But that's fantastic. I mean, so it means, and this has happened a number of times, is that we have to build the tools ourselves. But this is an area where I personally feel a bit guilty and certainly tenacious in wanting to be further alone.

Values and Hopes for the AI Era

SA: Last question. We talk a lot about the tools that people deserve and the sort of values that we want to express with our work. AI clearly will change the world in all sorts of ways, some really huge. What do you most hope that we get right as we go through this technological revolution, what do you most hope we discard or preserve or put out? I think that it would

JI: be that these tools, I mean, this is almost too naive-sounding to say it, but that they would make us happy and fulfilled and more peaceful and less anxious and less disconnected. I know I should care about productivity, and I do, but I just think that for the last 20 years, and you know the data, you know... I mean, when I said we have an uncomfortable relationship with our technology, I mean, that's the most obscene understatement. And I see that we have a chance to... And every bone in my body believes this. we have a chance to not sort of just redress that, but absolutely change the situation that we find ourselves in. That we don't accept this has to be the norm. That we don't accept that we can just, we have to extrapolate where we've been to get a sense of where we're going. That we just reject that. And I really believe that we can. I really do.

OpenAI is deliberately stacking powers: funding and coordinating the physical layer, converting that into a trusted, best-answer experience, and binding it together with identity, memory, and apps—run by a product machine.

…one of the biggest differences I've seen between people that really know how to build agentic workflows compared to people that are less effective at it is the ability to drive a disciplined development process, specifically one focused on evals and error analysis.

If you're going to really exploit the capabilities of these new tools, you need to be operating at the top of your game. You're not just responsible for writing the code - you're researching approaches, deciding on high-level architecture, writing specifications, defining success criteria, designing agentic loops, planning QA, managing a growing army of weird digital interns who will absolutely cheat if you give them a chance, and spending so much time on code review.

The unit economics of SaaS create a predictable organizational structure: ARR compounds through new bookings and net retention, so companies optimize ruthlessly around those metrics. This means the roadmap becomes a sales tool—every feature maps to a deal or a churn save. The best orgs maintain product integrity while doing this. The median org ships the top 10 feature requests from their biggest customers and calls it a strategy. From Ben Thompson's excellent interview with Bret Taylor:

Broadly speaking, sales and marketing matter a lot in enterprise software companies and tend to be the gravitational center... Because at the end of the day, annual recurring revenue is an annuity, it throws off cash every year, you basically need to add to that annuity by signing new recurring revenue contracts or adding to them, and you subtract attrition. That's just how software as a service works.

With that, when you stop growing, the business model breaks down. That's where things like the Rule of 40 come from, where your growth rate and your EBITDA margin sort of need to be in lockstep. You can grow very profitable with that business, when you stop growing, no one wants to work at one of those companies.

As a consequence, you end up with a really customer-centric, go-to-market-centric orientation, and product serves that. The best enterprise software companies, that voice of the customer dictates their product roadmap, and they can really meet that demand.

The worst ones stop innovating on the product and hold your feet over the fire in the sales process. Depending your level of cynicism, all enterprise software companies have been guilty of each at different points.

Another interesting thing about this is how the business model shapes culture. When you're growing 80%, product can lead and sales follows. When you're at 15%, sales leads and product follows. It's hard—the companies that stay interesting are the ones that maintain product velocity even as they scale GTM.

Living off borrowed time, the clock tick faster.

Anthropic is riffing on Madvillainy and I'm here for it.

• You learn more per unit time because you make contact with reality more frequently
• Going fast makes you focus on what's important; there's no time for bullshit
• "Slow is fake"
• A week is 2% of the year
• Time is the denominator

Three major AI video launches in three weeks, and each company picked a radically different target audience:

...compelling AI video generation is very much here, and it's widespread: over the last three weeks we have actual AI video products from Google, Meta, and now OpenAI.

...it feels like each company has an entirely different target audience: YouTube is making tools for creators, Meta is building the ultimate lean back dream-like experience, and OpenAI is making an app that is, in my estimation, the easiest for normal people to use.

In this new competition, I prefer the Meta experience, by a significant margin, and the reason why goes back to one of the oldest axioms in technology: the 90/9/1 rule.

• 90% of users consume
• 9% of users edit/distribute
• 1% of users create

If you were to categorize the target market of these three AI video entrants, you might say that YouTube is focused on the 1% of creators; OpenAI is focused on the 9% of editors/distributors; Meta is focused on the 90% of users who consume.

...it's striking how this target market evaluation tracks with the companies themselves: YouTube has always prioritized creators, while OpenAI's business model is predicated on people actively using AI; it's Meta that has stayed focused on the silent majority that simply consumes...

I used to think network effects would win out, and we'd end up with just one or two stables, but I don't believe that anymore. Cross chain swaps are getting more efficient by the day, as is intra-chain inter-stablecoin swapping. In a year or two, I think many intermediaries in crypto will just show your deposits as generic "dollars" or "dollar tokens" (rather than USDC or USDT) and they'll guarantee you redeemability in a stablecoin of your choice.

Already we are seeing this with a lot of the fintechs and neobanks that are product rather than crypto first. They care most about UX and would rather offer clients the best experience instead of honoring crypto traditions. So they just show your balance in USD and manage the reserves on the back end.

OpenAI partners with Stripe to launch instant checkout within ChatGPT.

It has been clear for a long time now that ChatGPT is a powerful tool for commerce, and the company has been taking baby steps to close the loop, with things like product carousels, an in-app browser (which lets OpenAI capture web usage and attribution), and the overall concept of having a router deciding how to answer every query. This, however, is an explicit step into the transaction itself — and it's almost certainly not the last.

…Meta can still build a good product. Vibes may not end up being for everyone — I hesitate to generalize too broadly from what I like, given my rather esoteric media consumption habits — but I think it is executed very well, with a strong and unique point of view. That sort of product sensibility is something that has generally been missing in AI.

..Meta has the courage and conviction to lean into what AI can uniquely do. There are a lot of things to wring one's hands about when it comes to Vibes, but Meta is leaving that hand-wringing to the peanut gallery, and actually shipping. There's something to be said for that.

2026 will be a pivotal year for the widespread integration of AI into the economy:

• Models will be able to autonomously work for full days (8 working hours) by mid-2026.
• At least one model will match the performance of human experts across many industries before the end of 2026.
• By the end of 2027, models will frequently outperform experts on many tasks.

“Scaling laws are real, ofc. ++ model power keeps raising the ceiling. But capability ≠ adoption. The moat in workflow AI is bottling institutional context and making it useful in the moment of work.”

…

“Pretrained models are powerful, but not turnkey. Extracting Quinn’s win rate, UCSF’s pathways, or Zappos’s ethos is closer to RL than pretraining. You shape behavior with constraints and feedback until it fits the environment. Workflow products do that shaping.”

…

“The Bitter Lesson is obvi right: scaling keeps raising the baseline. But raw capability isn’t adoption, because no user wants to give superintelligence full context all the time. App winners are the ones who bottle context into systems trusted at the point of action.”

Early GDPval results show that models can already take on some repetitive, well-specified tasks faster and at lower cost than experts. However, most jobs are more than just a collection of tasks that can be written down. GDPval highlights where AI can handle routine tasks so people can spend more time on the creative, judgment-heavy parts of work. When AI complements workers in this way it can translate into significant economic growth.

“In AI, we talk about ‘product overhang’, and this is what we discovered with the prototype. Product overhang means that a model is able to do a specific thing, but the product that the AI runs in isn’t built in a way that captures this capability. What I discovered about Claude exploring the filesystem was pure product overhang. The model could already do this, but there wasn’t a product built around this capability!”

…

“Building and testing 5-10 prototype ideas in a day is possible with AI agents. Prototyping used to be so time consuming that if there were two days to prototype, it was lucky to have two distinct prototypes built by the end. But now, agents can build prototypes very quickly, so tests of 5-10 prototypes per day are easily done, as the Claude Code team did.”

…

“Claude generates Markdown which is rendered in the terminal. Before the release, Anthropic engineers complained that nested lists don’t really look that good, and that the spacing between paragraphs was off. The problem was with the markdown renderer. Boris tried all the existing ones, but none looked good inside the terminal. The day before public release, Boris used Claude Code to vibe code a Markdown parser and renderer. And the result was better than anything preceding it, so they shipped it! Boris says he thinks no other terminal has the same Markdown rendering.”

I can’t overstate what a massive opportunity this is: every item in every YouTube video is well on its way to being a monetizable surface. Yes, that may sound dystopian when I put it so baldly, but if you think about it you can see the benefits; I’ve been watching a lot of home improvement videos lately, and it sure would be useful to be able to not just identify but helpfully have a link to buy a lot of the equipment I see, much of which is basically in the background because it’s not the point of the video. It won’t be long until YouTube has that inventory, which it could surface with an affiliate fee link, or make biddable for companies who want to reach primed customers. More generally, you can actually envision Google pulling this off: the company may have gotten off to a horrible start in the chatbot era, but the company has pulled itself together and is increasingly bringing its model and infrastructure leadership to bear, even as Meta has had to completely overhaul their AI approach after hitting a wall. I’m sure CEO Mark Zuckerberg will figure it out, but Google — surprise! — is the company actually shipping.

“…the company successfully navigated one paradigm shift, and is doing much better than I originally expected with this one. Larry Page and Sergey Brin famously weren’t particularly interested in business or in running a company; they just wanted to do cool things with computers in a college-like environment like they had at Stanford. That the company, nearly thirty years later, is still doing cool things with computers in a college-like environment may be maddening to analysts like me who want clarity and efficiency; it also may be the key to not just surviving but winning across multiple paradigms.”

Our whole thesis is that for every company in the world, their AI agent will be their most important digital interface they have, as important as their website or their mobile app. Just like mobile apps didn’t replace websites, and it also differs but not every retailer even has a mobile app. Some do, some don’t. You’ve covered this a lot, just like do you have the brand equity to warrant that?

I think across the board, your AI agent, whether it’s published in your mobile app or available via WhatsApp, or available over CarPlay or whatever it might be, that will be the primary digital interface to your company. We’re really focused on customer experience more broadly. What’s exciting for us, if we do our job, we’re the most important AI agent you make, we’re the one that actually powers your customer experience.

“Upon opening a project or folder, you’re likely to jump into editing files. This means Cursor needs to generate autocomplete suggestions, which the Cursor team calls tab suggestions. A low-latency sync engine powers the ‘tab model’. This generates suggestions that are greyed out and can be accepted by hitting the ‘Tab’ key. The suggestions need to be generated quickly in ideally less than a second. This ‘tab model’ must be as fast as possible, and data transfer as low as possible. There’s always a tradeoff between how much context to send, and the quality of the suggestions: the more relevant context Cursor can send, the better the suggestions. However, sending lots of context can slow down the display of suggestions, so getting this right is one challenge for Cursor’s engineers.”

I had this goal, which was we needed to somehow create a more imaginative world. I mean, one of the biggest risks in the world I think is a collapse in belief, a belief in ourselves, a belief in the future. And part of that I think comes from a lack of imagination, a lack of imagination of what we can be, lack of imagination of what the future can be. And so this imagination thing I think is an important pillar of something that we need in the world. And I was thinking about this and I saw this, I'm like, "I can turn this into a force that can expand the imagination of the human species." It was what we put on our company thing now. And that felt realistic. So that was really exciting.