The Minus One podcast explores pivotal decision points in the lives of interesting people, focusing on the "minus one" moment - the time before launching something new when you have to decide what to work on next. Host Aditya Agrawal, partner at South Park Commons, introduces guest Mike Krieger, who co-founded Instagram in 2010 with Kevin Systrom, grew the business through its Facebook acquisition, later founded Artifact in 2021, and recently joined Anthropic as Chief Product Officer.

"Welcome to the Minus One podcast where we ask the most interesting people in the world about times in their lives that they've had to answer a question we all eventually struggle with: what's next? Because before you launch at zero, you have to decide what to work on at minus one."

Timestamp: [0:00-1:04]

Mike grew up in São Paulo, Brazil, where his early interest in computers began when his father brought home an IBM computer. His formative computing experience came from playing a simple game featuring gorillas throwing bananas at each other (an early version of Angry Birds). The pivotal moment occurred when he discovered he could type "edit" to see the game's source code in BASIC, sparking his interest in building software.

"What I remember is at some point I figured out that you could type 'edit' and it was just like it was written in QBASIC or something, you could see the source code. I had no idea what that meant, but I was like 'oh, like that's what makes this thing happen.' So like early seed that like maybe building something was going to be interesting to me."

This early curiosity blossomed in high school when Mike became deeply interested in web development, exploring tools like Dreamweaver that allowed him to see and edit web page source code. Despite his growing passion, he noted that he had "no role models in my life saying 'oh yeah, like you can go build companies or build software,'" seeing coding as merely "a cool thing I do on the weekends."

Timestamp: [1:05-2:42]

Mike's earliest web creations reflect his dual interests in creative expression and technical exploration. His first significant website was a fan site dedicated to the band Weezer, featuring song tabs, band news, and participation in a "Weezer web ring." His online identity was even tied to this passion, using the screen name "Mike Weezer."

"I was really, really into Weezer, band. So I had like a whole Weezer fansite, Weezer tabs, you know, tabbed out the songs, you know, news about Weezer. I was in a Weezer web ring, do you remember web rings?"

His other early web project demonstrated more academic applications of his skills - a Brazilian history project for school utilizing image map technology, where clicking on different states of Brazil would reveal their specific histories. This project showcased Mike's early exploration of interactive UI elements, something he now reflects on as being "Web 2.0 before there was Web 2.0."

Aditya noted the universal impulse to use new technologies for self-expression, sharing that his first website was "geocities.com/metaris496" where he similarly showcased his favorite music and movies. This conversation highlights how fundamental the desire for digital self-expression was in the early internet era.

Timestamp: [2:43-4:26]

Mike and Aditya shared a nostalgic conversation about early web technologies that were foundational to their generation of builders. Beyond creating personal websites on platforms like Geocities and Tripod, they reminisced about features that defined the early internet experience.

"There's that desire to express and then the desire to connect, and at that time it looked like you know, you could put some probably pretty sketchy HTML code and it would create a forum on your, you know, at least a comment section on your website along with your visit counter, very important as well."

The conversation highlighted several iconic elements of early web culture:

• Comment sections created with basic HTML
• Visit counters that proudly displayed site traffic
• Tripod's "neighborhood" feature that grouped sites by interest
• Web rings that connected related sites in a navigable loop
• Forum integrations through simple code snippets

These early technologies, while primitive by today's standards, represented the first attempts at creating online communities and interactive experiences, laying groundwork for the social web that would later emerge.

Timestamp: [4:27-4:56]

Mike's decision to pursue higher education in the United States rather than Brazil was driven by his uncertainty about his career path. In Brazil, students declare their degree when applying and are locked into that track, but Mike was torn between journalism, programming, and music. Coming to the U.S. gave him more flexibility to explore his interests.

"I decided to come to school in the states and that was primarily driven by not really knowing exactly what I wanted to do. When I was 18, in Brazil, you declare your degree as you apply and then you're locked in. So if you change your mind, you're starting all over. I wasn't sure yet... I thought I wanted to be a journalist, but I had this programming stuff that I liked. I was still into music. I'm like, I can't decide right now."

At Stanford, Mike chose to study Symbolic Systems, an interdisciplinary program that combined computer science, philosophy, artificial intelligence, cognitive science, and linguistics. He loved the breadth of the program, which allowed him to explore many different interests while still focusing on technology.

Mike noted that several notable tech figures also studied Symbolic Systems, including Marissa Mayer and Reid Hoffman, and that the program tends to produce people who are both creative and highly technical - a rare combination. He reflected that this interdisciplinary background was particularly valuable for his future work:

"If I had done pure CS, I would have gotten very good at the how, but less good at the what and the why."

His academic advisor encouraged him to challenge himself with difficult CS classes to ensure he would graduate with strong technical skills alongside his interdisciplinary knowledge - building what he called the "verticality" of his "T-shaped" skillset.

Timestamp: [4:57-7:51]

After graduating from Stanford, Mike faced additional constraints as an immigrant that shaped his career planning. Unlike some of his peers who could immediately start companies, he needed to navigate visa requirements, which influenced his approach to entrepreneurship.

"Being an immigrant, there's some constraints you have around building and how entrepreneurial you're going to get. I had friends in college like 'oh I'm going to leave and start a company.' I'm like, 'I need to extend my F1 OPT to be able to get a job and then eventually get an H1' - like that whole dance."

Mike developed a strategic 10-year plan: first, join a medium-sized company (around 40 people) that would sponsor his visa; then join an early-stage startup (first 8 employees); and finally start his own company. This methodical approach reflected both immigration realities and his desire to learn the full spectrum of company-building.

His first job was at Meebo, an online chat startup with about 40 employees that was doing innovative work in asynchronous networking and browser-based UIs. At Meebo, Mike gained valuable experience:

1. Learning cutting-edge technology that was "ahead of the curve"
2. Observing founders manage boards, fundraising, and company pivots
3. Working with a complementary founding team where each member brought different strengths

"I got to see founders do the real work of running the company, which is not just like 'oh it's fun to found it.' It's like 'we're managing our board, we're managing fundraising, we're bringing the company along, we're figuring out when to pivot.'"

Even while working at Meebo, Mike continued experimenting, spending weekends prototyping projects, particularly focused on the iPhone and the early App Store. He specifically mentioned working on augmented reality crime mapping, showing his continued passion for side projects.

The dynamics of Meebo's founding team - with distinct but complementary areas of expertise - would later influence his successful partnership with Kevin Systrom at Instagram.

Timestamp: [7:52-9:54]

• Early exposure to computer source code sparked Mike's interest in building software, demonstrating how childhood curiosity can shape career trajectories
• The desire for self-expression naturally leads to creation when given access to new technologies
• Studying interdisciplinary fields (Symbolic Systems) provided Mike with a unique combination of technical skills and creative thinking
• Educational flexibility in the U.S. allowed Mike to explore multiple interests before committing to a path
• Immigration status significantly impacts entrepreneurial journeys, requiring strategic career planning around visa requirements
• Working at established startups before founding your own provides crucial learning opportunities about company operations
• Weekend side projects allow for continued exploration and skill development even while employed full-time
• Complementary skill sets and distinct areas of expertise formed the foundation for Mike's successful founding partnerships
• T-shaped skill development (breadth with depth in key areas) is crucial for product builders

Timestamp: [0:00-9:54]

People:

• Kevin Systrom - Co-founder of Instagram alongside Mike
• Marissa Mayer - Mentioned as another Symbolic Systems graduate
• Reid Hoffman - Mentioned as another Symbolic Systems graduate
• Chris Cox - Mentioned as potentially another Symbolic Systems graduate
• Seth - Mentioned as one of the founders at Meebo
• Elaine - Product work colleague at Meebo
• Sandy - Backend colleague at Meebo

Companies/Products:

• Instagram - Social media platform co-founded by Mike in 2010
• Facebook - Company that acquired Instagram
• Artifact - Company founded by Mike in 2021
• Anthropic - AI company where Mike is now Chief Product Officer
• South Park Commons - Aditya's investment firm
• Meebo - Online chat startup where Mike worked after college
• Geocities - Early web hosting service mentioned by Aditya
• Tripod - Early web hosting platform that had "neighborhoods"

Technologies/Concepts:

• QBASIC - Programming language Mike first encountered
• Web rings - Early internet navigation method connecting related sites
• Image maps - Technology Mike used for his Brazil history project
• Dreamweaver - Web development tool mentioned
• F1 OPT/H1 visas - Immigration pathways mentioned as constraints
• Symbolic Systems - Interdisciplinary degree program at Stanford
• T-shaped people - Concept of having both breadth and depth of skills

Entertainment:

• Weezer - Band that Mike was a fan of and created a website about
• Gorillas game - Early computer game Mike played (compared to Angry Birds)

Timestamp: [0:00-9:54]

While working at Meebo, Mike continually explored his interests through weekend projects. One of his creations was an augmented reality crime tracker app that allowed users to visualize crime data in their surroundings - a somewhat alarming but innovative use of the emerging AR technology and San Francisco's open city data.

"It was a terrifying app actually, you would like swing your camera around and it was like 'Ah! [crime] sitting 30 feet away.'"

Another significant side project was a Polaroid-style photo app that featured an animation where the photo would "shake" to reveal the image - an early precursor to what would eventually evolve into Instagram. These weekend explorations, while seemingly disconnected at the time, created the foundation for future opportunities.

"I think this will connect to some themes that we'll talk about today, but you don't know that that weekend side project you did is going to connect to something later, but it can and often it will. Just being out there building creates that opportunistic moment."

Mike emphasizes that consistent creation - even without knowing where it might lead - opens doors to unexpected connections and opportunities. This philosophy of continuous building became fundamental to his approach to innovation.

Timestamp: [10:00-10:41]

Mike's pivotal "minus one" moment happened at Coffee Bar (now Ernest restaurant) in San Francisco, where he was spending his weekends working on side projects. There, he repeatedly encountered Kevin Systrom, whom he knew slightly from Stanford. Kevin was in his own "minus one" phase, having left his job to prototype a location check-in platform called Bourbon that combined photos, games, and leaderboards.

"It was a classic meet cute. Like, 'Oh, what are you working on?' 'Oh, you're working on side projects too? Cool, you're doing a company? What text editor you use? I use TextMate.' 'Oh, I'm using like Sublime.' You know, and we got talking over that."

Their initial connection formed over shared interests in technology and building products. Mike was drawn to the problem space Kevin was exploring, and shortly afterward, Kevin decided to raise funding and invited Mike to join as co-founder. The investors were specifically looking for a technical co-founder to complement Kevin's skills.

However, Mike's immigration status created immediate challenges. They had to prove to the government that Instagram, then just a concept with one person and $500,000 in seed funding, was a "real company." Humorously, they were asked to provide a business plan, which they hadn't created before raising their angel round based solely on a prototype and demo.

Despite these hurdles, Mike was confident in his decision to skip his original plan's middle step (joining a small company) and instead leap directly into founding a company with Kevin.

Timestamp: [10:42-12:19]

Aditya highlighted a core principle of South Park Commons: it's difficult to effectively navigate the "minus one" phase without actively creating something. He emphasized that while the medium might vary - writing code, pushing pixels, or creating Figma prototypes - having a canvas for self-expression beyond just PowerPoint is crucial during this exploratory phase.

"It's very hard to do minus one unless you're actually just producing something now. My kind of tool of the trade is like writing code, but either you're pushing pixels, you're writing code, maybe nowadays you're doing like a Figma prototype, but I think it's really important in that stage to have a canvas to express yourself."

Mike strongly agreed with this philosophy, explaining that without creating tangible products, all you have are unvalidated hypotheses:

"If you're not creating things, you have all hypotheses and no actual contact with what's even doable from a building perspective, what's even possible from getting other people to use the thing if it's anything social, anything that requires adoption."

He emphasized that validation isn't necessarily about immediately determining business viability, but rather about deciding whether there's energy worth continuing to invest in a particular direction. For Mike, having "actual contact with the world" through creating and sharing something real has always been vital to his process.

The conversation underscored how building something tangible provides crucial feedback that theoretical exploration alone cannot - it tests technical feasibility, user interest, and most importantly, whether the creator feels sustained passion for the direction.

Timestamp: [12:20-13:56]

Aditya observed that Instagram stood out in its early days for being noticeably faster and more responsive than other mobile apps of the time, creating a significantly better user experience. He suggested this advantage likely came from Mike and Kevin's hands-on building process, which revealed optimization opportunities that theoretical planning might have missed.

Mike confirmed this assessment, explaining that their performance optimization strategy was grounded in real-world constraints rather than hypothetical scenarios:

"You could do hypothetical performance work, or you could build for a real use case. For me, it was I had the iPhone 4, which at the time was pretty... I mean I think that's still near peak industrial design for the iPhone. I'm glad that basically the new iPhones just look like that one. But Kevin had the 3GS, which had its charm, but his was extremely slow for some reason... and it was great because we did all our performance work for that device. So if it worked fast enough on that one, it was going to be great on the other ones."

This practical approach extended to network performance as well. Living in San Francisco in 2009 meant dealing with AT&T's notoriously poor coverage, especially in areas like Dolores Park on weekends where signal was virtually nonexistent. This real-world limitation pushed them to implement robust networking features, error recovery systems, and offline capabilities that many competitors overlooked.

"All the things we did around network performance, error recovery... if you were in a beautiful room with great Wi-Fi, you wouldn't have seen that. But we were out there every day testing the thing and then iterating really, really quickly."

Their experience highlights how building and testing in challenging real-world conditions - rather than ideal environments - led to a more resilient, responsive product that performed well across various devices and network situations.

Timestamp: [13:57-15:48]

Aditya noted that both Mike and Kevin had independently explored and discarded numerous projects before settling on what would become Instagram. Kevin had Bourbon and other ventures, while Mike had been working on various side projects. He asked whether abandoning these ideas caused feelings of loss and regret, or if it was actually easier to let go of ideas in their early stages.

Mike reflected on how he approached different types of projects with different expectations:

"I realized that there were projects that I was doing where the goal was traction, and that didn't make sense without traction. And ones that were more... I think Robin Sloan talks about apps as a home-cooked meal. Sometimes you just want to cook, and you're not opening a restaurant, you're just like, 'I want to express some idea through this thing' and then be like, 'Great, that was an idea, got it out there, people enjoyed it.'"

He emphasized that not every project needs to aim for massive scale or venture funding - some are simply expressions of ideas or learning experiences with natural endpoints. Mike cited a recent viral example of someone who created a website consisting solely of checkboxes that could be toggled on and off. After experiencing a surge of viral attention and learning about scaling challenges, the creator simply shut it down after two weeks, having accomplished what they wanted.

"That thing did not need to be a venture-backed or anything-backed business."

When reflecting on his own abandoned projects, Mike occasionally revisits his GitHub repositories to see old creations, such as a visualization tool for displaying tweets at conferences. While these projects had their moment and then faded, each contributed valuable learning experiences about networking, APIs, and other technical aspects that later proved useful.

Both Mike and Aditya shared similar experiences of building weekend projects (like a Chrome bookmarks manager in Aditya's case) that fulfilled their purpose as learning experiences or personal tools without needing to evolve into businesses.

Timestamp: [15:49-18:51]

The conversation shifted to how AI tools are changing the development process. Mike observed that modern AI tooling allows developers to create functional prototypes much faster, enabling quicker validation of ideas and providing that satisfying feeling of building something substantial with less effort.

"With a lot of AI tooling now, you can kind of get to something feeling real faster, which on one hand I think will let you validate those things much faster and just feel like you did something of some substance as well. I think that is almost entirely good."

However, Mike also expressed an interesting counterpoint about the psychological value of struggling through technical challenges:

"I do wonder, in the same way that people feel psychologically more affectioned for furniture that they've assembled from IKEA than stuff that you... Like, there is a grind where you're like, 'And then my Python packages were conflicting, and then I went... like Yak shaving and I did my two hours of that.' I don't know, maybe I'm nostalgic for bad developer experiences too."

Aditya agreed that even with AI assistance, developers will still encounter fundamental challenges - the tools might just shift where the difficulty lies rather than eliminating it entirely:

"That recursive nature of finding the hard part of the problem, even at least with the AI stuff today, it still exists. You'll probably just have to figure out where that long pole is."

He compared the situation to how JavaScript and Python developers might be viewed by those who write lower-level code in C, suggesting there's always a spectrum of abstraction in development tools, with AI representing a new layer that still preserves core problem-solving challenges.

The discussion highlights an interesting tension in development: tools that make creation easier and faster are generally positive for innovation, but there may be something valuable in the struggle and deep engagement that comes from working through technical challenges manually.

Timestamp: [18:52-19:59]

• Weekend side projects often contain seeds that can unexpectedly connect to future opportunities
• Serendipitous meetings in the right environments (like Coffee Bar in SF) can lead to transformative partnerships
• Immigration status creates unique challenges for founders that require creative navigation
• Building actual products is essential for validating ideas beyond theoretical hypotheses
• Testing in real-world conditions (like slow devices and poor networks) leads to more robust products
• Not every project needs to scale or become a business - some are valuable as learning experiences or creative expressions
• Learning when to abandon projects is an important skill in the innovation process
• Modern AI tools accelerate prototyping but may change the psychological relationship with the creation process
• The struggle with technical challenges may create deeper affection for and understanding of the final product
• Physical gathering spaces like coffee shops play an important role in fostering connections between creators

Timestamp: [10:00-19:59]

People:

• Kevin Systrom - Instagram co-founder who Mike met at Coffee Bar while working on Bourbon
• Robin Sloan - Mentioned for his concept of "apps as home-cooked meals"

Places:

• Coffee Bar - San Francisco coffee shop where Mike and Kevin met, now Ernest restaurant
• Ernest - Fancy restaurant that replaced Coffee Bar in San Francisco
• Dolores Park - Mentioned for having poor AT&T coverage in San Francisco

Companies/Products:

• Bourbon - Kevin Systrom's pre-Instagram project combining location check-ins, photos, games, and leaderboards
• Instagram - The photo-sharing platform that evolved from Mike and Kevin's collaboration
• South Park Commons (SPC) - Aditya's organization focused on the "minus one" phase
• AT&T - Mobile carrier noted for poor coverage in San Francisco in 2009
• IKEA - Referenced in analogy about valuing things we've struggled to build

Technologies/Devices:

• iPhone 4 - Mike's phone when building Instagram, noted for its industrial design
• iPhone 3GS - Kevin's slower phone that became their performance testing benchmark
• HTML5 - Technology that made Bourbon easy to build and iterate on
• TextMate - Code editor Mike mentioned using
• Sublime - Code editor Kevin mentioned using
• GitHub - Platform where Mike stored his past projects
• Chrome bookmarks manager - Project Aditya worked on as a side project
• Python - Programming language mentioned in context of package conflicts
• JavaScript - Programming language mentioned in comparison to lower-level languages
• C - Programming language mentioned as more low-level than JavaScript/Python
• Figma - Design tool mentioned as a modern prototyping platform

Concepts:

• Augmented reality (AR) - Technology Mike used in his crime tracker app
• Open city data - San Francisco data that Mike used for his AR crime tracker
• Yak shaving - Programming term for getting caught in recursive problem-solving
• Home-cooked meal apps - Robin Sloan's concept of apps built for personal satisfaction rather than scale

Timestamp: [10:00-19:59]

Mike elaborated on the crucial "minus one" decision to pivot away from Bourbon, their initial product, which he described as being at "negative 0.5" - not quite at the starting point of a successful venture. This decision exemplifies one of the toughest challenges founders face: abandoning "good" ideas in pursuit of potentially "great" ones.

"We started working on things and about three months in we had Bourbon, the HTML5 thing. We were distributing it in private beta but trying to spread it, and it was hard to explain what we were doing because it had become too many ANDs."

The original Bourbon app had become a convoluted product with multiple features - location check-ins, photo sharing, video attachments, games, invitation systems, and planning tools. Mike reflected that post-Instagram, he realized that whenever a company pitch contains "a lot of conjunctions," it's probably heading in the wrong direction.

Despite having over a thousand users who expressed enthusiasm for the product, including a native iPhone app version, Mike and Kevin struggled to see a path to substantial growth. The pressure of limited runway - both financial (their $500,000 seed funding) and Mike's visa situation - created urgency to find a viable direction.

The pivotal moment came when they realized one of their investors had mentally written them off, believing they were stuck on an unviable idea. This external perspective helped them turn "the mirror on ourselves" and recognize they needed to reassess whether Bourbon was truly worth tripling down on, or if they had simply "stacked a bunch of somewhat related stuff" without a coherent vision.

Timestamp: [20:04-23:09]

Aditya observed that one of the most challenging situations for founders is dealing with moderate traction - not a complete failure but not a breakout success either. This middle ground can be particularly difficult to navigate because it provides just enough positive feedback to keep founders invested in an idea that may ultimately be limited.

"That takes courage. I mean, I don't think that's easy. I think people, even today, I see it is so hard to essentially give up on an idea that has that middle amount of traction. It's almost like that is the fate you don't wish on a founder because it's much better to actually get zero traction, then you have clear, easy signal, right? Like, go and move on to the next thing. It's when you have a little bit of traction, then you get into trouble."

Mike agreed, noting that emotional ties to users and positive feedback loops make it even harder to walk away:

"Especially if you've built the feedback loops where that traction is like 'Oh, my friends like it' or 'I've met people through it' and people are using it. I think that's the hardest thing."

He then shared a counterintuitive insight from his experience:

"The things I used to believe about product that I no longer believe is that good retention is indicative of a thing that will do well in the long run. You can have very well-retaining, low-growth products. That's the maybe unintuitive result that I've found."

Aditya confirmed this observation, noting that while conventional wisdom suggests that 40% retention at the end of month two puts you in the top 1% of products, high retention alone doesn't guarantee high growth. Both metrics are necessary for true success.

This conversation highlights a crucial lesson for founders in the "minus one" phase: being able to objectively evaluate moderate success and recognize when it's time to pivot, despite encouraging signals from a small but enthusiastic user base.

Timestamp: [23:10-24:15]

When Aditya asked about the timeline between abandoning Bourbon and launching Instagram, Mike revealed that the transition happened remarkably quickly. The team made a crucial observation about user behavior on their original platform:

"This product is pretty good, but it's HTML5 mostly, only a few people using the iPhone app. And the photos part works even though it's really hard. And I think that's interesting - if people are finding a way to use your product for something that is like built for but not built that well for, there's something there. There's like a desire there."

This insight - that users were gravitating toward photo sharing despite it not being Bourbon's primary function or strength - provided a clear signal about where to focus their efforts. They decided to extract this single feature and build an entirely new experience around it.

The urgency of their situation accelerated the decision-making process. Mike was paying himself the H-1B minimum wage (the "prevailing wage" required for visa compliance), and their runway was limited. As he put it, "the clock is ticking" and they needed to find something viable quickly.

Interestingly, they maintained Bourbon while launching Instagram, keeping it as a fallback option if their new direction failed. This reflects their uncertainty about whether this pivot would succeed, but also their commitment to finding a viable path forward:

"We didn't unship Bourbon and we were ready to turn off Instagram if it didn't work and go back to trying something else on Bourbon. But at least [we had] conviction that we needed to move on something."

Their approach demonstrated a balance between decisive action and strategic flexibility - committing fully to Instagram while maintaining the option to return to their previous platform if necessary.

Timestamp: [24:16-25:33]

Mike shared a guiding principle that he and Kevin had learned at Stanford that helped them navigate uncertain decisions: "We may not be right, but at least we're not confused." This philosophy became crucial to their approach at Instagram - embracing clarity of direction even amid uncertainty about outcomes.

"That is a really guiding principle I think for exploration. It's like you're not sure if you're right, you won't know actually maybe potentially for years, but don't confuse yourself by hedging and being like 'Well, we're going to build a thing that's a little bit of this, little bit of that' or 'We're going to keep two prototypes going.'"

Mike expressed his strong belief in picking a clear direction with an "eyed view" about the specific problem you're solving, then periodically reevaluating whether that direction remains correct. He contrasted this with a pattern he's observed in other founders:

"I've also seen founders be in like hedging land where they're working on two or even three things at once, and I've never seen that be the path to success."

Aditya expanded this observation beyond product development to career and life decisions, noting that people often attempt to hedge in major life choices as well:

"That hedging pattern ends up repeating itself in many different ways. It's not just in products. I think people end up doing it with their careers, people often end up doing it with like big life decisions."

He advocated for committed choices with scheduled reevaluation points rather than concurrent hedging:

"It's okay to reevaluate, but you cannot be reevaluating concurrently. You can reevaluate every 6 months, 12 months, like give yourself some time, but it's just so bad for your psychology if you're reevaluating and living multiple versions of your life at the same time."

Mike connected this principle to executive hiring, sharing advice he'd received: commit fully to new hires for an initial period (like three months) as if they're perfect for the role, then have a designated checkpoint to reassess. This approach ensures that if things don't work out, it wasn't due to lack of belief or investment.

Timestamp: [25:34-27:31]

Expanding on Mike's comments about executive hiring, Aditya shared insights about the universal pattern he's observed when integrating senior leaders into companies. He described a predictable emotional journey that occurs with virtually every executive hire:

"It is literally every single executive that I have had the chance of working with who got hired into a company - they have everybody goes through that initial euphoria, and then everybody goes through the trough of despair because you're like, 'Hey, this person isn't perfect,' you know, and 'Why do they have so much to learn? Why haven't they fixed all my problems yet?'"

This observation highlights the unrealistic expectations that founders and teams often place on new executive hires, expecting immediate transformation and perfect execution from day one. Aditya emphasized that integration takes time, especially when working with strong founders who have deeply ingrained ways of operating.

The conversation reveals an important lesson about leadership transitions: there's an adjustment period that almost always includes an initial honeymoon phase followed by disappointment before reaching productive equilibrium. Understanding this pattern can help founders set more realistic expectations and provide the necessary support and patience during an executive's onboarding process.

This insight connects to the broader theme of avoiding hedging - by committing fully to new executives (as Mike suggested) while recognizing that the integration process includes predictable ups and downs, companies can create more successful leadership transitions.

Timestamp: [27:32-28:03]

When asked about technical decisions that made a significant difference in Instagram's first three to six months, Mike immediately highlighted their relentless focus on performance optimization - particularly for less capable devices like what he jokingly called "the diseased iPhone 3GS."

This commitment to performance wasn't a one-time effort but an ongoing priority that shaped their engineering culture:

"We were renting a very small office and we would do this exercise where we would take all of our endpoints and measure their average latency, and then we'd make a goal to like, by the end of the week, cut all that in half. Because it's early - there's always stuff to cut, they're always doing something too inefficiently."

Mike emphasized that this performance obsession directly contributed to user engagement:

"100% I believe that stuff accrues where like, people open the app and it feels snappy, it's an app they want to open up if they have a minute before... I was going to say their Uber arrives. It was no Uber back then. Their Muni bus arrives. They have something to do that's not going to take a whole minute to do."

He summed up this philosophy with a memorable phrase: "Don't make people wait while they're waiting." This insight recognizes that mobile apps often fill brief moments of boredom or transition in people's lives, and introducing friction through slow performance defeats that purpose.

Beyond performance, Mike identified another crucial technical priority: ensuring that building the "next adjacent product thing" would be easy. This focus on developer velocity and maintaining quality code architecture allowed them to iterate quickly and adapt to emerging opportunities - a capability that would prove essential to Instagram's continued evolution.

Timestamp: [28:04-29:55]

• Products with "too many ANDs" in their pitch often lack a clear focus and are harder to execute successfully
• The most difficult position for founders is dealing with moderate traction - enough to provide hope but not enough to indicate breakout success
• High retention alone doesn't guarantee success - a product can retain users well but still fail to achieve significant growth
• Pay attention to what users are doing with your product, not just what you built it to do - user behavior can reveal unexpected opportunities
• "We may not be right, but at least we're not confused" is a powerful guiding principle for navigating uncertainty in startups
• Hedging across multiple products, directions, or life choices creates psychological strain and dilutes focus
• Commit fully to choices with scheduled reevaluation points rather than continuously questioning decisions
• Every executive hire follows a predictable pattern: initial euphoria followed by a "trough of despair" before reaching productive equilibrium
• Performance optimization was crucial to Instagram's early success - making the app feel snappy encouraged casual, frequent usage
• "Don't make people wait while they're waiting" - recognizing that mobile apps often fill brief moments in people's lives
• Building with developer velocity in mind enables faster iteration and adaptation to new opportunities

Timestamp: [20:04-29:55]

Products/Companies:

• Bourbon - Mike and Kevin's original product with multiple features before pivoting to Instagram
• Instagram - The focused photo-sharing app that emerged from Bourbon
• South Park Commons (SPC) - Aditya's organization, mentioned in context of product philosophy
• Uber - Mentioned as a comparison for waiting scenarios (didn't exist during Instagram's early days)
• Muni - San Francisco's public transportation system, used as an example of waiting scenarios

People:

• Kevin - Kevin Systrom, Mike's co-founder at Instagram

Devices/Technology:

• iPhone 3GS - Older iPhone model that helped set Instagram's performance standards
• iPhone 4 - Mentioned in comparison to the 3GS
• HTML5 - Technology used to build Bourbon

Concepts:

• "Too many ANDs" - Concept describing overcomplicated products with too many features
• "We may not be right, but at least we're not confused" - Stanford principle about clarity in decision-making
• Hedging - The practice of pursuing multiple directions simultaneously
• Developer velocity - Focus on enabling quick iteration in software development
• Endpoint latency - Technical measurement of API response time
• H-1B visa - Work visa mentioned in context of Mike's immigration situation
• Prevailing wage - Minimum salary requirement for H-1B visa holders
• The executive adjustment curve - Pattern of initial euphoria followed by "trough of despair"
• "Don't make people wait while they're waiting" - Mike's philosophy on app performance

Places:

• Stanford - University where Mike and Kevin learned important principles

Timestamp: [20:04-29:55]

Mike expanded on Instagram's technical foundations, highlighting specific engineering practices that helped them move quickly while maintaining quality. Despite being a tiny team of just two founders, they implemented professional development practices that were unusual for early-stage startups:

"We had great unit tests, which was pretty rare I think for a company that was just me and Kevin. We used Django as our backend... we had factored it in a way where all the different pieces were pretty decoupled. We used their signal system - all this stuff that was definitely to go fast at the beginning by going slow, or optimize your workbench, sharpen your tools before going in."

Mike contrasted their approach with counterproductive patterns he's observed in other startups:

"I've seen companies fall into the trap like 'we wrote our own programming language' and you're like, 'it's two years in, have you shipped anything?'"

Finding the right balance between quality and speed was crucial to Instagram's success - creating a development environment that supported "high quality, nimble iteration speed." This wasn't just an early Instagram phenomenon - Mike noted they recreated this engineering culture at Artifact, his later venture:

"I've talked to multiple people that were at Artifact that are now at other companies. They're like, 'Man, I really miss our dev environment. Iteration felt really, really, really good there.'"

The third key technical practice Mike highlighted was their attention to qualitative network monitoring - looking beyond simple metrics to understand what was actually happening on the platform. This became especially important when they began exploring machine learning features:

"Very easy to get caught in stats and say like 'Photo uploads went up by 10% this week' or 'our new ML algorithm on Instagram Explore increased time spent by 5%' - but what's actually happening under the hood? We had a really good pulse and built internal dashboards around just keeping pulse on what the health of the network was at a more qualitative level."

Mike noted that maintaining this qualitative perspective becomes increasingly challenging as a company scales, requiring deliberate investment in tools and processes.

Timestamp: [30:00-31:30]

Aditya strongly resonated with Mike's point about maintaining qualitative awareness alongside metrics, highlighting an important distinction in product development philosophy:

"Something that especially with consumer products keeps on coming up over and over again is that you can—I believe you should not be data-driven, but you should really be data-informed. The data should be in some ways informing your qualitative sense of what's going on and not the other way around."

This perspective represents a nuanced approach to using analytics in product development. Rather than letting metrics dictate decisions (data-driven), a data-informed approach uses quantitative information to enhance and validate qualitative understanding.

The distinction is subtle but profound:

• Data-driven: Metrics lead decision-making; the numbers determine direction
• Data-informed: Metrics support decision-making; numbers provide context for human judgment

For consumer products especially, this balance helps teams avoid the trap of optimizing for metrics that might improve numbers but damage the overall user experience or product integrity. The conversation suggests that Instagram's success was partly due to this balanced approach—tracking performance indicators while maintaining a strong qualitative understanding of how the platform was actually being used and experienced.

Timestamp: [31:31-31:59]

Aditya recalled an iconic image from when Instagram first joined Facebook—approximately 12 team members in what he remembered as "a big hangar," which Mike clarified was actually a garage-style space with open doors on Facebook's old campus. This visual perfectly captured the stark contrast between Instagram's small team and Facebook's massive organization.

Aditya wondered if this transition felt like another "minus one" moment—rebuilding something within a larger context. Mike's response revealed the dual challenges they faced:

"I think it's simultaneously going back to minus one in terms of 'how do we operate, how are we going to grow this team,' but also having to grow up much quicker. So I think we were stretched on both ends."

On one side, they were re-establishing their foundation—building new relationships, adding team members, and shifting from their extremely short-term planning approach:

"Up to 2012, we were so understaffed that we weren't doing a lot of product planning. We were looking at where we were, where we wanted to be two to three weeks from now, and then building to that. It was very much Sprint to Sprint to Sprint to Sprint. And then it was like, 'Okay, we've got to start planning and thinking about things more thoroughly.'"

Simultaneously, they faced the demands of operating within a large public company with established processes:

"All of a sudden we had a lot more scrutiny because Facebook had the legal teams, the privacy teams, and all these things that come with the expectations of being larger, so we had to grow up very quickly in a way that was difficult."

Mike shared a specific example that illustrated this rapid maturation:

"All of a sudden the FBI has a request of your data, and it's Facebook—you can't play the 'Oh, we're a small startup, give us a few days, we've never done this before.' It's like, 'No, no, no, there's a process and we expect you to have this response in this time.'"

One of the most challenging aspects was that this organizational scaling happened behind the scenes:

"You're doing it in a way that doesn't reflect itself in the product. Nobody sees that you're having to basically revamp all of your internal processes and internal tooling to be a bigger company, but you have to do it."

As Aditya summarized, it was like "how do you grow up to be an adult in like one year," capturing the accelerated organizational maturation Instagram underwent following the acquisition.

Timestamp: [32:00-34:24]

Aditya shifted the conversation to Mike's decision to leave Instagram after eight years, acknowledging that while Mike and Kevin had become "rock stars" and "celebrities in the technology world," making this decision must have been emotionally challenging. He asked Mike about how he processed the emotional and intellectual signals that led to his conviction to leave.

Mike reflected on the complex personal journey of separating his identity from the company he co-founded:

"I think it was hard. This was now eight years into the Instagram journey. My role was always more internally focused than Kevin's—he had both internal and external focus and did more interviews. But my role had shifted to have some more external presence, whether it was being part of a longer magazine interview."

To illustrate the difficulty of separation, Mike shared a story about his efforts to connect Formula 1 (a personal passion) with Instagram:

"Right after Liberty had bought Formula 1, I was like, 'Oh great, this is an opportunity to actually get the drivers on social media because they couldn't really use anything before.' I was sitting there a couple months later as we were thinking about leaving, I was like, 'Man, that was such a cool door to open.' Today Instagram and F1 are a really good combination, so it was some good seeds that were planted there."

This experience highlighted a profound identity question:

"You start realizing how much of your public persona is associated with the thing that you started and how many friends you've made via that. A lot of coming to reckoning with that decision was just like, 'Who am I if I'm not the CTO of Instagram?'"

Mike acknowledged the temptation to take a defensive stance ("if they don't like me after I leave Instagram, they weren't my friends anyway"), but admitted, "you're only human, so you still have to think about this stuff."

Beyond identity, another major challenge was leaving the team they had built:

"Breaking off from the team was the other big thing. We built this team from two people to the 12 people in the Facebook garage to, at the time we left, we were a thousand. It was a big chunk of people that I knew very well and all of a sudden wouldn't get to see every day."

In summary, Mike identified the "softer, interpersonal, reputational, public persona pieces" as the hardest parts about leaving Instagram—concerns that went far beyond the work itself.

Timestamp: [34:25-37:38]

Aditya shared insights from his experience counseling founders transitioning out of their companies, identifying three consistent patterns of loss that people experience:

"I kind of tell them those same three things. I'm like, you will miss your identity, because no matter how strongly you feel you can disassociate—you know, in Silicon Valley, we are often an expression of the work that we do, and you will miss that identity."

This loss of professional identity can be particularly acute for founders whose names become synonymous with their companies, as was the case with Mike and Instagram.

"You will not actually miss the work, and you know, it's rare to find someone who, like, three months later is like, 'Oh man, I really miss that exact meeting or that thing I was doing.' It's very rare to hear that."

This second observation highlights an interesting paradox—while founders may struggle with losing their professional identity, they rarely miss the day-to-day work itself once they've moved on.

"But you will sorely miss the people, right? Like, you make connections, you make dear, emotional connections, and that gets better with time, but I still miss some of the people that I used to work with because those are my fondest memories of my time at Facebook and Dropbox."

Mike acknowledged that he wished he'd had this framework during his own transition:

"I wish I had you back then. I definitely felt like I did not have a playbook. It was like, 'I'm making this leap, and I think I'll be fine.'"

He also noted how his personal circumstances affected his experience leaving Instagram versus his later departure from Artifact:

"At that point [leaving Instagram], we had like a dog, we were married, but it was very different actually leaving Artifact, which we can talk about in a bit. At that point, we had two kids, so it was like, 'Okay, I have something I can feel like is occupying a lot of my emotional and time investment rather than the thing that I was working on.'"

This comparison highlights how personal life circumstances can significantly impact the experience of professional transitions—having children provided an emotional anchor during his later career transition that wasn't present when leaving Instagram.

Timestamp: [37:39-39:07]

Aditya asked Mike about his experience in the immediate aftermath of leaving Instagram – what his first 30 days looked like, whether he had a plan, or if he felt a sense of disorientation without his usual routine and purpose.

"Did you have a plan? Did you know what you were going to do when you got up? Or did you feel like a profound sense of unmooring, like, 'Hey, what am I supposed to do with my time?' Or had you planned for that?"

Mike revealed that he had actually created a buffer between his professional departure and confronting that existential question:

"I deferred my unmooring entirely. We had a trip planned to Europe that we'd been planning for months, so it was actually great."

This pre-planned trip provided a natural transition period, allowing Mike to step away from his professional identity before having to confront the question of "what next." He described a serendipitous moment during this trip that seemed to affirm his new chapter:

"One of those things where just life, the universe is showing you something. One of the visits that we had on this trip, we went to Barcelona, and I had this pre-scheduled meetup with the chef behind El Bulli, Ferran Adrià."

This brief glimpse into the beginning of Mike's post-Instagram journey suggests that creating space and distance through travel was an important part of his transition strategy, allowing him to decompress and gradually explore new possibilities without the immediate pressure to define his next professional identity.

The conversation was headed toward Mike's encounter with the renowned chef Ferran Adrià, promising insights about how this meeting might have influenced his perspective during this transitional period.

Timestamp: [39:08-39:59]

• Investing in technical infrastructure and best practices (like comprehensive unit tests) from the beginning accelerates development speed in the long run
• Finding the balance between quality and iteration speed is crucial for startups - avoid both over-engineering and technical shortcuts
• Being "data-informed" rather than "data-driven" allows for better product decisions by using metrics to support qualitative understanding rather than letting numbers dictate direction
• Acquisition transitions create dual challenges: building new interpersonal foundations while rapidly maturing organizational processes
• Post-acquisition scaling often involves substantial behind-the-scenes work on processes and systems that isn't visible in the product but is essential for operating at scale
• A founder's identity becomes deeply intertwined with their company, making separation a profound identity challenge
• External relationships often change after leaving a high-profile company, requiring emotional preparation
• People rarely miss the actual work after leaving a company, but deeply miss team relationships and connections
• Having alternative sources of identity and purpose (like family) can significantly ease the transition out of a company
• Creating physical and mental space through travel or breaks can help founders process transitions and discover new directions

Timestamp: [30:00-39:59]

People:

• Kevin - Kevin Systrom, Instagram co-founder
• Ferran Adrià - Chef behind El Bulli restaurant, whom Mike met in Barcelona
• Chase Carey - Mentioned in context of Formula 1 leadership after Liberty Media acquisition

Companies/Organizations:

• Instagram - Social media platform co-founded by Mike
• Facebook - Company that acquired Instagram
• Artifact - Company Mike founded after Instagram
• FBI - Mentioned regarding data requests to Facebook
• Liberty Media - Company that acquired Formula 1, mentioned in Mike's story
• Formula 1 (F1) - Racing organization Mike worked to connect with Instagram
• Dropbox - Company Aditya worked at, mentioned in context of missing former colleagues
• El Bulli - Famous restaurant run by Ferran Adrià

Technologies/Development Tools:

• Django - Web framework used for Instagram's backend
• Django signal system - Specific feature used for decoupling components
• Unit tests - Testing methodology implemented at Instagram
• ML/Machine Learning - Technology Instagram began exploring for features like Explore
• Internal dashboards - Tools built to monitor network health

Places:

• Europe - Where Mike traveled after leaving Instagram
• Barcelona - City Mike visited where he met Ferran Adrià

Concepts:

• "Going fast by going slow" - Principle of investing in infrastructure to enable speed later
• Data-informed vs. data-driven - Distinction in approach to using metrics
• Identity loss - Challenge faced when leaving a company
• Unmooring - Feeling of loss of direction after a major transition
• Organizational maturation - Process of developing processes and systems for larger scale

Timestamp: [30:00-39:59]

During his post-Instagram European trip, Mike had a serendipitous meeting with Ferran Adrià, the legendary chef behind El Bulli, once considered the world's best restaurant. This encounter became a profound moment of inspiration during Mike's career transition.

El Bulli was famously closed at its peak by Adrià, who chose to reinvent rather than simply continue his success. Mike reflected on the parallel to the tech industry:

"You hit number one in the world's best restaurants, you're like three or four years from some crazy reboot, which maybe tells us something about how to think about this stuff in tech as well."

The meeting took place in Adrià's warehouse, which Mike described as both a "workshop for his next idea" and a "temple for El Bulli" - simultaneously preserving the legacy of what came before while creating space for future innovation. What struck Mike most was a neon sign in the space:

"He had a neon sign that says 'We closed El Bulli to open El Bulli,' which I think was such a cool thing to read and experience right when I was just leaving Instagram. It's like 'I left Instagram to start Instagram'—not Instagram again, but whatever is going to happen next... It could not have come at a better moment of transition for me."

This powerful statement about purposeful endings leading to new beginnings resonated deeply with Mike at this pivotal moment in his career. The chef's philosophy of deliberate closure to enable reinvention provided a framework for Mike to process his own transition away from Instagram.

Timestamp: [40:01-40:59]

Upon returning home from Europe, Mike faced the challenge of structuring his newfound freedom. True to his nature as a self-described "list maker" and "bullet point guy," he opened a blank document in WorkFlowy ("bullets as a service") to plan his next year.

"I remember sitting down the week I got back and it was an empty document. I love making lists... and I was using WorkFlowy, which is like bullets as a service, right? You can just do bullets on bullets on bullets. And I was like, 'Alright, I'm going to make my plan now for the next year.'"

Mike developed what he now humorously calls a "not great idea" - pitching Netflix on a series where he would explore his various curiosities by embedding himself in different organizations:

"I'm curious about Formula 1 racing. I'd love to go basically be an intern on their software aerodynamics team. I'm like weirdly passionate about Disney; I want to go work with the Imagineers. I was going to pitch Netflix on a series where they follow me around and I'll do like a month on each of those things."

He quickly recognized the impracticality of this concept, but the exercise revealed his wide-ranging interests. The list included "the year of learning," reflecting his college interest in neuroscience, which led him to experiment with brain-computer interfaces:

"I got one of these 3D printed things and I spent a week hooking it up. I had the whole thing hooked up and I was measuring alpha and gamma waves and meditating. I was going to build this meditating VR app, you know, just like ideas."

After a week or two, his enthusiasm waned: "I don't feel like doing this anymore. This is not that interesting, it's not that rewarding." Looking back at his ambitious list later, Mike was struck by how few items he actually pursued, and came to accept that this was "fine" - the value was in the exploration process itself.

This period of wide-ranging exploration, though seemingly unproductive in the moment, was an important part of Mike's career transition, helping him discover what truly engaged his interest.

Timestamp: [41:00-43:16]

After Mike mentioned watching TV during his time off, Aditya made a passionate recommendation for "The Bear" Season 3, drawing an insightful parallel between restaurant excellence and startup creation:

"You should watch The Bear Season 3 if you haven't. It is the perfect articulation of, in my opinion, creating greatness in a restaurant as a metaphor for starting a company - the level of craft, attention to detail, the insanity of like inter-team dynamics. It is actually a work of art. It might be the best TV I've seen in a while."

This recommendation resonated with the themes of their conversation about craft, excellence, and the challenging interpersonal dynamics in high-performance environments. The comparison between restaurant creation and startup building highlights how both domains require obsessive attention to detail, mastery of craft, and navigation of complex team dynamics under pressure.

Mike confirmed he had watched the first two seasons and was planning to watch the third, showing their shared appreciation for quality storytelling that reflects entrepreneurial themes.

This brief exchange not only revealed their shared interest in television but also demonstrated how creative works can provide valuable frameworks for understanding the entrepreneurial journey.

Timestamp: [43:17-43:53]

Mike reflected on how he initially viewed his post-Instagram career options in a surprisingly limited framework:

"I had in my head, I'm like, 'Okay, I've done this company, and now I need to either ride off into the sunset never to be seen again, start another company, go be an exec at a different company, or become an investor.' And it felt, you know, that was kind of like, 'Well, that's my option set now.'"

This reductionist framing led him to approach his exploration with a specific destination already in mind, particularly focusing on finding his next company:

"Maybe I need to be thinking about the next company, so why don't I start prototyping and building things that might be the next company? And that was not that satisfying because then you're not building to scratch an itch or out of intellectual curiosity. You're building to build, or building with this very specific destination in mind."

Aditya noted that this exact same thinking had influenced Ruchi Sanghvi's founding of South Park Commons:

"Ruchi, when she talks about the founding of South Park Commons, uses quite literally the same words. She's like, 'Listen, I'd had a good run in Silicon Valley, so I could have started a company, been an executive at a company, started my venture fund, joined a venture fund. But what I really wanted to do was: what is the stuff that I don't know could be interesting? What do I need to go tinker with and learn?' And that was literally the original inspiration for South Park Commons."

Mike shared that he worked with a coach during this period who recommended a book about transitions, which emphasized the importance of sitting in the "messy" uncertain period rather than rushing to the next thing. Through coaching exercises focused on what gave him energy, Mike realized:

"I like a lot of ideas. I don't like just doing one, clearly—I can't decide, it was a theme of this podcast and of my life. I love building, I love being around builders."

His coach suggested he consider creating "some way of starting a lab or some kind of incubation space," revealing parallels with the eventual founding philosophy of South Park Commons, though Mike ultimately took a different path.

Timestamp: [43:54-46:17]

With limited time remaining in the interview, Aditya asked Mike to explain how he eventually converged on starting Artifact. Mike described this journey in three distinct chapters:

"For us it was like three, I would say three chapters. Chapter one was we had this hypothesis: one of the four options is you invest, but we didn't want to go join a venture fund, still wanted to preserve some optionality. So let's do angel investing, but take it seriously."

Rather than joining an established fund, Mike and Kevin created their own structured approach to angel investing:

"We had our deal flow, we reached out to a lot of investors that we liked and said, 'Hey, if we could add value to a deal, let us know.' For founders that have done well and will hopefully add value, it's fairly easy to get a small allocation in rounds."

They pursued this seriously for about a year, and while Mike was proud of their investments, something was missing:

"I missed building. I knew I missed building."

The pandemic arrived just as Mike was considering starting another company, which he now views as a "silver lining":

"COVID was instantly, you know, so challenging and awful and completely disruptive, but the silver lining for us was, 'Okay, nobody's expecting us to start a company right now. The world is just crazy. We can do things that are hopefully helpful without it needing to be a company.'"

This led them to create Rt.live, using machine learning to predict COVID spread by state and eventually by country. The project allowed Mike and Kevin to explore their interests without pressure:

"Both of us had interest in machine learning. Kevin went really deep on machine learning for that project. For me, it was a way of reconnecting with front-end engineering, which I hadn't done for a while. Again, we weren't doing it thinking like our next company is going to have front-end engineering and machine learning. It was like, 'No, this is a thing that we want to do right now.'"

Mike also built a "whale community notifier" with his wife Kaitlyn while staying in West Marin, combining the Twitter API with mobile development—another "fun project" that wasn't intended to become a company.

As the pandemic began to ease, they started thinking more deliberately about their next venture:

"You're kind of still in COVID, but the wave has gone down. You can start thinking about what the next thing is. We were still very interested in machine learning. We'd been thinking about news and in general like how do people get their information. And we really wanted to move beyond the photo-photo-photo, which was so a core part of Instagram."

These interests converged into what would become Artifact. Their approach was notably informal at first:

"We spent the first three, four months—there was no LLC, S Corp, C Corp. It was just two of us hacking on things and collaborating. You can have that trust when you've built something with somebody, that we'll figure out those details in a little bit."

Mike contrasted this with first-time founders who prioritize incorporation before building anything. Through prototyping, they discovered: "This is fun. It's really fun to build. We're in a fun space. It's really fun to explore what machine learning is doing."

This evolutionary approach—from angel investing to COVID response projects to informal prototyping—eventually led to the formal creation of Artifact, with each stage building naturally on the skills and interests developed in the previous phase.

Timestamp: [46:18-49:59]

• Deliberate endings can enable powerful new beginnings, as exemplified by Ferran Adrià's philosophy "We closed El Bulli to open El Bulli"
• Seemingly serendipitous encounters and experiences can provide profound guidance during career transitions
• List-making and structured exploration can be valuable during transitions, even if many ideas are never pursued
• Building with a predetermined outcome in mind ("this must become a company") can limit creativity and satisfaction
• Successful founders often feel boxed into limited post-success pathways (start another company, become an executive, become an investor, or disappear)
• Working with a coach or guide can help navigate the uncertainty and emotional challenges of major career transitions
• The "messy middle" period of transition is uncomfortable but necessary before finding a new direction
• Angel investing alone may not satisfy builders who derive energy from creating products
• Crises like the pandemic can create unexpected opportunities to explore interests without external pressure or expectations
• Building trust through previous collaborations allows for informal experimentation before formalizing a new venture
• The path to a new venture is rarely linear and often involves multiple phases of exploration
• Projects undertaken for intrinsic interest or social good can unexpectedly build skills and relationships that lead to new ventures
• First-time founders often focus on formalities (like incorporation) before building anything of substance

Timestamp: [40:01-49:59]

People:

• Ferran Adrià - Chef behind El Bulli restaurant
• Kevin - Kevin Systrom, Mike's co-founder at Instagram and Artifact
• Kaitlyn - Mike's wife, worked with him on the whale notifier project
• Ruchi Sanghvi - Founder of South Park Commons

Companies/Organizations:

• El Bulli - Former world's best restaurant, closed at its peak by Ferran Adrià
• Instagram - Social media platform co-founded by Mike
• Netflix - Streaming service Mike considered pitching a show concept to
• Formula 1 - Racing organization Mike was interested in
• Disney - Company Mike was passionate about, specifically Imagineers
• South Park Commons (SPC) - Organization founded by Ruchi Sanghvi
• Artifact - News app founded by Mike and Kevin after Instagram
• Rt.live - COVID tracking project created by Mike and Kevin during the pandemic

Places:

• Europe - Where Mike traveled after leaving Instagram
• Barcelona - City where Mike met Ferran Adrià
• West Marin - Where Mike and his wife spent time during COVID and built the whale notifier

Products/Technologies:

• WorkFlowy - Note-taking app described as "bullets as a service"
• Brain-computer interfaces - Technology Mike experimented with
• EEG - Technology for measuring brain waves that Mike explored
• Alpha and gamma waves - Brain activity patterns Mike measured
• VR - Virtual reality, mentioned for potential meditation app
• Twitter API - Used in Mike's whale notifier project
• Machine learning - Technology explored in Rt.live and later Artifact

Media:

• The Bear (Season 3) - TV show Aditya recommended as metaphor for startups
• Book about Transitions - Unspecified book recommended by Mike's coach

Concepts:

• "We closed El Bulli to open El Bulli" - Adrià's philosophy about reinvention
• Angel investing - Initial post-Instagram activity for Mike and Kevin
• The four post-success options - Start another company, join as executive, become investor, or disappear
• "Sitting in the mess" - Concept from transitions book about embracing uncertainty
• Whale migration patterns - Subject of Mike and Kaitlyn's community project

Timestamp: [40:01-49:59]

Mike explained how their experimental prototyping evolved into a formal company. The decision to move forward came from a combination of personal enjoyment and excitement about the technology space:

"It was fun. It's really fun to build. We're in a fun space. It's really fun to explore what machine learning is doing. The product we're building has significant other-market fit, which is all you can ask for at the very beginning."

This "other-market fit" - where the founders themselves find genuine value in the product - provided the confidence to move forward. Mike identified the key factors in their decision:

"I think primarily that we wanted to build together again, and that was the highest order bit. And the second one was this space of personalization and machine learning was really interesting."

He noted that this was around 2020-early 2021, before the current AI revolution: "AI was not yet what it is today, so we were talking about machine learning versus AI."

The transition from informal experimentation to officially forming a company came from a deeper emotional need:

"The leap from that to 'let's form a company and go hire people' came from a concern that both of us had... We've done the company thing before, we want to do a company, I'll be uneasy if I'm not working on something great, but how do we actually make it feel real? That's where the leap came from - let's bring other people on and actually hire people."

This reveals how the formalization of Artifact was driven not just by the product opportunity, but by their psychological need to be building something substantial with a team - moving from hobbyist experimentation to a committed venture.

Timestamp: [50:04-51:01]

Aditya raised a sensitive question about the psychological burden of launching a second company after massive success: "Did you guys fear how people were going to perceive it?" He acknowledged that ego plays a role for everyone, noting "it's not easy to have a second act in public."

Mike candidly shared both the psychological and practical challenges:

"Psychologically, you're like, 'Well, can I build something as big as Instagram?' Like, probabilistically no. That thing is like, you know, top three-four apps ever. The odds that you will hit that are quite low, no matter how good you are. And even if you get lucky a second time, are you going to get lucky in the same way? It was unprecedented, and it really blocked me for a while."

He revealed a personal realization he had never shared publicly before:

"This is going to sound funny, I don't think I've ever actually said it out loud, especially not on a podcast, but hitting a million users was very cool, hitting 10 [million] was really cool, 100 [million] was interesting. Between 100 and a billion, it sounds awful, but it was like 'cool, another milestone'—but it becomes so abstracted at some point."

This perspective helped him approach Artifact with a healthier mindset:

"It might not hit a billion users—an absurd number of people using anything including a product—but that wasn't why I was still working on it. I was working on it because I loved building with the people I was with. Every time I drifted and was like, 'Oh, but what if this doesn't work?' I went back to 'What do you actually want to be doing?' and just regrounding myself in that as much as I could."

Beyond the emotional aspects, these pressures affected their product development approach:

"It also had an effect on how we thought about launch. Here's the bar for launch, right? It's not Instagram V1 where we didn't have collapsable comments... We didn't summarize our likes... For the longest time we didn't have commas in our numbers in the stats part of your profile."

Mike explained that the bar they held themselves to on product quality, social features, and machine learning was extremely high—perhaps too high:

"I think it led us to launch much later than we should have. We built for about a year and a half before launching, and the last six months we were building things at high quality, I think, but not getting that same contact with the real world that we were talking about."

This perfectionism had consequences: "We got more data into the system in the first 12 hours of being out in the world than we had for the previous six months." For a machine learning system that thrives on data, this delay was significant.

The cumulative effect was starting their journey from a place of exhaustion rather than excitement:

"When we got to the start line... we went minus one to zero, but it was a much more tiring journey than it would have been otherwise. We didn't start bright and bushy-tailed at zero... We didn't start with the same freshness as I think we did with Instagram, and I think that made it harder to iterate from that point forward."

Timestamp: [51:02-54:46]

Aditya responded to Mike's reflection about building Artifact with a deeply personal observation about his own approach to career fulfillment:

"I'll come back to one thing that you said that really resonated with me, and it's something I often think about in my own life, which is that at some point, even if you've built a bunch of things that have done well and you're proud of them, I'm just like, 'But I still need something to do every day for the rest of my life that I like doing.' That is almost the point."

This perspective shifts the focus from outcomes and achievements to the daily experience:

"For me, a lot of it boils down to: I like building things with smart people. I also enjoy it when they're successful, but that's afterwards. On a day-to-day basis, I like having something to do with smart people—the act of building. Once I am able to reduce it down to that, then I'm just like, 'You know what, I can be a happy camper because I know I can do that if I want to every single day.'"

Mike enthusiastically agreed with this framework:

"100%. And that's what I was choosing. Getting to do that day-to-day... Did we need to do another company? Not from a career perspective, but from a 'I am very happy to be engaged on product' perspective."

He highlighted how Artifact allowed him to be hands-on technically, which was personally gratifying after his later years at Instagram involved more management:

"I got to write so much code through the Artifact journey. I did a bunch of the iOS stuff, did a lot of the backend, some machine learning infrastructure. It was so cool and fun to be able to do that after being at Instagram where I was managing a really big team. Just being back in there was great."

This exchange reveals a profound insight about sustainable career happiness: while outcomes and impact matter, daily engagement in activities that bring personal satisfaction may be the more reliable path to fulfillment, especially after achieving significant success.

Timestamp: [54:47-56:13]

Aditya noted the rarity of founders who decisively shut down consumer apps at Artifact's stage, especially when many tend to continue as "zombies." He asked Mike about the decision-making process that led to their bold move to end the product.

Mike explained it was driven by two key factors. First, it wasn't about absolute scale:

"It's funny because people are like, 'Well, you just shut it down because it didn't reach Instagram scale.' It wasn't that it wasn't on the trajectory for a billion users—again, nothing really is—but it was more that you can feel it when something is growing."

He shared a concept he calls "energy in the system":

"I have this half-baked notion that I need to formalize at some point, which is: is there energy in the system or is there not energy in the system? With Artifact, we poured Blood, Sweat, and Tears into the product, into the product design. Kevin did so much deep work on machine learning and personalization... I think our recommendations were phenomenal, especially if you had enough data in the system. But the energy was not in the system."

Despite having dedicated users who loved the product (Mike mentioned that he still has people tell him they miss Artifact), the app wasn't achieving the growth trajectory they needed.

Their decision-making process was methodical rather than impulsive:

"It's not like we woke up one day and were like, 'This is it.' We were like, 'All right, we have a hypothesis of three more things that we think could move the needle here. Let's try them, let's give them a little time, let's reevaluate.' They set a new baseline for growth, but they didn't inflect it."

At that point, they faced the critical question: "Is it the thing we want to spend another year on?" The answer was no, in part because of opportunity cost:

"You can think of opportunity cost in this abstract way, like there's all these things I could be doing. In reality, I look now like, 'Man, I ended up having this crazy opportunity to join Anthropic and be CPO.' That window is a window... It's so cool that I got to go through that window, and that was only because we were able to be deliberate about Artifact."

Aditya strongly endorsed this decision-making approach:

"I've been in Silicon Valley for 20-plus years now. I can't think of too many people who have shut down a consumer app of their own accord at the stage that you guys did... I wish more people did that because it would allow them to go and do the next thing."

He articulated a core philosophy: "The earlier you can shut down something that doesn't have energy, the better it is for all involved... The worst state for a startup is when it's growing slowly. That is actually the worst spot. It's way better for a startup to be dead or to be growing fast."

Mike acknowledged that being self-funded made this decision easier:

"We were fortunate to be self-funded because it meant that decision was just between me and Kevin. It is much harder once you've raised, and it's especially harder if you've raised a lot."

Timestamp: [56:14-59:54]

• "Other-market fit" (founders genuinely loving using their own product) provides initial validation before broader market fit
• The decision to formalize a company often comes from emotional needs beyond just product potential - the desire to build something "real" with a team
• Following extraordinary success creates unique psychological pressures for founders embarking on second ventures
• Setting unrealistically high standards for a new product based on previous success can delay launch and limit valuable real-world feedback
• The metrics that matter at scale (hundreds of millions of users) become abstract and less personally meaningful than early milestones
• Daily enjoyment of the work itself provides more sustainable fulfillment than focusing solely on outcomes or scale
• Hands-on building and coding can be deeply satisfying for technical founders, especially after periods of primarily managing teams
• Successful founders develop an intuitive sense for whether a product has "energy in the system" - momentum that will lead to growth
• Methodical testing of improvement hypotheses before making shutdown decisions creates confidence in the final choice
• The "worst state" for a startup is slow growth - being definitively unsuccessful can be better than limbo
• Self-funding provides greater decision-making freedom, especially around pivoting or shutting down
• Recognizing opportunity cost is critical - shutting down a moderately successful venture can enable pursuit of more impactful opportunities
• Having a loyal core user base doesn't necessarily indicate a product has the growth potential needed for venture-scale success

Timestamp: [50:04-59:54]

People:

• Kevin - Kevin Systrom, Mike's co-founder at Instagram and Artifact
• Aditya - Podcast host, partner at South Park Commons

Companies/Organizations:

• Artifact - News personalization app founded by Mike and Kevin after Instagram
• Instagram - Social media platform referenced as Mike's previous success
• Anthropic - AI company where Mike is now Chief Product Officer (CPO)
• South Park Commons (SPC) - Aditya's organization, mentioned regarding startup philosophy

Products/Technologies:

• Machine learning - Core technology behind Artifact's personalization
• AI - Distinguished from machine learning as more recent technology
• iOS - Mobile platform Mike coded for during Artifact development

Concepts:

• "Other-market fit" - Founders genuinely enjoying using their own product
• "Energy in the system" - Mike's concept for detecting true product momentum
• Startup stages - "Minus one to zero" referenced regarding startup journey
• Self-funding - Approach Mike and Kevin took with Artifact
• Opportunity cost - Factor in deciding to shut down Artifact
• Zombie companies - Startups that continue without meaningful growth
• "The act of building" - Process that provides daily fulfillment
• "Inflection point" - Growth pattern sought but not achieved with Artifact

Product Features Mentioned:

• Collapsable comments - Feature initially missing from Instagram
• Like summarization - Feature initially missing from Instagram
• Number formatting (commas) - Initially missing from Instagram statistics
• Recommendation system - Core of Artifact's personalization technology

Timestamp: [50:04-59:54]

Mike and Aditya continued their discussion about shutting down startups, expanding to the broader venture ecosystem. Mike observed that many companies that raised large rounds in 2021 and early 2022 are now wrestling with difficult realities:

"I'm starting to see like, 'Well, how about we buy back all the shares?' There's some reconfiguration that's happening. In some cases, the companies are still really good—they just need a refactor. But in other cases, you're like, 'Man, is this ever going to be the company that you thought it was going to be?'"

He raised the ethical dimension of this question: "Is this good for the world and for you to keep working at it? Maybe not."

Aditya provocatively suggested that venture capitalists should be more supportive of founders moving on from struggling companies:

"I'm probably going to get in trouble with my VC colleagues for saying this, but I think they should actually be more supportive of letting these companies do something else, or actually just like returning some capital back."

He argued that this approach would be "healthy for the ecosystem" as a form of "creative destruction," challenging the conventional venture capital mindset that pushes founders to persist indefinitely with companies that have little chance of achieving their original vision.

This exchange highlights a tension in the startup ecosystem: while VCs typically want founders to continue pursuing growth at all costs once they've invested, both Mike and Aditya suggest that a more humane and ultimately more productive approach would allow for graceful conclusions and redeployment of talent and capital to more promising opportunities.

Timestamp: [1:00:00-1:00:40]

Aditya shifted the conversation to Mike's recent move to Anthropic as Chief Product Officer, praising Claude 3.5 as his "favorite model" that he uses "for everything." He noted the interesting contrast between Anthropic's technological maturity and product infancy:

"A company like Anthropic is both the result of many, many years of seminal work in terms of the math and the statistics and the actual machine learning theory, but also of distributed systems... but then on the product side it still feels like minus one right now, in my opinion. It's early days—like how do we actually use this thing?"

Mike described how Anthropic represented an unexpected career path that challenged his previous assumptions:

"I had talked myself into believing that the only way to have a Greenfield product opportunity is to start a company, which is kind of depressing because starting a company, as I've gotten to experience, is really hard. And I didn't want another company. I was like, 'Am I going to go and be a product person somewhere that already has a really mature product?'"

He had a specific set of criteria for his next role:

"I had missed from the Instagram days showing up and having eight things going on—again, the theme of our conversation, I like doing a lot of things at once. So I liked at Instagram that we had multiple product teams and multiple engineering teams doing a bunch of stuff."

Mike was looking for "a fairly mid-stage or growing company that needs me but has a lot of Greenfield product." He acknowledged that these requirements "narrowed it down to not very many companies in the world," making Anthropic a surprisingly perfect fit:

"It ended up working out so perfectly in that I came in and there's the beginnings of a product and product initiatives, and they'd hired product people, but I could come in and help drive things forward and bring all the things that I've learned from the consumer side, from the ML product set on Artifact, but also apply it to this company."

The opportunity to work on both Anthropic's API business and consumer products like Claude and Claude Pro aligned perfectly with Mike's interests and experience. This serendipitous match reinforced his earlier points about remaining open to opportunities:

"It goes to show, if you're having that FOMO, if you're thinking about restarting things—if you put yourself out there and you're having the conversations, those doors can appear, and then you can just be ready to go through them."

Timestamp: [1:00:41-1:03:35]

After eight weeks at Anthropic, Mike shared his early observations about Artifacts, a co-creation feature that allows Claude users to create documents, mini-games, and websites. He was quick to clarify:

"People love Artifacts. I promise there was no 'Oh, my thing was Artifact, let's call this thing Artifacts'—total coincidence."

What excited Mike most about this feature was how it fundamentally reimagines the concept of "no code" development:

"For all the talk and promise of low code and no code, it was still this abstraction of either you're working on top of some kind of workflow builder or drag-and-drop thing, or you're doing code but in this pseudo-code way. The thing that unlocked for me when I saw Artifacts is: this is truly no code."

He described his experience building within the system:

"I've built several things in Artifacts where I didn't even look at the code it generated. I don't actually know if it used React or Svelte or neither, or if it just pulled... I don't know. And if you're building large software, of course it matters. But if you're building these prototypes or demos or just web apps to help you build an intuition about something, how it's built doesn't matter, I think, at all."

Mike identified two key directions for evolving Artifacts:

1. Deeper data integration: "How do they start getting access to more data or more knowledge that they can go deeper on? Whether it's things you're thinking about, documents you've written... eventually data stores as well. Connecting that to your Google Docs would be really interesting—one of the ways it could actually start reading and writing for more than just this ephemeral kind of space."

2. Publishing and permanence: "What's the publishing story? Right now you can share them and you click a link and you're in a fully functioning web app that somebody built by just prompting the AI, but could you imagine them being something that has a little bit more permanence? Maybe inside a company: 'Hey, I prompted Claude to pull a list of our top 10 customers every day, and I didn't have to bug our BI person.'"

This vision extends to how these creations might be shared and improved within organizations: "How do I interact with that internally, and how do people remix it?"

Mike observed that this external creativity has a profound internal effect as well:

"It's leading to a lot of really cool creativity outside, and then that's the case—I learned this with Instagram—it breeds a lot of creativity inside too, because you're seeing all these use cases that are now possible that weren't possible, you know, three weeks ago."

Timestamp: [1:03:36-1:06:04]

Aditya built on Mike's observations about Artifacts with an insightful critique of how low-code/no-code tools have historically evolved:

"One of the big weird turns we took in low-code/no-code was to make it look more like software engineering, whereas in reality, low-code/no-code is more like messy Excel stuff. It's just like you use it, you throw it away. It can be messy, it should be dirty, and you don't really care that much."

This perspective challenges the conventional wisdom that has driven tool development in this space, suggesting that the best approachable creation tools shouldn't try to mimic professional software development practices, but instead embrace the quick, disposable nature of everyday problem-solving.

Aditya articulated why AI-powered creation tools like Artifacts feel so revolutionary:

"As we use these—whether it's Artifacts or some kind of similar system—you can express yourself quickly without caring what actually happens underneath, and you're comfortable throwing it away because you know you can build it again easily. I think that makes a huge difference."

Mike agreed, while acknowledging current limitations:

"There's no technical debt, it's getting recreated. There's limitations, of course. I'm not saying that you're going to build a software suite on that, at least not right now, but it points the way to something really interesting."

This exchange highlights a fundamental shift in creation paradigms: from carefully crafting programs that must be maintained over time to quickly expressing ideas through AI that can be easily regenerated or modified as needed. The conversation suggests that AI may finally deliver on the long-promised democratization of software creation by embracing impermanence and regeneration rather than trying to make everyone into software engineers.

Timestamp: [1:06:05-1:06:53]

As the conversation drew to a close, Aditya offered a heartfelt reflection on Mike's career journey and approach to decision-making:

"I'm just struck by how many times you have operated from a place of both conviction but also vulnerability and courage. Thank you for being so upfront with what your journey has been. I hope that listening to some of the decisions you have made and how you've approached them will inspire people out there as they do their own minus one journey."

This observation encapsulates a central theme of the conversation: that navigating pivotal career moments requires both the strength to make difficult choices and the vulnerability to acknowledge uncertainty and learn from experiences.

Mike expressed appreciation for the conversation while reflecting on his motivation for sharing so openly:

"I feel like we could talk for another two hours and have lots to talk about as well, but I was trying to channel the things I wish I had heard when I was making these decisions. At the time, there was nothing really out there to help guide me through it."

This final exchange highlights the value of sharing authentic career stories, especially around challenging transitions and decisions. Mike's willingness to discuss not just his successes but also his doubts, fears, and recalibrations provides a more complete picture of entrepreneurial journeys than typically shared.

The conversation concludes having provided listeners with a rare window into the emotional and intellectual processes behind major career transitions, offering guidance for others facing their own "minus one" moments.

Timestamp: [1:06:54-1:07:46]

• Venture capitalists should potentially be more supportive of gracefully winding down companies that aren't achieving their potential, returning capital when appropriate
• "Creative destruction" in the startup ecosystem allows talent and resources to be reallocated to more promising opportunities
• Finding the right next role after major success may involve challenging assumptions about what career paths are available
• The ideal role combines elements that energize you (like having multiple simultaneous challenges) with opportunities to apply your unique experiences
• Being open to unexpected opportunities is crucial - "those doors can appear, and then you can just be ready to go through them"
• AI-powered creation tools like Anthropic's Artifacts represent a fundamentally different approach to "no code" development
• True "no code" means not needing to understand the underlying implementation at all, focusing purely on expressing intent
• Future AI creation tools will likely evolve toward deeper data integration and more permanent, sharable artifacts
• Traditional low-code/no-code tools erroneously tried to make everyone into software engineers instead of embracing quick, disposable creation
• The regenerative nature of AI-created tools eliminates technical debt concerns - you can always recreate or modify as needed
• External user creativity drives internal product development by surfacing unexpected use cases and possibilities
• Navigating career transitions requires both conviction to make difficult decisions and vulnerability to acknowledge uncertainty
• Sharing authentic career stories, including doubts and fears, provides valuable guidance for others facing similar transitions

Timestamp: [1:00:00-1:07:46]

People:

• Mike - Mike Krieger, former Instagram/Artifact co-founder, now CPO at Anthropic
• Aditya - Podcast host, partner at South Park Commons
• VCs - Venture capitalists mentioned in context of startup closures

Companies/Organizations:

• Anthropic - AI company where Mike is now Chief Product Officer
• Instagram - Referenced regarding Mike's experience with multiple product teams
• Artifact - Mike's previous company, noted as distinct from Anthropic's "Artifacts" feature

Products/Technologies:

• Claude 3.5 - Anthropic's AI model mentioned as Aditya's favorite
• Artifacts - Anthropic's co-creation feature for building documents, mini-games, and websites
• Google Docs - Mentioned as potential integration point for Artifacts
• React - Frontend framework possibly used by Artifacts (Mike wasn't sure)
• Svelte - Frontend framework possibly used by Artifacts (Mike wasn't sure)
• Excel - Referenced as comparison for messy but functional tool creation
• Low-code/no-code tools - General category discussed regarding development paradigms

Concepts:

• Creative destruction - Economic concept referenced regarding shutting down startups
• Greenfield product - Term for new product areas without established solutions
• FOMO (Fear of Missing Out) - Mentioned regarding career opportunities
• Technical debt - Software engineering concept noted as absent in AI-generated code
• "Minus one" - Podcast theme about the decision phase before starting something new
• API business - Referenced as part of Anthropic's offerings
• BI (Business Intelligence) - Mentioned in context of data access within organizations

Features/Capabilities:

• Co-creation - Core capability of Artifacts feature
• Sharing/publishing - Current and potential capability of Artifacts
• Data integration - Future direction for Artifacts development
• Remixing - Potential capability for collaborative iteration on Artifacts

Timestamp: [1:00:00-1:07:46]