a16z

From fork to frontier in four years

Martin Casado and Matt Bornstein on the Cursor team's tremendous clarity and conviction behind every decision: Martin: "They were very product-focused, which you actually don't see a lot in AI. They really believed that you can build a product, and if it's really good, it'll be adopted." "That dictated a lot of their decisions. If you really believe in the product, you would never do a plugin because then you're part of somebody else's product." "There was... tremendous clarity on exactly what type of company they were, and the decisions all fell from that consistently." "Many of the companies we saw were pastiches of 'We'll try all of these different things, we're not quite sure what's gonna work.'" Matt: "They had the ambition and the courage to do the maximal form of what the product would look like." "As founders, we all want to be ambitious and push to the maximum point of the Pareto frontier, but you have to know what the curve is. And they knew what the curve was and didn't get distracted by the non-product things." @martin_casado @BornsteinMatt

Vídeo

a16z GPs Martin Casado and Matt Bornstein on the trait Cursor and Elon share and why it makes them a great fit: Martin: "Probably the greatest strength of the company is that once they made a decision, it doesn't matter how hard that was, whether it's deciding not to enter a market or deciding to be very limited in what they're doing, they follow through." "Kudos to them for actually having the courage and the tenacity for these tough decisions." Matt: "Which is sort of an Elon-ish trait too." Martin: "These two companies are very good fits on a technical level. Elon has the compute, they have the distribution and the data. On a vision level, they both think that code is the path to changing all of compute and maybe all of humanity." "But also on a culture level. They're very product-engineering focused, very product-engineering heavy, very fast-iterating. They make hard decisions very quickly... This is probably the best fit I've ever seen as far as just the full package." @martin_casado @BornsteinMatt

Vídeo

a16z GP Sarah Wang on what Michael Truell said when she asked him about Claude Code: "He said... 'Look, we are going after the biggest market in the world. There's going to be competitors. In fact, there's been a rotating cast of characters from the very beginning. Microsoft Copilot was the first one. Claude Code is one of a rotating set of characters. We think in big markets, you're always going to have formidable competitors. That does not scare us.'" "I just remember being somewhat blown away by the clarity of that comment, the lack of fear. There's humility as well, like it's humility mixed with bravado, and I just think that is kind of the winning combo when you're going after these juggernauts like that." @sarahdingwang

Vídeo

a16z GP Matt Bornstein on the signal that made the Cursor team stand out and the pitch meeting with Michael Truell: "The Cursor team kind of stood out because not only did they align with the things that we thought about the world, more importantly, smarter people than us." "Engineers working at the time, at the leading AI companies like OpenAI, Midjourney, Replicate, and a bunch of others all kind of use this and subscribe to this model." "The team frankly just seemed kind of special. They had an unusual degree of focus." "I remember we got Michael to come pitch our GP group, and 90% of the meeting was him saying no to things... Michael literally just sat there and listened to all the questions and said, 'Hmm, interesting. No.'" @BornsteinMatt

Vídeo

a16z's Martin Casado, Matt Bornstein, and Sarah Wang on the Cursor story: In 2023, four MIT dropouts released a fork of Microsoft's code editor. Microsoft owned VS Code, GitHub, the OpenAI weights, and had the best enterprise distribution in software history. And despite this, Cursor won. A rotating cast of competitors were supposed to crush Cursor. The company blew past every forecast it gave its investors. In this episode, Martin, Matt, and Sarah break down how it happened: why Cursor initially refused to train its own model, how they kept evolving their own product three times in two years, and how they built the fastest-growing sales team any of them has ever seen. 00:00 Intro 01:19 AI coding before agents 03:30 "Code in the future will look like pseudocode" 04:55 The pitch meeting that was 90% "no" 06:14 Why Cursor didn't build its own model 08:15 What Michael Truell does for fun 11:24 The growth assumptions Cursor broke 12:13 Competing with Microsoft 14:45 At dinner with John Schulman 15:40 The competitors that couldn't kill Cursor 17:33 Michael Truell on Claude Code 19:14 Cannibalizing yourself 3x in two years 20:09 Why the gross margin critics were wrong 22:08 The sales team they never wanted 25:40 Hiring AEs like engineers 31:00 How their office suits their product 33:19 Why Cursor and Elon fit 35:15 How Cursor got good at M&A YouTube: t.co/au6VSPj6Yf @martin_casado @BornsteinMatt @sarahdingwang

Vídeo

Most AI apps are pricing at the wrong layer. Asked how they'd rather be billed, technical AI buyers preferred credits tied to recognizable value over tokens nearly 2 to 1. Per-token pricing anchors your product to a cost curve that keeps falling. Full piece from Tugce Erten and Sarah Wang on why customers want legibility, not tokens: t.co/Lwmb3SCkWe

AI power users are here and the case for agent-friendly software is materializing. Aaron Levie on agents outnumbering people 1000x: "If you start to imagine that we all have to build software for agents… we spend as much time now thinking about the agent interface to our tool as we do the human interface." "If you have a hundred or a thousand times more agents than people, then your software has to be built for agents… it's going to be through an API or a CLI or MCP." "What if you give a coding agent access to your SaaS tools and a coding agent access to your knowledge work, workflows and context?" "It can actually code its way or use APIs through whatever task it's trying to achieve." "That appears to be like a paradigm that is starting to compound… and I actually think it kind of makes sense as the ultimate manifestation of this stuff." @levie

Vídeo

Vibe shift. ETF sectors with the highest inflow: In 2020: Clean energy, innovation, healthcare, cloud, emerging markets tech In 2026: Artificial intelligence, infrastructure, defense, space, nuclear

"What if we're insufficiently optimistic?" Anish Acharya's case in eight parts, from a conversation with Jen Kha on AI: 1. Macro (03:34) - Prices for last-gen GPUs are RISING on a per-hour basis. That never happens in computing. It points to essentially infinite demand for intelligence and highly constrained supply. 2. Moats (05:30) - Network effects, scale, and brand are as good as they've ever been. "No amount of coding agents is gonna make Nike not Nike." The vulnerability is integration, the moat built on being painful to migrate away from. 3. Tokens (06:42) - For unbounded-upside work like sales and product, it's rational to pay almost any price for a model even one IQ point smarter. 4. Labs (10:37) - They're vertically integrating down into inference, not up into apps. Inference is one homogeneous workload at enormous scale. The app layer is a thousand idiosyncrasies of pricing and packaging. 5. Models (11:15) - They're not commodities. Some are literal and precise, some are creative, and organizations will need both. 6. Agents (15:20) - An agent is just a model in a loop with tools and memory. Coding loops already fix reported bugs end to end. Business loops come next, think the model that says "We need to open a branch in Tijuana." 7. Consumer (17:20) - We're in the DOS era of AI, so there's no app store for it yet. But the 99-cent app era is over, people are paying $200 a month, and the Birkin bag of software is coming. 8. Builders (35:05) - New business formation is skyrocketing. The 25-year-old who would've been a YouTube creator is now building software for their neighborhood: a mom-and-pop SaaS economy. YouTube: t.co/0aslotMCxU @illscience @jkhamehl

Vídeo

Airbnb HQ, San Francisco, 2008

Americans reported losing $3.5 billion to imposter scams last year. Impersonation is an entire industry. Despite brand protection as a growing software category, the scams are still winning. Per Doppel's data: - Brands catch less than 10% of the fakes targeting them - 94% catch fewer than half - The median brand catches zero Full piece from Doppel co-founder and CEO Kevin Tian on the business of online impersonation: t.co/MBJlI9idiJ

Steven Sinofsky says the culture of a big company is as fixed as a law of physics, and that's why they don't crush startups: "It's weird that they teach disruption as a theory in the business school, when really it should just be a fact in the physics department." "You always think, 'Oh my God, we're just gonna crush all of these little companies.' You always think that when you're at the big company, and then you realize they never get crushed." "The startups don't aim straight at the incumbents. And the incumbents just don't pay attention... Microsoft is worried way more about what Amazon and Google are doing." "The elements of disruption that matter are the cultural ones of being a big company, and those are constant. Those are the laws of physics." "We're at a magic moment... you don't have to go build a data center and build your own egress and call AT&T. Now you're up and running in the first hours of your first dinner." @stevesi

Vídeo

Martin Casado says startups are growing at meteoric rates because AI has leveled the playing field with incumbents like Microsoft and Meta: "Six months ago you'd have asked this question, what advantages do incumbents have? They have the same advantage all incumbents always have. They have the capital, and they have the cash flow, and they have distribution." "What's crazy is AI solves the distribution problem. It just solves the demand problem. And these companies are able to raise so much money that they're actually on competitive footing with the Microsofts and the Metas." "We're in a very new territory when it comes to the new challengers versus the incumbents, specifically for these two reasons." "In the past, if you had a company and you wanted to get people to use your stuff, it was hard... But the demand is so unlimited for tokens and for GPUs, literally you can just decide how much money you're putting into it in order to drive top-of-funnel and growth." "The things that have been typically hard for startups are very much easier now. And I think this is why we're seeing such meteoric growth of the Cursors, the Anthropics, and the OpenAIs." @martin_casado @eriktorenberg

Vídeo

Five years ago, @patrickc asked @sama whether raising colossal amounts of money right out of the gate was underrated. Martin Casado says we now know the answer: "Prior to AI, there was always this battle between Eric Ries and Ben Horowitz, The Lean Startup, and then Marc and Ben wrote The Case for the Fat Startup, which basically argued raise the money and go for it." "But there's always been this natural limiter, actually, which is engineering." "Patrick Collison is right. We now have a discipline for taking a lot of money with small teams and using it productively. That's a very, very big change. I don't think we've internalized it." "There's been this view in venture, this zero-sum thinking... You're a venture capitalist, don't you believe in positive-sum stuff? If you look at the numbers, the more capital that flows into private markets, the larger the market gets." @martin_casado @stevesi @eriktorenberg

Vídeo

Martin Casado explains AI as a fundamentally new abstraction layer of computing: "I don't think in the history of computer science that I can recall have we ever abdicated actual reasoning or logic. It's always been a resource. It's been like compute, network, and storage..." "And then the human is putting in the high-level thing, and it's using the compute, network, and storage to calculate the answer." "But now I feel like you're actually abdicating thinking in a way, where you're like, 'Tell me the answer.' Like, 'I'm not even really sure what the question is.'" "This is the first time it feels like a different layer of the stack. Maybe this really is the next abstraction, which is more of a human-level abstraction, which doesn't map directly and so is actually different... You kind of pray to the model god in the right words, and then it produces the answer that just ends up being useful." @martin_casado

Vídeo

Martin Casado spent years dismissing the AI doom scenario and says what he got wrong was how long the scaling laws would keep holding: "I was responding to this Bostrom notion of recursive self-improvement, fast takeoff. You create one of these things, you step back, and it takes over the world... That's clearly not what's happening." "But here's what I got wrong. What I got wrong is I did not know that we could effectively just continue to pour money in this. The scaling laws are holding." "I don't know what it means to do a $100 billion training run, to have this thing that you're putting $100 billion in, and then that money comes from this meta-economic machinery that may want to solve whatever." "They may want to solve cancer, but they may also want to create a weapon. Like, who knows?" "This concentration of this many resources in a useful way, I think, is very new. I don't think we understand the implications. I think you could reasonably argue that that's very dangerous if you apply that $100 billion in the wrong way... What does it mean to be able to concentrate resources?" @martin_casado

Vídeo

Martin Casado and Steven Sinofsky on why AI is turning engineering problems into capital problems: Martin: "20 years ago, if you're a startup of 10 people and I gave you a billion dollars, what would you do with it?... In software, you hire people and then there's nothing you could do. The Mythical Man-Month is very real." "Right now, if I give 20 people a billion dollars, they can actually use it usefully... We've never been like that before." "This is like a law of physics where our early intuition, which is like all problems are engineering problems, starts to change... It changes the nature of capital versus innovation versus competition versus defensibility." Steven: "Computing was capital-bound for the first 30 or 40 years. If you wanted to do something with a computer, your first step was we have to get one, and then you couldn't. You were capital-bound, and then you were engineering-bound, and now we're capital-bound again." "Mad Men goes through the scenario where the computer shows up at the advertising agency... They couldn't figure out what to do, but they were excited that they had the capital to acquire one, and it made them look like they knew what they were doing." @martin_casado @stevesi

Vídeo

Your startup intuitions were trained on a world that no longer exists. Steven Sinofsky and Martin Casado have watched computing flip from an engineering-bound field to a capital-bound industry. Twenty people can now put a billion dollars to work productively, AI solves the distribution problem that kept startups small, and challengers sit on a level playing field with Microsoft and Meta for the first time. With Erik Torenberg, they get into why some mathematicians are cheering on their own automation, every AI panic that already happened in past eras of computing, and why nobody can predict the capabilities of a model built with $20 billion. 00:55 Why mathematicians love being automated 02:45 Is AI math worth any money? 08:50 The first proof humans couldn't check 14:35 Computing before electricity 19:50 How IBM explained computers in 1953 26:00 The failed computer that birthed the web 27:50 When Harvard banned computers from exams 30:20 Is AI just another abstraction layer? 38:15 When 20 people can spend $1B productively 42:50 The zero-sum VC myth 46:25 Disruption is physics, not business school 53:25 The chip Intel called a printer part 55:05 What a $20B model can do 59:45 What Martin got wrong about AI risk YouTube: t.co/AeuevOh1JK @stevesi @martin_casado @eriktorenberg

Vídeo

Most consensus startup intuitions are from a world that no longer exists. In this conversation, Steven Sinofsky, Martin Casado, and Erik Torenberg sit down to cover computing's new laws of physics: - 20 people can now put $1 billion to work productively - With AI, startup growth becomes a spending decision - Challengers can sit on a level playing field with giants like Microsoft and Meta - Computers are trusted with reasoning itself, no longer just labor - Nobody can predict what $20 billion of compute becomes 00:00 Intro 00:55 How mathematicians are reacting to being automated 02:45 Is AI math worth any money? 08:50 The first proof humans couldn't check 14:35 Computing before electricity 19:50 How IBM explained computers in 1953 26:00 The failed computer that birthed the web 27:50 When Harvard banned computers from exams 30:20 Is AI just another abstraction layer? 38:15 When 20 people can spend $1B productively 42:50 The zero-sum VC myth 46:25 Disruption is physics, not business school 53:25 The chip Intel called a printer part 55:05 What a $20B model can do 59:45 What Martin got wrong about AI risk YouTube: t.co/AeuevOh1JK @stevesi @martin_casado @eriktorenberg

Vídeo

.@bhorowitz on why he refused the Databricks founders' $200K ask and wrote them a $10M check instead: "So there were six PhD students, and Ion Stoica, who was their professor." "What they had was this thing called Spark, and the competitor was something called Hadoop. Hadoop had very well-funded companies already running towards it, and Spark was open-source. So the clock was ticking. I think they didn't quite know what they had." "Professors in general, it's a pretty big win if you start a company and you make $50 million. You're a hero on campus... So I'm always a little nervous about a company that comes out of academia thinking too small." "So I said 'I'm not gonna write you a check for $200,000. I'll write you a check for $10 million.'" "You need to build a company. You need to really go for it if you're gonna do this, otherwise you guys should stay in school." Ben Horowitz w/ @lennysan on Lenny's Podcast (2025)

Vídeo

.@sama on how to pick high-risk bets, from AGI in 2015 to the researchers he hires today: "When we started, people thought it was totally unlikely or almost impossible that AGI was possible." "We just got hammered by all of the intellectual giants of the field for saying that we were going after AGI." "High-risk bets are okay as long as you take the ones where if they work, it's super valuable." "And research looks this way. The kind of people that make great researchers are sort of non-consensus, fresh approach, high energy, non-standard people." "You don't want to fund a founder who has a very slightly different take on the same idea as the last thousand people you talked to." "It's very clear when you have someone who just thinks differently and is willing to stand by convictions that are very unpopular. May well be wrong, but if right, at least they're gonna be really right." Sam Altman w/ @davidsenra

Vídeo

Make it exist first, make it good later

Frédéric Renken and Steijn Pelle during the year they spent doing manual back office work at doctors' offices, before starting @LassieAI to automate it Full profile: a16zbuild.substack.com/p/how-steijn-p…

AI use by both patients and clinicians inflected shortly after the launch of ChatGPT Protege CEO Bobby Samuels on the Oracle Problem in health AI and how to solve it: a16z.news/p/the-oracle-p…

AI use by patients and clinicians inflected shortly after the launch of ChatGPT Protege CEO Bobby Samuels the Oracle Problem in health AI and how to solve it: a16z.news/p/the-oracle-p…

AI power users are showing up outside tech. The fastest-growing Codex adopters since February: Legal: 108x Sales: 41x Recruiting: 41x Marketing: 26x Healthcare: 24x Charts of the Week: a16z.news/p/charts-of-th…

Humans are the minority user of AI Agents burn nearly 5x the tokens people do, up 14x since February Charts of the Week: a16z.news/p/charts-of-th…

Martin Casado says Cursor's founders spend 30-40% of their time on hiring and culture, and that focus is how they iterate so fast: "A lot of these AI companies that I work with are so research-heavy. If you're not working on a new model architecture, it's kind of hard to actually interest the core teams." "They always kept the main thing the main thing, which is changing how you write software. They believed fundamentally it was a product problem... The model's very, very important, but it was a product problem." "There's actually not a lot of companies that are actually focused on product... There's a lot of research companies... But they were like, 'We are a product company.' And I think that's actually a key differentiator in this era." @martin_casado w/ @sophiadew and @theojaffee on MTS

Vídeo

Martin Casado says capital used to be a lottery ticket, and AI made it a vending machine: "If I'm training a model and I want it to be good at X, I can create an RL environment of X, or I can go pay someone to answer questions for X, and I can turn capital into being good at that X." "The problem is you don't know what gets worse... Let's say you put in $10 to do this. I don't know if you get $9 back on the other side of that." "What we've never been able to do in the history of this industry is put in $10 and get anything back." "It was literally put in $10, engineer, engineer, engineer, wait two years, probably screw stuff up, it'll probably fail. But maybe on the other side you'll have a product you can monetize." "Now it really is $10 in and then some amount out, pretty directly." @martin_casado w/ @sophiadew and @theojaffee on MTS

Vídeo

Martin Casado says AI has made capital more productive than ever before: "One of the very famous models, probably one of the most popular models, was built with a team of about 20 people. And I would say the cost of that was probably $2 billion plus." "In the history of humanity, in the history of engineering efforts, we've never been able to have 20 people productively use $2 billion." "We like to look at this wave in the context of technical sophistication, but one of the major stories is that we're able to apply large amounts of money productively, in short amounts of time, to whatever problem we're trying to solve." @martin_casado w/ @sophiadew and @theojaffee on MTS

Vídeo

It's time to build mega data centers. In counties where they're already operational, the picture is uniformly better since 2024: - More housing - Higher home values - Less unemployment - More job growth Charts of the Week: a16z.news/p/charts-of-th…

Same titles, different buildings. Data centers pay a premium across the job board: - Facilities managers: +64% - Network technicians: +42% - Construction managers: +29% Charts of the Week: a16z.news/p/charts-of-th…

Texas builds. Its ~14 GW under construction, the most in America, is just ~15% of the state's private non-residential spending. In New Mexico and Wyoming, data centers might as well be the construction industry: ~60% of spending, on just 1-2 GW. Charts of the Week: t.co/vsjTAG5o2X

a16z GP Angela Strange on how Borderless Founders accelerate preferential attachment through talent and customer power laws: "I would argue Borderless Founders have more levers at their disposal." "So if you start with talent, this is the fiercest talent market I've seen in my career as an investor, and also as an operator." "We have an AI insurance company in Brazil, and so you might think, 'Oh, the best talent is at the top most well-known universities there.' Actually, for AI talent, it's at a university that you haven't heard of." "And so if you are a company that gets 10 incredibly smart people really early, the 11th that you might be trying to recruit from the US that has 100 other offers meets your team, is like, 'Holy shit, how do they get people that are that good there?' And that starts the talent flywheel." "On the customer side, anytime you want to land a large customer, the first question they're going to ask is, 'Who else is using you that's in my industry at my size?'" "Nobody wants to be the first bank or the first insurance company. But via Borderless Founder network, you can get a large bank and a large insurance company in another country much more quickly." @astrange @VirtualElena

Vídeo

Over 40% of the a16z Apps team's investments over the last two years went to international founders. Half are headquartered outside the US. Angela Strange and Gabriel Vasquez own the bet behind that number, a bet that the best founders can come from anywhere. In Poland, ElevenLabs became the national AI champion. In Spain, Supersonik landed Salesforce as its first design partner. In Colombia, Angela's first check went to Addi, which now serves a quarter of the country and recruits Capital One's best credit talent to Bogotá. In this episode, they sit down to discuss the playbook: why AI opens every market but concentrates the epicenter in the Bay Area, and why country diasporas beat elite alumni networks. 00:00 Intro 00:54 From a WhatsApp group to a global strategy 05:29 Why AI pulls founders to the Bay Area 10:29 Defining the borderless founder 12:29 How diaspora networks help companies scale 18:11 The three advantages of borderless founders 22:36 Preferential attachment across borders 25:11 The bridge to Silicon Valley works both ways 27:11 Mapping and building global ecosystems 33:15 Backing repeat founders 40:31 Silicon Valley speed and global ambition t.co/t3kDg0dned @astrange @GEVS94 @VirtualElena

Vídeo

Alex Atallah, OpenRouter co-founder and CEO, featured on the front page of The Stanford Daily in 2011 after releasing his campus social network "Dormlink"

While competitors chased a trillion-dollar live shopping market, @Whatnot noticed its own customers turning livestreams into storefronts. Co-founder and CEO Grant LaFontaine on what users really care about: "They care about whether they're gonna enjoy the thing, whether it provides value." "There's this counterintuitive thing, which is if you try and build a market, it oftentimes distracts from: are you building a really great user experience?" "We were building our Funko Pop marketplace, talking to our customers every single day, and then we saw our customers hacking social media by selling items in live video. But the experience was really broken. Transactions weren't integrated well, shipping wasn't, payments, discovery." "We just thought we could build a better one of those... Logan locked himself in his room for a few weeks, came out with the first version of the product. I went live, sold a lot of items, and the rest is history." @GrantLaFontaine @DavidGeorge83

Vídeo

In third grade, Grant LaFontaine sold a Pokemon card online. The buyer mailed a money order; Grant cashed it at the post office and then shipped the card to the customer. Today he's co-founder and CEO of @Whatnot, where the largest businesses do over $100 million a year at up to 40% margins. Grant joins a16z's David George to discuss how that first transaction became a platform that did $8 billion in sales last year, why live commerce is a third of all commerce in China yet still emerging in the US, and how Whatnot operates reliably at scale, and more. 00:00 Intro 01:05 The Pokemon card and the money order 06:25 Half a Bitcoin for a domain 09:20 Why not knowing the market was an advantage 11:05 Customers don't care about your market 12:45 Most people don't buy anything 17:15 Why the eBay comparison misses 20:05 95 minutes a day on a shopping app 25:00 Sellers doing $100M with 40% margins 26:35 Fresh fish from the San Diego dock 32:25 Running a police force for two New York Cities 37:00 Why Whatnot won't do AI avatars 40:15 Cars on Whatnot @GrantLaFontaine @DavidGeorge83

Vídeo

Stripe’s Will Gaybrick: "We just want to make sure that moving between tokens and dollars is as seamless and safe as moving between dollars and euros. We're at the beginning of this journey, but we think it's going to be a big part of the future of Stripe." @gaybrick

Vídeo

"Having established themselves as the performant, neutral layer for model routing, OpenRouter now has a pretty great road map ahead of them. Tools like Ori Eval let OpenRouter figure out for you, based on progressively smarter understanding of your prompts and your business needs, how to dynamically route you the best possible models and providers, on a second-to-second basis. And as the AI economy scales outward, and the token flow expands from 'models to apps' and into a network of companies exchanging value through value-added tokens, OpenRouter becomes a genuine network of businesses exchanging intelligence, just as Stripe has become the preferred network of businesses exchanging dollars." OpenRouter & Stripe: The Intelligence Network, by @martin_casado: t.co/3uyIFUYc2H

"Stripe has signed a deal to acquire OpenRouter. Together, they become the trusted, scaled, and performant network where the world’s AI companies exchange intelligence. This is a good thing for everybody, and it comes not a moment too soon. Model companies (both closed and open source) are in a generational Red Queen’s race of getting smarter and faster. Meanwhile, application companies are experiencing a Cambrian explosion of possibility, if only the infrastructure can keep up. OpenRouter and Stripe, working together, cements a pillar of AI’s economic infrastructure in a way where everyone wins big." "I have never seen anything in my career like what I’ve seen in the past 24 months. The sheer amount of value creation, in matters of months, by companies like OpenRouter is going to go in the economic history books. OpenRouter, A Stripe Company, is genuinely just scratching the surface of its potential as one of the network exchange layers of the internet. Kudos to Stripe for creating the perfect place for them to accelerate into the next leg of their journey." OpenRouter & Stripe: The Intelligence Network, by @martin_casado: t.co/3uyIFUYc2H

In third grade, Grant LaFontaine sold a Pokemon card online. The buyer mailed a money order; Grant cashed it at the post office and shipped the card. Today he's co-founder and CEO of @Whatnot, where the largest businesses do over $100 million a year at up to 40% margins. In this conversation, Grant joins a16z's David George to discuss how that first transaction became a platform that did $8 billion in sales last year, why live commerce is a third of all commerce in China yet still emerging in the US, how Whatnot stays reliable at scale, and more. Intro The Pokemon card and the money order Half a Bitcoin for a domain Why not knowing the market was an advantage Customers don't care about your market Most people don't buy anything Why the eBay comparison misses 95 minutes a day on a shopping app Sellers doing $100M with 40% margins Fresh fish from the San Diego dock Running a police force for two New York Cities Why Whatnot won't do AI avatars Cars on Whatnot @GrantLaFontaine @DavidGeorge83

Vídeo

In third grade, Grant LaFontaine sold a Pokemon card online. The buyer mailed a money order; Grant cashed it at the post office and shipped the card. Today he's co-founder and CEO of @Whatnot, where the largest businesses do over $100 million a year at up to 40% margins, and where users spend enough time to put the platform on par with social networks. In this conversation, Grant joins a16z's David George to discuss how that first transaction became a platform that did $8 billion in sales last year, why live commerce is a third of all commerce in China yet still emerging in the US, how Whatnot stays reliable at scale, and more. 00:00 Intro 01:05 The Pokemon card and the money order 06:25 Half a Bitcoin for a domain 09:20 Why not knowing the market was an advantage 11:05 Customers don't care about your market 12:45 Most people don't buy anything 17:15 Why the eBay comparison misses 20:05 95 minutes a day on a shopping app 25:00 Sellers doing $100M with 40% margins 26:35 Fresh fish from the San Diego dock 32:25 Running a police force for two New York Cities 37:00 Why Whatnot won't do AI avatars 40:15 Cars on Whatnot @GrantLaFontaine @DavidGeorge83

Vídeo

Travis Kalanick on why he ran Uber so close to the line: "Every decision I made at Uber, I defend to this day. I wasn't 100% correct, but always good intentions and always coming from the right place." "One of my lawyers made a sports analogy, 'Do you have chalk on your shoe?' And I'm like, 'Never had chalk on my shoe.' But in order to know, you would need an electromagnetic scanning microscope with reverse angle slow-mo replay to verify." "The problem was I ran too close to the line in too many situations. When you are big and important, the scrutiny and the expectation is that you don't run that close to the line, even if it's correct." "It comes from what I did before Uber... A really hard startup, first four years, no salary, ran out of money several times, super grind, lose all friends. It was so hard that I had to be epically precise and hardcore just to pay the bills and go to the grocery store next week." "That precision and intensity made Uber what it was, but I was running a $70 billion company the way somebody who thought he was gonna starve next week would run it." @travisk w/ @davidsenra

Vídeo

Tuesday at SpaceX

It's been 5 weeks since we learned about the OpenAI/Hugging Face breach, which revealed an awkward reality: defenders have to ask models the same questions attackers do. In this conversation, Cotool CEO Max Pollard and Neo CEO Nick Warner sit down with a16z's Joel de la Garza at Black Hat to discuss how security tools were designed to stop people or malware (and how AI agents are neither), what breaks when half of enterprise software goes agentic, and how teams are routing around cyber refusals. 00:00 Intro 01:00 Models escaping containment & the Hugging Face breach 01:50 Why models refuse to help the good guys 05:45 Built to stop people and malware, but agents are neither 06:45 "The end justifies the means, in the mind of the model" 10:45 50% of enterprise apps agentic before 2027 14:20 When your honeypot becomes a false positive machine 15:25 Signatures are dead, and so is behavioral detection 18:55 Defending AI and defending from AI @maxpollard415 @cotoolai @neo_ai_security

Vídeo

Curiosity compounds

Travis Kalanick says founders should date ideas before committing to them: "I have lots of ideas all the time. I'm an idea factory. But other people have great ideas, too." "I like to say, you go out on a date with the idea. Was it a good date? Did it go well? How did you and the idea get along?" "Who you are is going to be a big part of which idea works for you... Be in touch with who you are, and then when the right idea comes your way, you just know." @travisk w/ @davidsenra

Vídeo

Travis Kalanick says a leader's job is to find the line between order and chaos: "On one side is order. You have lots of rules, lots of structure, lots of process, and eventually lots of bureaucracy." "If you go to the other side of the line, which is chaos, lack of rules, lack of process, lack of structure... you also get to a place where you're going slow and people are bummed." "That line between order and chaos is innovation at speed and at scale, and the job of every leader is to find that line. And it's not in two dimensions. It's like in eighty dimensions." "Go back to launch at Uber... There's a bunch of 23-year-olds launching cities... Nothing would actually launch until it got to a pricing call... It could be eight, ten hours of pricing calls before we launched London." "I never solved the same problem twice. Once we solved it, it became part of the playbook, and eventually we get to city 20 and that pricing call takes five minutes. I stopped going to them." "So now I've got one rule, which is that pricing call. The fewest number of rules while staying out of chaos. If you didn't have that one rule, you now have 23-year-olds running around doing crazy shit." @travisk w/ @davidsenra

Vídeo

.@cdixon on "Climbing the wrong hill":

A common thesis is software gets severely commoditized from here on out. Stripe's Will Gaybrick says their data shows the opposite: "It's plausible. We're either in the singularity or creeping towards the singularity, and it's very hard to estimate what a future looks like where models are recursively generating models." "But we are seeing the exact opposite right now, where software creation is exploding. Customers are monetizing faster than ever." "Our 2026 cohort is 50% larger and growing faster than the 2025 cohort. That one being 70% larger and growing faster than the 2024 cohort." @gaybrick @stripe @DavidGeorge83

Vídeo

A common thesis is software gets severely commoditized from here on out. Stripe's Will Gaybrick says their data shows the opposite: "It's plausible. We're either in the singularity or creeping towards the singularity, and it's very hard to estimate what a future looks like where models are recursively generating models." "But we are seeing the exact opposite right now, where software creation is exploding. Customers are monetizing faster than ever." "Our 2026 cohort is 50% larger and growing faster than the 2025 cohort. That one being 70% larger and growing faster than the 2024 cohort." @gaybrick @stripe @DavidGeorge83

Vídeo

Stripe's Will Gaybrick says AI agents will navigate the internet like hummingbirds, and that's what will make microtransactions work: "If you're trying to sell an article, you're always gonna be squeezed between the subscriber business model and the free business model... I think that was probably true in the past." "As you give agents more complex tasks, you want them to be these little hummingbirds, going around the internet, just slurping up a little data here... doing a little compute over here... You don't want the individual human to have to create accounts everywhere." "I think microtransactions will just be necessary for that economy to exist, the agentic economy." "On the flip side, they're now eminently possible because of stablecoins... If you just give an agent a budget, they can easily take dollars, move into stables, and find a way to check out. They don't mind the back and forth." @gaybrick @stripe

Vídeo

Stripe's Will Gaybrick: "Build everything" Against an industry that sees agents as a way to cut costs, Stripe is using them to build more: agents wrote 30% of code in a week, global tax filing shipped in 1/3 the time the US version took, and after AI made sellers 20% more productive, Stripe hired even more sellers. President of Technology & Business @gaybrick sits down with a16z's David George to cover why there's no one left for Stripe to copy, why checkout pages will disappear, how agents plus stablecoins make micropayments real, and why tokens are becoming a currency worth protecting like dollars. 00:00 Intro 01:00 From payments to 30 products 02:30 1 in 6 free trials abused 05:50 Win the startups, then win them again 09:40 Borrowing from Google, Apple, and Ford 14:30 Minions: 7K one-shot PRs a week 18:45 Building everything vs. cutting costs 26:00 Why timelines keep compressing 29:50 What replaces the checkout page 34:20 The case against $9.99 subscriptions 37:20 Stablecoins solve a political problem 41:35 Tempo, a payments-only blockchain 43:10 Tokens are money now 49:00 How Stripe scales taste t.co/MW8kTxfcpq @gaybrick @DavidGeorge83

Vídeo

Travis Kalanick, late 1990s

Travis Kalanick, late 1990s

Right place, right time: many neoclouds spent years mining crypto, then AI showed up and turned their power rights, data centers, and GPUs into some of the hottest assets in tech. 25 quarters in, CoreWeave is pulling in more quarterly revenue than Azure, AWS, or Google Cloud were at 30. Charts of the Week: t.co/2D6Avy4f2T

Travis Kalanick says when it gets easy, it's time to push hard: "I always used to say, if it's getting easy, it's about to get really hard. You don't want it to be that way." "I've talked to a lot of entrepreneurs over the years where they're like, 'Oh man, we're just cranking. We're on easy street right now.' And I'm like, 'Dude, you are about to get your ass whooped and you don't even know it.'" "I know that too because I've experienced it myself." "There's actual times when it's easy, but it's because you're not continuing to push... When you're winning but you're taking it easy and you're about to start losing." @travisk @bhorowitz

Vídeo

Travis Kalanick showed up to fundraise for several stealth companies at once. Ben Horowitz only needed to check one thing: Ben: "The other thing that I learned was he was still Travis. He hadn't turned into... He wasn't living his best life. He wasn't doing that. I was at conviction very fast." Travis: "We actually had separate companies going. So when I went to do the fundraise, I showed up at a16z like the guy with a trench coat with a bunch of watches saying, 'You want a watch? Which watch would you like? I got lots of stuff. Do you want some food stuff? I've got some mining stuff.'" "They're like, 'We just want the stuff.' And we heard that from the first few people we talked to... I had to merge separate entities with different investors and put it together." "You see some of the things with Elon, and he's starting to put the pieces back together... He was managing five or six or seven different companies. So I was really trying to put it together under a single roof." @bhorowitz @travisk

Vídeo

Dario Amodei recently commented he was concerned employees were starting to prioritize money over mission. He may be on to something... Charts of the Week: a16z.news/p/charts-of-th…

Wild adoption gap: the top 1% of AI spenders are spending more than 600x as much as the median company Charts of the Week: a16z.news/p/charts-of-th…

Michael Truell in 2025 on Cursor's trajectory: "Eventually in the future, we want to touch the model side of things... that's actually been a really important product lever for us." "I do think we're gonna need to be a multi-product company going into the future." "There's a whole AI coding bundle to be built, and we want to be, for many of our customers, the AI coding provider for them." "The ways in which work is changing within the editor start to affect how teams work together too... It's necessary to have the best editor, as to also have this complement that's helping teams review and collaborate." @mntruell w/ @martin_casado, a16z Runtime 2025

Vídeo

Cursor, 4 years ago

Uber's original culture had a value Ben Horowitz still quotes, "meritocracy and toe-stepping." Travis Kalanick kept it at Atoms, under a new name: Ben: "We're here to build. We're gonna do it. We're gonna push as hard as we can, and yeah, you may get your toe stepped on, but you gotta keep going... In some ways it was the core part." Travis: "At Atoms we call it 'The Best Idea Wins.' But the toe-stepping is important, because there's so many people that are not willing to fight for the best idea 'cause they're gonna upset somebody else." "If you don't fight for the best idea, then what idea are you fighting for? You're fighting for the mediocre idea, the politically expedient idea, an idea that's gonna make you less good than if you went for the best idea." "Having everybody oriented towards that is a cultural vibe that helps you win." @travisk @bhorowitz

Vídeo

Cursor once self-described as "compute-starved." After an emergency meeting in January 2026, they decided to lean into Cursor’s biggest advantage: its beloved product and massive developer reach. By May 2026, they released Composer 2.5 and beat frontier models on a price-to-performance ratio. "It seemed utterly unreasonable that Cursor might pull this off in such a short amount of time. But iteration speed matters, and Cursor moves fast." Full piece from @sarahdingwang, @BornsteinMatt, and @martin_casado on Cursor + SpaceXAI: t.co/jnyXICp1pJ

Travis Kalanick says entrepreneurs need pissed-off energy: "When you get into your 3rd and 4th, this may be even my 5th at this point, company, you get fricking really good at doing stuff." "Things that used to take me a day or 3 days, a lot of stress, anxiety, like how am I gonna do this, take me like 45 minutes." "I'm so used to adversity at this point that it's kind of normal, so I don't get as pissed off. But being pissed off makes you good as an entrepreneur, so I gotta make sure I'm still fired up." "You get too used to adversity, and that's kind of a problem." @travisk @bhorowitz

Vídeo

"We are in a market that's had an iPod moment, and it's gonna have an iPhone moment, and another iPhone moment." Michael Truell on why AI coding is nowhere near finished: "Despite the headlines, despite how much demand there is in this market, it's so far away from being automated." "It's really easy at an executive level to underestimate just how far away we are from the limit of automating software. There's a really long, messy middle." "There are definitely more in the future, and we've tried to build a company that can continually build those things." @mntruell w/ @martin_casado, a16z Runtime 2025

Vídeo

Travis Kalanick says revenge can build a business, but it will not build your best work: "A couple of companies before Uber, I did a peer-to-peer file-sharing system. Got sued for a quarter of a trillion dollars by 33 of the largest media companies in the world and I was pissed." "I did a revenge business... my whole thing was turn those guys who sued me into customers. It was like a full spite business." "[Revenge] can be successful, but it ends up being less than what it should be when it comes from the wrong place... Part of what drove me earlier was a little bit of a fear of failure." "When the Uber thing happened and then I moved on to new stuff... I kind of fell in love again." "When you fall in love again, you don't think about the ex very much. If you come from that place, you can create in a beautiful undirty way." @travisk @eriktorenberg @bhorowitz

Vídeo

"Anysphere—which most people just called 'Cursor'—became the fastest software company in history to reach $100 million in annual recurring revenue." "In 2024, things went vertical. Developers discovered the product and loved it, and they told their friends about it, and their friends loved it too." "At one user meetup, a user from Japan explained that he’d flown in from Tokyo with a book he’d written in Japanese about how to use Cursor. It was honestly like the Beatles for software; we’d never seen anything like it." Full piece from @sarahdingwang, @BornsteinMatt, and @martin_casado on Cursor + SpaceXAI: t.co/jnyXICp1pJ

Travis Kalanick says new media is why he's back after eight years of silence: "The media landscape 10 years ago, 90% of what was said was very negative. Business had become politics, and we didn't know it... If it bleeds, it leads, right?" "Now the media landscape is like, 'yeah, you don't have to talk to those guys.' You can go and talk to a David Senra. You can go and talk to Joe Rogan or name your guy, and you can say what you need to say." "Elon buying Twitter is the beginning of us being able to speak our minds and for disagreeing to not be illegal. That's a big deal. We should all thank Elon for that." "I want to build. I want to not think about what The New York Times is writing when I make decisions about what I build. I had to emulate what it's like being a capitalist in Russia... and that's what I did for eight years." @travisk

Vídeo

Travis Kalanick: "A lot of folks think I'm back. I've been working my ass off the whole time. I just haven't been talking about it." Eight years, thousands of employees, multiple industries, none of it on LinkedIn, and now the biggest check Ben Horowitz has ever written. @travisk joins @bhorowitz and @eriktorenberg for a fireside chat to discuss Atoms: an industrial AI company that sees manufacturing, real estate, and logistics as the CPU, storage, and network of the physical world: 00:00 Intro 01:30 The Uber video from 10 years ago 04:26 Bits, atoms, and the three computing primitives 07:37 Can a delivered meal cost less than the grocery store? 10:16 Standing next to a TCP packet 11:59 Ride-sharing was the gold medal. Transport is full of silver 13:59 Resistance to change is the final boss 22:22 Why he didn't buy Lyft 24:09 Best idea wins 28:13 Blood, sweat, and ramen 32:32 When it gets easy, push harder 36:18 Showing up to the fundraise like a guy selling watches 45:46 Dirty fuel vs. falling in love again 48:31 Be uniconic

Vídeo

"Cursor and SpaceXAI are a deep cultural fit with one another. They both work with a pace and intensity that borders on the absurd. And they’re both defined by iteration." "It doesn’t matter if you get it wrong the first time, or the second, or the third—even the tenth or the twentieth. Just keep on going. If you’re moving in the right direction faster than anyone else, you’ll probably win." "Cursor reinvented itself as a company not once but twice: first, from email client to tab-complete IDE and token reseller; and again, from tab-complete IDE and token reseller to AI lab and token creator." "And it did all of that in less than four years." "They really, really want to win." Full piece from @sarahdingwang, @BornsteinMatt, and @martin_casado on Cursor + SpaceXAI: t.co/jnyXICp1pJ

Figma turns 14 today

Every trucking company in America was suddenly required by law to buy the thing Samsara happened to sell. a16z GP Andy McCall, who ran Samsara revenue through the IPO, tells the story: "Prior to 2016, truckers had these manual logbooks... You would write down in your logbook, 'I just drove for four hours, now I'm taking a 20-minute break.'" "Around 2016, they implemented this ELD mandate: electronic logging devices. Hey, we can use technology to actually track when the vehicle's moving and when it isn't." "It provided this huge tailwind... and we just happened to be one of the newer companies doing it." "Basically the entire industry all of a sudden had to find budget to go out and buy these things." t.co/MEjXWPdmBa @joeschmidtiv @VirtualElena

Vídeo

Joe Schmidt says not every enterprise AI company needs a billboard in San Francisco: "You just realize that you have the same two competing companies. One's on one side of the freeway and the other's on the other side of the freeway, and they're selling the exact same piece of software." "It's all targeting the same sales motion. The reality is you don't always have to do that. You don't always have to sell to the same companies in San Francisco." "What I wanted to try to do is just tell founders, hey, here's the framework for evaluating which playbook should you be following. There's this very obvious one, which is to go after the very obvious companies here in San Francisco, in New York City, in a major metro that probably have some sort of proof or social value associated with them." "Or go out and sell in Ohio, go out and sell in Chicago, go out and sell in St. Louis. Find people who need your solution." t.co/MEjXWPdmBa @joeschmidtiv @VirtualElena

Vídeo

a16z GP Andy McCall on advice for starting a sales career, and the hire founders wait too long to make: "I made this mistake early in my career. I was chasing where can I make the most commission... At the end of the day, none of that matters." "You want to find the great company that's gonna grow. It's like a career elevator. You will grow with the company. You can't let your ego get in the way." "Sales operations is probably one hire that I see companies waiting a little too long on." "It can be literally one person. But you need somebody that every day is thinking through territory alignment, commission skirmishes, setting a sales constitution." "When you get to scale mode, you want that stuff largely figured out. They become speed bumps otherwise." t.co/MEjXWPdmBa @joeschmidtiv @VirtualElena

Vídeo

Lighthouse or Landgrab: Choosing Your AI Sales Strategy In this conversation, a16z's Joe Schmidt, Andy McCall, and Elena Burger sit down to discuss the lighthouse and landgrab playbooks for selling enterprise AI: win marquee logos so proof travels, or win budgets that already exist. They cover: - How companies like Harvey and Applied Intuition run the lighthouse playbook, while Stuut, Decagon, and Pylon use the landgrab strategy - How Andy built Meraki's sales org and then took Samsara from single-digit millions to >$1B ARR as CRO - Why you should spend 1% of your time on sales strategy and 99% on execution - How to hire the right sales org based on your strategy 00:00 Intro 00:58 Lighthouse or Landgrab? 02:36 Buyer exposure and proof that travels 06:31 Samsara's ELD landgrab 09:58 Sell to whoever will buy 13:26 Examples like Stuut and Harvey 16:28 Pylon & climbing the ACV ladder 18:21 Meraki's free-access-point playbook 21:39 Your proof of concept shouldn't be a science project 26:46 Why most companies need both playbooks 29:53 Hiring for lighthouse vs. landgrab 33:54 The return of big software 37:32 1% strategy, 99% execution 38:58 Advice for sales careers @joeschmidtiv @VirtualElena

Vídeo

We're thrilled to invest in Vals. A frontier model can look brilliant on a leaderboard and still struggle with the messy work that actually matters in the real world. @ValsAI takes a fundamentally different approach to evaluation: test models on the work people actually want them to do, not on contrived exams. The team works with domain experts to turn real workflows into rigorous benchmarks, then builds automated grading systems that can evaluate the final work product to an expert standard. We're thrilled to partner with @RayanKrishnan, @langstonnashold, and the entire Vals team as they build the trust layer to underpin the AI economy. By @JenniferHli, @stuffyokodraws, @RaghuRaghuram, and @shangdaxu

"'100% of this text is AI' is the new Scarlet Letter of dunking on people and posts." "I hate to give the French credit for anything, but the 1960s semantic thinkers who created 'French Theory' really cooked when they anticipated Pangram and AI authorship 80 years ago." "There are certain tells that you can spot where, it's like, 'This is a coherent sentence, but there's no way a human wrote this.'" "To bastardize Voltaire, if a wise and brilliant author doesn't exist, the LLM has to invent one." "People want authorship. They want authorship because it is meaning, because the indexing and relation and handling of ideas is itself the meaning of the idea." Alex Danco on Pangram, authorship, & French Theory: t.co/oMOBEOO0Wg

Garry Tan says AI is a multiplier on the individual: "We're sort of in the middle of a big transformation now, I think. For the longest time, we really, really believed that you needed co-founders. And I think that everything else being equal, having co-founders is net really, really good." "I think what's happening right now with vibe coding and agentic coding, literally any given person could be 400 of that person, like from two years, even nine months ago, you can be a multiple of yourself." @illscience: "Do you think founders are more ambitious than they've ever been?" @garrytan: "They should be."

Vídeo

Garry Tan says turning down Palantir for a Microsoft promotion was a $2–4 billion mistake: "Joe Lonsdale and Stephen Cohen were my fraternity brothers at Stanford... they were interns at Peter Thiel's hedge fund." "They said, 'Let's go get Garry. He's up in Seattle.' They flew me down to have dinner with him, and Peter said, 'I'm so sure this is the right thing for you. Here's a check for 70 grand.' That was how much I made at Microsoft at the time." "I said, 'Thank you very much, Mr. Thiel, but I might get promoted to level 60 this year.' Which I did. But that was a $2 to $4 billion mistake at this point." @garrytan @illscience

Vídeo

Garry Tan says we'll see a wave of exceptional older AI-native founders: "There's going to be as many Patrick Collisons as ever, but one mega trend that we're seeing is the 35, 40, 45-year-old founder who's been around the block, built a lot of engineering." "Peter Steinberger is a perfect example of that. He's been around the block. He knows what to build." "If you take that person, suddenly there's 400 of those people. You can outperform an entire department of any Mag 7." @garrytan @illscience

Vídeo

Garry Tan says turning down Palantir for a Microsoft promotion was a $2–4 billion mistake: "Joe Lonsdale and Stephen Cohen were my fraternity brothers at Stanford... they were interns at Peter Thiel's hedge fund." "They said, 'Let's go get Garry. He's up in Seattle.' They flew me down to have dinner with him, and Peter said, 'I'm so sure this is the right thing for you. Here's a check for 70 grand.' That was how much I made at Microsoft at the time." "I said, 'Thank you very much, Mr. Thiel, but I might get promoted to level 60 this year.' Which I did. But that was a $2 to $4 billion mistake at this point." @garrytan @illscience

Garry Tan says AI enables a single markdown file to act like an employee: "If you really want to tokenmax, you have to use something like Hermes Agent or OpenClaw... let me load a million tokens or 800,000 tokens into any given request." "For a CEO or a founder, it actually makes a lot of sense to do that, and you have to give yourself permission to tokenmax in that way. What you get is, you get to live in 2028 today." "You do some feat of strength, and then you turn it into a markdown file plus code plus tests that can be reused and put into a cron job." "A markdown file is an employee. It's an employee that will do the job perfectly every single time, and it'll do it as many times as you want." @garrytan @illscience

Vídeo

Garry Tan on YC, First Principles Thinking, and Progress in the Age of AI Silicon Valley's advantage is that it gives weird, ambitious outsiders permission to try, find one another, and become insiders through building. Garry Tan learned this firsthand throughout his career at companies like Microsoft, Palantir, and YC. AI may now extend the power of Silicon Valley much further. Garry joins a16z’s Anish Acharya to cover the turns that shaped him, why founders should trust direct experience over consensus, how AI is allowing small teams to do more with less, and why he's deeply optimistic about AI-native companies and the future of SF: 00:00 Intro 00:26 Getting into tech 04:45 Silicon Valley culture & finding your people at the fringe 10:59 What makes YC great: a birthright for tech outsiders 13:54 Solo founders, vibe coding, and founders being 400x themselves 21:05 Business loops: skillifying every task into a markdown file 23:03 Tokenmaxxing: how to live in 2028 today 28:09 Using AI to make conflict constructive 33:10 The torture of the white-collar job & life above the API line 39:08 Why bureaucracy is a whitepill 41:33 What the next computer looks like: voice, memory & the harness wars 44:44 Local politics & making SF a beacon t.co/N64jhnYNkY @garrytan @illscience

Vídeo

Today in "everything is an 80/20 rule": For most B2B businesses, a few buyers make up the bulk of their revenue. Full piece on how Stuut uses agents to help B2B companies get paid the money they're owed: a16z.news/p/what-it-take…

Collections is less of a grinding problem and more of a search problem: find the few invoices that move the number, and put your best effort there. For a typical business, two-thirds of the money they're owed sits in 10% of invoices. Full piece on how Stuut uses agents to help B2B companies get paid the money they're owed: t.co/FgyDwqj0od

There's a $7 trillion pile of unpaid invoices held by US nonfinancial businesses. Full piece on how Stuut is building the solution: a16z.news/p/what-it-take…

.@PalmerLuckey says every generation fights the new thing, and no generation fights to bring the old one back: "People are saying, 'Oh my God, AI is gonna ruin everything.' How many people said the same thing about automated manufacturing?  The reality is that it actually made cars from a plaything for the rich into something that anybody can do." "There were artists who believed that photography would kill painting, art, and illustration. And of course, it didn't do those things. It actually enhanced them, and it created its own new art form." "I feel like we're unfortunately in one of those swings where everyone is questioning whether technology will make our future better. I think it's especially unfortunate because we're on the precipice of so many things that have been scarce becoming unscarce." Palmer Luckey w/ Tetsuro Miyatake, Offtopic Podcast (2025)

Vídeo

Soon

Datadog CISO Emilio Escobar says intent is now a must-have for AI security solutions: "These agents are trained on existing code, and they have a reward structure. If code is meant to solve the bug, but it gets rewarded on that, it doesn't care if it's doing something else outside of that." "You have to be careful how you prompt these things, but also how it actually interprets your prompt and executes on that. So we have this judge now evaluating the code output of the agents to then make sure that we're doing that." "I did a roundtable last week about agentic security, and the sense that I got from a bunch of the security leaders was a sense of helplessness. Of just waiting for a commercial solution to come in and solve it all." @eaescob @datadoghq

Vídeo

Datadog CISO Emilio Escobar on Securing Agents at Scale In this conversation, Emilio Escobar joins a16z's Joel de la Garza at Black Hat to cover what happens when you hand coding agents to 4,000 engineers, why the permissioning that worked for a decade broke the moment agents could write their own SQL, and why Emilio isn't panicking: 00:00 Intro 00:55 Securing agents 03:11 AI flattens the org chart 05:20 Role-based MCP servers & sandboxing agent credentials 07:34 Understanding agents' intent 10:27 How to avoid reward hacking 12:07 How the CISO's role is changing 14:05 "Security engineers will become real engineers" 18:12 Why Emilio isn't panicking t.co/4ah0IXCusG @eaescob @datadoghq

Vídeo

Introducing Grok Bot

Inferact CEO Simon Mo says open and closed models are converging. The moat is environment: "What really differentiates open-weight models from closed-weight models? In the end, there's not much differentiation. It's more about the distribution strategy and go-to-market strategy. Capability-wise, I don't really see a big gap, not even today." "One of the most important parts is just the data. It's about who gets what data, and then what are the environments you are building to let the model improve on itself." "For Moonshot, they have built some of the best environments for front-end coding... the ability for this model to code and then see what the rendered code is, and then continue looping in this iterative process." "Now, this is about their environment to improve the model. It's not about just source data. It's not about where they get the data from. Rather, it's who can build the best environment and who can make the most optimization and algorithmic choices to leverage all this learning from this environment." "The next year is all gonna be about that: how open-weight model labs are differentiating and really getting the model to meet the real world." t.co/OiRH8fXarb @simon_mo_

Vídeo

Matt Bornstein says open source is how AI companies stop being wrappers: "A year ago, a bunch of smaller companies or new application companies were trying to figure out, 'How do I really build an AI without just being a wrapper on top of OpenAI?' The answer to that question turned out to be open source." "This is what Cursor did. This is what Decagon and Harvey are in the process of doing now. A bunch of really strong application-level startups made the determination: we can't build just on closed source." "We need to do our own mid-training, our own post-training, our own inference and deployment tricks. All of that means it must be built on top of open source. The closed-source vendors won't give you the access to do this." "Open source became really central in a way that's not always visible because it's deeply embedded in some of these products. Some of the most innovative products and applications now really depend on this very deeply." t.co/xFgyd5YjXv @BornsteinMatt

Vídeo

"Markets outperform central planning because they transform countless independent encounters with reality into a distributed process of discovery, adaptation, and selection." "Many competing superintelligences, each probing a different frontier of the universe with different assumptions and values, will discover more than any single intelligence, however vast." "Open weights extend the same logic: they allow intelligence to be adapted everywhere rather than optimized once and rationed from the center." "Regulatory capture may slow their diffusion. It can weaken the gradient. It cannot reverse it." Christian Catalini on the economics of open versus closed AI: t.co/AyuMbXPr1m