Everyone wants to be a product manager, PMs want to be GMs, GMs want to be CEOs, and the CEO just wants to build product again ..

For years, we assumed consumers wouldn’t pay for software. Now people are happily paying $200+/mo for products like Claude, Grok 4 Heavy and Gemini. Not because they’re for everyone, but because they’re 100x better for someone. The free tier is the funnel. The real product is narrow.

“We will call seeds pre-seed, and price them like As”

Here is my 2 cents on OpenAI buying Windsurf + appointing a "CEO of applications". There is a concept that was popularized during the big telco era called "wholesale transfer pricing" which is effectively the ability of a supplier to extract margin from everyone else in the value chain, most notably the consumption layer. In extreme cases this power is so high that they can extract (largely) all of the profits from the rest of the stack, as has happened with record labels (unique suppliers of the Beatles, for example) and music streaming services. There is no need to vertically integrate when you're a unique supplier to an industry. I think this was the situation that OpenAI seemed to be in as they had the most cutting edge and dominant foundation models, so all they had to do was keep working on AGI and raising prices to capture all of the economics in the ecosystem. However, we're now living in a world where consumption layer apps are multiplexing and routing to different foundation models, which is disadvantageous for (foundation models) suppliers and limits their ability to extract down-funnel economics. In that world OpenAI must vertically integrate and own the consumption layer to protect their economics, thus the recent focus on AI Applications.

A surprising source of alpha is *actually using* new AI products - if you haven’t tried R1, Operator, DeepResearch, Cursor, Krea, Notebook LM etc you’re at a competitive disadvantage to builders / investors / consumers that have.

Impossible Tiny Desk🔈 Obviously this is not actually Christopher Wallace nor is it NPR Instrumentation from the very talented The Frank White Experience via YouTube 4o / Kling via @krea_ai / @hedra_labs / demucs

An open question was answered today - "what will the AI-native distribution channel be?" It looks like ChatGPT will be that channel with 800M active users + the Apps SDK. This is likely as important as Steve Jobs announcing the app store in March of 2008 ...

As much as I adore India and the many impressive founders+investors in the region, this is entirely fake news!

Deel has grown from a cold start to over $800M of revenue in just five years. But how exactly did they do it? While they may look like a typical Silicon Valley story the truth is that @Bouazizalex and @shuooo have defied many SV conventions in their approach including M&A, no "wedge", and enormous investments in internal tooling/automations. @justin_kahl and I spoke to leaders across @Deel's business to unpack the tactics and cultural tenets fueling the company, five powerful insights are covered below. Five powerful principles that we cover in depth 👇 - Vertically integrated product stack - Infinite product wedges - Strategic M&A early, and as a core accelerator - Automations - "Deel speed"

Bad board member archetypes: 1 - Too rich to care (doesn't do the work) 2 - Ideas guy (also doesn't do the work) 3 - Fake friend (doesn't tell the truth) The best board members are high truth, low anxiety

Haters hate, builders build. Big up to all the people working on good ideas that sound like bad ideas (the story of every startup). Also, there is a special place in hell for people that work at startups and punch down on other startups. It’s the opposite of a founder mindset.

Introducing the Abundance Agenda. We believe AI will turn luxuries into commodities - making consumers happier, healthier, and more productive than ever before. And a new generation of AI-native companies will lead the way. More from @a16z consumer 👇

Notion launched Meeting Notes today, which takes direct aim at Granola’s beloved AI note taking product. Two quick thoughts: 1 The “incumbents” in the horizontal prosumer / productivity category (Notion / Airtable / Figma) are far more capable + ambitious than the traditional incumbents in a product cycle, and are very awake to the need to replicate AI complements to their ecosystem. 2 Voice scribes have commoditized in < 1 year so any product built around this capability must rapidly build down into a system of record + build wide into a system of engagement uniquely enabled by the voice capability. Capable incumbents + magical new AI features means consumer AI will be more of a foot race than we’ve had in years.

Vibe coding platforms are blowing up. But the data shows it's more of a creation vs. a consumption story. Ex. @lovable - traffic to Lovable itself is 3x greater than all traffic to sites published on Lovable 👇 People are making things for themselves...not the world (yet!)

The bitter lesson is coming for app gen and the best teams are already thinking about which investments are durable vs fleeting. More: - Retention today is proportional to success rate - did the user get what they expected in a reasonable number of tokens (aesthetics / features / bug fixes / deployment)? - Many smart teams have engineered around the constraints of models - context mgmt, parallelism, system prompts, design systems etc and are seeing real alpha as a result - New models will be a tidal wave that overwhelms much of this work - teams will need to be ruthless and avoid sunk cost thinking + building new sources of alpha as needed - The best teams are already thinking about durability - integrations, re-usable components, and routing inference / payments feels safe; the smartest thing I heard this week was “we’re leaning into the bitter lesson” - The only thing I’m sure about at this point is that the world has 1% of the software it will have in 5 years If you’re building in the space I want to hear from you - anish at a16z

🧵💀 As @pmarca says “There is no substitute for knowing what you’re doing.” I’m convinced that most product managers / designers have no idea what to build and have instead created a bunch of process to mask this failure. I count myself in this group at times in my career 👇🏽

Thrilled to announce @a16z is leading @partyround's Seed - a bet on an incredibly creative and ambitious team disrupting how founders raise money 🎉

I’ll take the other side, here are a few approaches that advantage app layer startups: 1 Categories that benefit from being *multi-model* - labs and big tech will only ever be able to offer their 1P models and in coding / creative tools etc you get better results by multiplexing across many providers 2 Cornered resource (data) - you note this but there are many categories where startups have locked up a proprietary dataset and are 10x better than labs (open evidence, vlex) 3 Networks / compounding loops 4 Ecosystems that imply a ton of feature surface area (sure you can replicate Granola’s recorder but is OpenAI really going to build the entire ecosystem of productivity apps implied by it) The labs are ambitious and formidable but this is the same as saying Google will win everything on the web in 2004…

come work with me @a16z! we're hiring a partner on our consumer AI investment team A bit more about our consumer team @a16z: - we have the OG founder of consumer software on the team (yep, it’s Marc) - consumer investing is in our DNA - Coinbase, AirBnb, Pinterest, and Facebook are all a16z - we have a deep consumer AI portfolio @ a16z - krea, eleven labs, hedra, ideogram, mirage, cluely + many more our investment partners help identify and select the companies we invest in - they think critically about new technologies / models / products / market trends, can build a sophisticated POV on a business by analyzing the data, and work directly with GPs in the board room to help companies make progress across product, GTM, fundraising etc. a few things we look for: - relevant experience as a builder (engineering, product) or as an investor / analyst with banking experience - interest in start-ups - founder empathy is a key a16z cultural value, so having passion / experience or POV on the startup community is a must - technical / passion for AI - you've embraced AI in your workflows - highly numerate - you understand cohort retention, revenue quality, and selection bias - social network - you're at home networking with founders, and might be active on X or Substack if this role seems exciting the best way to connect to us is an intro through a mutual, but you can always fall back to consumerinvesting@a16z.com with X, resume and/or LinkedIn

Notes on app generation / "vibe coding" - All the best app generation / vibe coding products are now focused on creating+hosting production apps - If these platforms ultimately make money by "taxing" payments + inference and domain hosting for production apps then that revenue should be very high margin and overwhelm the economics of generation over time - Thus platforms for making production apps >> platforms for (just) prototyping - Very common and totally acceptable to have a zero GM free tier - I think of this as "CAC" that converts into higher margin "ultra" tier ($200/mo+) - Story for compounding advantage still unclear - could be app/platform feedback loop (likely on mobile) or composability via re-use of generated software - Models making exponential progress and will soon be able to generate code of arbitrary complexity; at this point value moves to ideation / refinement / extensibility / distribution; app layer hugely important - Labs will likely want to push from developer up and from the consumer down; vibe coding seems like a path for Anthropic to have a differentiated mass market consumer story - I believe this is a global consumer phenomena that will be as big as social media

Still the greatest Drake diss record of all time #kendrickvsdrake

Overdraft revenue is to banks what SMS revenue was to telcos - a line of business that's going to zero in the foreseeable future, yet prevents them from building stuff with 10x more profit potential and customer value.

Thrilled to announce we're leading the Series A in @uuuntitleddd! [untitled] is building the new operating system for musicians - finally giving them workflow tools that live up to their creative expectations. Why our team @a16z invested👇

This essay from @paulg is a must read for builders: paulgraham.com/greatwork.htm… There are many insights contained in it but I thought this one was particularly important, as it’s a trap that many founders (including me!) have fallen into: “Instead of making what they want, they try to make what some imaginary, more sophisticated audience wants. And once you go down that route, you're lost.”

Thrilled to share that I’ll be joining the @a16z team: a16z.com/2019/07/11/anish-ac… where I’ll be looking out for the next wave of fintech companies that are making money fair, efficient, and perhaps even joyful ..

I spoke with @cdixon on the a16z pod about what consumer founders should think about in the age of AI. Chris is one of the rare people who has seen these cycles from every angle. He was a 2x consumer founder himself before becoming one of the most successful consumer investors of the last two decades, backing companies including Coinbase, Oculus and Reddit. He has also been consistently early to new communities on the internet, predicted major platform shifts, and built a deep understanding of networks at a time when most people still underestimated their power. He has also written some of the most influential essays on these topics and is, in many ways, a true n-of-1 thinker on how technologies grow from toys to global platforms. My notes and takeaways from the episode:

Thrilled to announce I'm leading the Series B in Titan - this is a bet on a world class team building the future of active investing 🚀👇🏽 a16z.com/2021/07/20/investin…

This is the best use of App Clips I’ve seen so far by a long shot. App clips has long been a viral channel hiding in plain sight. Nice work @nikitabier !!!!

We're leading a $20M round in @Wabi. Wabi is a personal software platform where users can create, discover, remix and share mini apps, founded by @ekuyda. Why we're thrilled to be backing the team: - The internet drives participation, yet only 20M programmers have ever been able to create software…what if 6B consumers could create their own apps? - Eugenia is a genuine badass who was early+right to AI companions. @Replika was the first API customer of OpenAI - This is the YouTube moment for software, with one important twist: content decays over time while software compounds - Social is a primitive - the next gen app store will be multiplayer by default - Software no longer has to be durable, serious, or mass market - we're finally going to recapture the weird eclectic internet vibes that we had in the 1990s

Our team @a16z couldn't be more excited about voice in the era of AI. AI can finally take the "voice data" we produce (or hear) all day, and use it to improve our work and our lives. Our thesis on AI scribes ⬇️

Yep, we just invested in @ArcBoats the "Tesla for Boats". Why? - Repeat founders + team from SpaceX, TSLA & Boeing - Electric is a 10x for every part of boating: no maintenance, cheaper, quicker, and unlocks access to protected waters -Aerospace design practices & A+ software 👇🏽

You know the vibes are spectacular when the @tbpn guys are in the mix

I’ve been thinking about this for some time. The way I’ve crystallized it is we can now build disposable software. When software was expensive and hard to build, we mostly made it for serious things and expected it to last (like payroll systems, tax tools, ERP software and consumer networks). Every line of code had to justify itself. But not all software needs to be permanent or practical anymore. More and more, people are creating small, personal apps and tools that only make sense for you or maybe a couple of friends. You can build something you never would’ve justified before because the economics didn’t work. But now they do. I built a math game to help my kids earn screen time. I spun up Catsagram -- a mini Instagram for my kids to post pictures of our cat (this is real, and you can sign up). These aren’t products or startups. They’re not made to scale or monetize. They’re just little apps custom-fit for a specific audience, useful in a narrow context, and fun to build. To Justine's point, this wasn’t really possible before. You wouldn’t write a thousand lines of code just to share photos with three people. It wouldn’t have made economic sense. But that’s changed. With LLMs and AI-native runtimes, it’s now easy to spin up custom software with almost no effort. What used to take weeks can now take an hour or less. In many cases, you barely need to write any code at all. Building small, throwaway apps is starting to feel like doodling in a notebook and that shift changes why we build software in the first place. Software creation used to be constrained by ROI. Now it’s constrained only by imagination and that’s a much more interesting limit.

💥 New post today on three key lessons I learned at @Creditkarma in moving from a generalist consumer internet PM to a specialized consumer fintech PM: a16z.com/2020/02/26/fintech-…

Software creation used to be constrained by ROI. Now it’s constrained only by imagination and that’s a much more interesting limit. Welcome to the era of disposable software.

1/n 🎯New post from @matt_haf and I about the rise of active trading + why it's here to stay, inc catalysts and the 🔑 approaches: - Distributed hedge funds - Individuals as the largest asset managers - Low code trading IDEs - Bottoms up communities a16z.com/2021/03/29/active-t…

The post bank fintech stack will look like.... - Every person is a business of one ☎️ - Per card swipe auction to give you the lowest cost of debt 💳 - Income streaming, configurable rewards, yield chasing across asset classes ⚙️ - Multiplayer👯‍♀️ + More below..

Thrilled to announce that we're leading a $4.5MM seed investment in Runway, a new product with the vision of giving every leader in an organization access to valuable financial data and insights. Excited to work w/ the awesome founders @blader and @a_a! a16z.com/2020/06/25/investin…

1/3 Every founder faces a tradeoff between prioritizing exploration (build new stuff) and exploitation (build more of what's working). This is the multi-arm bandit problem, where the answer is normally some combination of the two, but this misses something fundamental.

App generation is one of the fastest growing AI spaces. Cos are scaling to $100M+ ARR at record pace, seemingly in a race to dominate the market. But we think it won't be "winner takes all." The market will segment, with different winners in different areas. Our thesis 👇

Thrilled to announce that we're leading the $14MM Series A in Mosaic, a software platform that delivers beautiful, personalized homes at a lower cost and higher speed by partnering with local homebuilders. Excited to work w/ Salman Ahmad and Sep Kamvar! a16z.com/2020/09/02/investin…

VCs trying to figure out who @ArfurRock is

Field notes for Product Managers in the Age of AI Product managers have always been in the business of solving for ambiguity — they insert themselves to resolve uncertainty around execution, product craft, analytics, and customer needs. As AI advances, so does the argument that the PM’s role is being made obsolete by increasingly capable models. However, debate around the viability of the function misses an important point: there isn’t any less ambiguity involved in bringing products to market and scaling them today, but the tools and opportunities are completely new. Product managers that ignore this dynamic risk irrelevance. The current class of product leaders came of age during the mobile/web transition and were trained in mobile app-specific methodologies like growth accounting, mobile-first product craft, and owning an app icon on the homescreen. There was a lot of discussion around how to avoid injecting desktop-era product holdovers like hover state cues and click-first design metaphors into mobile products. What are the new skeuomorphic metaphors that will constrain mobile product management when building in AI? And what are some of the best practices to consider? A few come to mind: 1. “Interview” your models alongside your customers. Large models are inherently probabilistic, meaning that given the same input, the outputs are varied rather than deterministic. Because AI models have these stochastic outputs and exhibit emergent behaviors — unexpected actions that aren’t explicitly programmed — PMs now have to spend as much time “interviewing” their models as they do their customers, probing to understand the models’ capabilities and constraints. PMs should be asking: What types of “noise” or unpredictable outputs do these models generate that I can leverage in my product? How does the model respond to edge cases, and when do I need it to be stable? Honing this intuition can lead to products like Websim, the AI-powered simulator that generates strange, unexpected websites that make you feel like you’re peering into the model’s mind. Instead of reining in the model to produce polished, conventional outputs, Websim’s builders leaned into the weird. To find the value in the unexpected, PMs also need to become skilled at writing evals: structured tests that help you see where your model performs well and where it’s falling short. Evals aren’t just about measuring accuracy, they’re identifying and assessing emergent capabilities to inform your product design. 2. Don’t shy away from extreme products at extreme prices. We’re now seeing a class of software products that can do things that were unimaginable just a few years ago: an AI nurse that calls patients with information and reminders before surgery; a tool that generates complicated web applications from a single prompt; a product that performs sophisticated research and analysis that previously would have required a team. In this world, there really isn’t a ceiling on how much you can charge. When ChatGPT launched its $200/month subscription last year as a mass consumer product, it seemed like wishful thinking. Now it’s a go-to, daily tool for power users. Similarly, AI products like Krea, Cursor, Midjourney and many others have been successful in aggressively exploring price ceilings, rather than optimizing for price floors. Our belief is that software will be the #1 category of discretionary spending for consumers in the near future. Against that backdrop, the PM’s product design prompt should be: “What does the $1,000/month version of our product look like?” Then work backwards to deliver it. 3. The elusive AI moat: first and fast. Consumer AI companies should be intentional about exploring new moats, whether through LoRAs, proprietary workflows, integrations with other software, or new channels like voice and phone. I believe one of the most overlooked moats is emotional: simply being willing to build products that carry emotional valence. Apple, Google, and the like have a thousand committees designed to ensure that the messy aspects of the human experience (disagreement, persuasion, sexuality) are never surfaced in their products. Language models function as mass “averaging machines” optimized for consensus, leading to products that can be bland, uninspiring, or outright bad. Startups, on the other hand, can build around those edges — emotion, friction, intensity — to create products that feel unique in the marketplace. Network effects remain the gold standard of software moats. However in the competitive landscape of AI, where the volume of products being spun up is so enormous, many of the traditional frameworks for establishing moats may not apply. For example, systems of record can now be indexed via vision models + RPA, potentially rendering the strength of this moat less powerful. As builders gain access to the same models and infrastructure, “soft” moats — like mindshare and momentum — that once seemed too weak to sustain a competitive advantage are becoming increasingly important in consumer AI. A playbook we’re seeing more of: first and fast. Leading founders are the first to build a product, then stay at the front of the pack by continually shipping new features and capabilities. 4. Models are platforms, not products. The first generation of AI products were really webpages in front of models, with the foundation models doing the heavy lifting of generating images, writing poems, and delighting users with new capabilities. As these foundation models increase in number and sophistication, users will increasingly require opinionated workflows around the model to make the most of them. For example, text-to-app products like Replit, Lovable, and Bolt are a miraculous experience for prototyping new ideas. But moving from prototype to production will likely require more advanced interfaces that support fine-tuning and customization. Thus, we believe the next generation of large-scale AI products will be opinionated and sophisticated products built around foundation models. 5. Reflexive AI use is tipping from differentiator to default. You can’t productize a system you don’t understand. That means it’s not enough to dabble in ChatGPT, you need to understand the difference between a language model and a reasoning model. Have you tried Deep Research, Operator, Gemini Flash, custom GPTs, and GPT-4o in multimodal mode? Have you read up on chain of thought, or observed it when using DeepSeek or any of the other reasoning models that expose it? The single most important intuition-builder for PMs is reflexively using AI products every day, in every part of their job. This view is quickly tipping into consensus, as the CEOs of Shopify, Duolingo, Box and many more declare their companies pivoting to AI-first in all efforts. Over time, PMs should naturally incorporate AI products into their daily habits and become beacons for best practices at their company.

Responding to comments on this: - When Apple launched the App Store they had 6M iPhones in distribution. ChatGPT has 800M actives. The "what if it works" case for the Apps SDK is serious. - Memories + discovery + monetization would be a 10x here, I suspect all of that is on the roadmap. - Yes this is directionally similar to Custom GPTs but is far more ambitious in how apps show up / what tools they can use etc. Pessimists sound smart and optimists get paid :)

Thrilled to announce our $14MM Series A investment in Deel, a payroll platform enabling companies to grow their global workforce and providing financial services to these workers. Excited to work w/ the awesome founders @copernicussw and @Bouazizalex! a16z.com/2020/05/21/investin…

The case for consumer: 1 - Most of the biggest software companies in the world are consumer companies. 2 - Most of these companies were built around a product cycle (internet / mobile / web3 / AI). 3 - Generative consumer companies like ChatGPT + MJ are some of the fastest in the world to > $100M of revenue, and this was just year 1 of the product cycle.

.@cdixon explains that software can scale from nothing to hundreds of millions of users in a way no other industry can because of exponential forces. He points to three: Moore’s law, composability and network effects. Network effects are the gold standard but Moore’s law applied broadly (exponential improvement in the underlying platforms) and composability (exponential increase in software utility through open source code reuse) are just as powerful. 1. Moore’s Law: Anything built on top of a platform that keeps getting exponentially cheaper and faster has a built-in tailwind. It began with chips doubling in power every couple of years, then extended to storage and bandwidth, and today includes foundation models. When the base layer improves this quickly, whole categories of products suddenly become possible, and the companies that build for the future rather than the present capture the upside. The first iPhone was basic and limited, but Apple could see where the platform curve was going and designed for it. 2. Composability: Software compounds because it can be reused. Open source makes this especially powerful: once code is written, anyone can pick it up, improve it, and combine it with other pieces like Lego blocks. Each library, API, or framework becomes shared infrastructure that lowers the cost and speeds up the next project. That’s why Linux, which began as a small side project, became the world’s default operating system, and why today’s apps are built on stacks of shared, reusable components. 3. Network Effects: Networks become more valuable as they add more people. Facebook started as a small digital network at Harvard. Then it spread school by school, until suddenly it was the default way to connect online. Once a product clears the hurdle of initial adoption, growth feeds on itself in a way that compounds quickly, often overwhelming slower competitors. This is why tech doesn’t look like other industries. Full conversation with @cdixon coming next week.

Customer acquisition has never been harder and many of us product-oriented founders have struggled here - it’s rare to have a superpower around customer acquisition *and* product craft. If you’re thinking about this here are five ideas to chew on: a16z.com/2019/09/24/how-fint…

Back in 2006, the average person looking at YouTube would have seen amateur dancers, aspiring musicians, and home videos. An enterprising person would’ve seen an entirely new internet-native business model, made possible by creative people circumventing gatekeepers in traditional entertainment. YouTube (and later Instagram and TikTok) began as a place for amateurs to post videos, but evolved into a platform for passionate people to build brands and host internet-native formats (unboxing videos, “get ready with me” reels) that never would have made it on primetime. It’s one of many success stories for the long-tail creator on the internet. We believe we’re at a similar turning point for software itself. Just as YouTube democratized content creation, @Wabi is democratizing software creation, and we’re thrilled to be leading Wabi’s $20M pre-seed. Long-tail entertainment has seen success on the internet. But it wasn’t until LLMs that long-tail software was feasible: it was simply too expensive to hire developers to build software with an unproven track record. Now LLMs are becoming more mature, and three major use-cases are emerging: chat, companionship, and code. But even as coding has reached escape velocity, the only people who can successfully build end-to-end applications today are those who already have some familiarity with backends, frontends, and deployment. While barriers may come down eventually, most amateurs today stay stuck in “localhost” purgatory, because there is no AI-native GUI with an intuitive interface that doesn’t look like a DOS command line. This is why we’re so excited about Wabi. Wabi is a platform that elegantly solves the UI problem by delivering a single platform for consumers to generate, remix, and consume personal software - no syntax required. Wabi allows you to prompt an app into existence, and iterate directly with the UI instead of the code. We’ve already seen people build apps around things like weight-lifting, clip-art generation, and fasting, and early users are amazed by the speed and simplicity of creating software built just for them. Even better, Wabi is a network, which means you can access, use, and modify any app built on the platform. Anything is possible, from creative mini-apps, to personal CRMs, to social applications. Wabi is founded and led by @ekuyda, a formidable entrepreneur who has a track record of being early and right to emerging consumer behaviors. She previously founded @replika, which was years early to the companionship market and boasts 40M users today. She’s intensely committed, wildly creative, and exactly the kind of founder we like to back. We believe we’re moving into an age when software becomes a default mode of creative expression, on the same plane as video, and that Wabi will lead that charge. We have not yet seen an AI-native product demonstrate network effects: most AI tools today are great at single-player use-cases, but don’t look like multiplayer networks. Wabi delivers a compounding feedback loop for software, where creation and distribution are coupled. People used to criticize the creator economy as an investment area because it was hard to see network effects and durability in the space, with the exception of a few massive personalities. Creators themselves have to find clever ways to scale, and Wabi is showing us what a platform for the software-native creator looks like. Importantly, content decays over time while software compounds, and the value of the Wabi platform alongside it. New kinds of creators will emerge, and Wabi will be the place to discover them. Just like we look back on the pre-YouTube era, and marvel that just a few TV networks monopolized content, we may look back on the pre-Wabi era with bafflement that just 20 million developers dictated the behavior of the software that we used. We’re thrilled to be backing Wabi, and can’t wait to see what you make.

Thrilled to announce that we're leading the $9MM Series A in Silo, an operating system and system of record for wholesale food distributors that’s bringing the produce supply chain online. Excited to work w/ Ashton Braun, Antonio Bustamante and the team! a16z.com/2020/09/24/investin…

Thrilled to announce we're leading @happyrobot's Series A! 🤖 HappyRobot is reinventing communication for logistics - starting with AI phone agents that handle track and trace, payment updates, negotiations + more. Why our team @a16z invested 👇 (@omooretweets, @seema_amble)

🚀 Thrilled to announce our investment in @krea_ai! Krea is the platform where creativity and design meets AI spanning image, video, 3D, realtime and more. With 20M+ users from hobbyists to teams at Pixar and LEGO, Krea is the #1 tool for creatives. Art loves entropy! 🎨✨

When we first backed @deel, it was 10 people and a bold idea. Today, they’re serving 35,000+ companies in 150+ countries — and are approaching 3 years of profitability. 6 years later, @bouazizalex and @shuooo continue to execute on their vision. Deel is becoming the default infrastructure for global work.

1/11 - GM friends -> New post below on GenAI + “probabilistic products”; specifically how should product managers+designers think about building on a platform with non-deterministic outputs? More here and below: a16z.com/2023/05/23/generati…

Around every big product idea is a dozen smaller ideas that make it work. By the time the entire playbook is obvious it’s usually too late to catch up. Thus the value of getting input usually far exceeds the value of secrecy.

1/3 SO EXCITED about this - congratulations to the founders, investors, and most of all of my fellow Karmanauts who were involved, both past and present: businesswire.com/news/home/2…

The world is short software. Entire categories of software were never built, simply because of insufficient ROI, high costs, or because the preferences of ~20 million developers dictate the products we all use. But now it’s easy to prototype, build, and ship entirely novel applications, using new app-generation tools. We’re about to see what the long tail is capable of. You might liken this to what YouTube did in 2006. Before YouTube, people would’ve looked at the ~100 cable channels available and thought “this is enough.” But YouTube proved that there were entire niches of entertainment that creators could build a business around. The same thing is happening now for software. Until large models, it wasn't possible for a passionate non-technical creator to build a business around software. Even as storage, CPU, and bandwidth costs came down, software engineering costs remained high. Now we have new app-generation tooling, which dramatically collapses the cost (and the risk) involved with shipping new software. As a result, we’re beginning to see the first fully AI-native creators come to market. Welcome to the YouTube moment for software. Read more on Substack.

The “AI wrapper” hate never made sense to me. Engineers argue, skeptics complain. Meanwhile: – Founders keep building – VCs keep writing checks – Customers keep paying Critics aren't seeing the next steps:

Fantastic, old-school energy this morning at @ycombinator demo day.

Blog post -> How new-school fintech companies are building more durable lending businesses (either direct or affiliate) and moving towards being the “money'' button on your phone: a16z.com/2019/09/12/fintech-…

Eugenia is a legend and an A+ human she’s building Wabi, a very cool platform for app generation + consumption - personal software is coming and this is the way the mass market will produce and consume it

My big idea for 2024? AI finally unlocks voice as a platform for consumers to interact with tech. 👇🏽

Comet is underpriced at $200/mo..

We need TBPN Disrupt

love my friend @signulll who is making a great, albeit orthogonal, point: - yes putting a model in front of proprietary data is very smart and makes companies like Open Evidence enormously valuable vs ChatGPT which simply doesn’t have the data - yes this potentially deepens OpenAIs moat if they capture the data and use it to extend their models HOWEVER - the app/platform feedback loop has always had this dynamic - platforms have an incentive to attract their complements (ie apps) in the near term and an incentive to compete with their complements long term - yet app developers make the deal with the devil because of the scarcity/value of distribution and the favorable early platform dynamics (i.e. get your bag!) - App Store was no different - sure they didn’t have “general intelligence” or the ability to dynamically generate new apps, so instead they commoditized their complements through App Store fees + replicating the top apps as system apps TLDR - the value of being first to a massive new platform will hugely benefit early entrants in the near term and hugely benefit the platform itself in the long term; this opportunity typically only comes once per product cycle so builders ignore it at their own peril

From Series A to Series E - it’s been incredible to watch @Bouazizalex, @shuooo, and the @deel team build one of the most durable, global companies in tech. And they’re just getting started.

there are no functional needs, just emotional ones - people open an app because they want to feel a feeling, not just to get something done the best consumers products today provide emotional interfaces to functional work, and allow functional work to have emotional valence

INDUSTRIES NOT MARKETS - The investing mistake we keep making is looking at categories like AI Code like markets (with one winner) instead of industries with a dozen - As these categories evolve expect to see horizontal winners at every part of the stack and then vertical winners that are specialized - Evidence: Codex, Cursor, Replit etc are all working and winning - The same trend is true is many AI categories - i.e. legal where Harvey seems like a horizontally dominant player while Eve, Crosby and others are running away in their specialized verticals

Poke has gotten a lot of buzz lately and speaks to where AI interfaces are going - an emotional overlay to otherwise functional work like email / calendar / on-boarding etc. Perhaps the next gen interface is less pixels and more personality ….

New post 🚨 -> Using gambler’s fallacy and financial automation to help consumers save: a16z.com/2019/11/07/fintech-…

We’re doubling down on @deel and co-leading their Series E at a $17.3B valuation. Deel is powering global work + achieving exceptional growth and profitability. We're thrilled to continue the journey with @Bouazizalex, @shuooo, and team. Let's GO!!!

3/ Thus the best approach once you've found PM-fit is to prioritize *taking the entire market* - focus on what's working (80%) and leave a smaller portion of your time for new stuff (20%). Many caveats, obv, but a guiding principle for me.

The evolution of consumer fintech competitive advantages: Past: 1. Marketing (we acquire online!) ✉️ 2. Economics (we offer the highest yield!) 💵 Present: 3. Product (we provide the most utility!) 🛠 Future: 4. Networks (better with friends!) 🍻

Music models are musical instruments🎛️ As both a DJ and a consumer investor, I’ve thought a lot about what’s the right product experience for generative music models. I’ve landed on the concept of a “generative instrument” - you’re designing a next gen guitar, synthesizer, drum machine, turntable etc. What are the necessary product features of this AI musical instrument? 1. Audio prompts - audio prompts unlock creativity in making music because people are terrible at articulating the music they love with words - and instead reference other music (everyone knows what “give me a bassline that has Thriller vibes” means). Additionally, producers build songs around single ideas / samples / bars, and audio prompts wrap this workflow perfectly. 2. Feedback loops + extending generations - related to audio prompts, a player must be able to take existing generations and use them (or components) as input into new generations - this is how one would create, for example, a jazzy rhodes piano solo that would fit with an existing generated track. 3. Character consistency - the natural extension of generating a song is generating an album, and that will require having a consistent voice. Perhaps a natural extension of this will also be an “—sref” feature which captures a specific vibe without further prompting. 4. Native stem separation - obvious, but it must be real time, high quality, and as easy as EQing / filtering in existing workflows. 5. In-painting - the idea here is to have models take a direction and generate the details, which closes the gap between a novice piano player and an expert (for example) What’s incredible about these instruments is they are the first musical instruments in human history that don’t require any technical skill to play. They just require taste and passion. In the near future if you love music you will make music - and I for one am incredibly excited for that day.

I joined @stevesi and @eriktorenberg on This Week in Consumer to talk about where AI is actually useful, especially at the consumer layer. A few ideas I’ve been thinking through lately:

Burn in hell @comcast

Voice will be the first way most people meaningfully interact with AI. And for the next generation, voice AI will be the first way most people meaningfully interact with any technology! On the @a16z podcast, we discussed why startups are winning this race vs. incumbents ⬇️

Ding dong, the witch is dead 🤪 With the change of guard at the CFPB we now enter a golden era for consumer fintech product builders - with a fair regulatory environment and new AI primitives that are uniquely suited to enabling fintech: ⁃First reasoning models - these models mean we can come to deterministic answers to complex problems around credit building, tax, debt settlement etc. Language models gave us the heart, reasoning models give us the brains. ⁃Second, model features like “computer use” allow for read/write programmatic access to all your financial accounts - every UI is now an API and builders can do everything from aggregating brokerage accounts to programmatically refinancing credit card debt ⁃Finally a fair regulatory environment where hostile, opaque government forces aren’t dead set on paternalistically “protecting” consumers by preventing innovation and making harmful systems changes with no regard for second order consequences (ie the changes to make credit reports less accurate) IMO The failure of many of the last generation of consumer fintech efforts was two-fold: insufficient product ambition (perhaps because of a lack of primitives) with too much focus on customer subsidies via rewards, loosened underwriting standards, unsustainable yield payments etc and a hostile regulatory environment with a CFPB that consistently made choices which harm the subprime / non prime consumers they claimed to protect. Now is the time to build extreme, weird, fabulously ambitious consumer fintech products that transform the financial lives of consumers. It sounds like hyperbole but the opportunity is there 😀

1 Schwab cutting fees was first positioned as a disaster (look at all the lost $$) and then a victory (look at all the new customers). What really happened? a16z.com/2019/11/22/the-case…

1/5 Though there may be some discussion to be had on long term market dynamics for Neobanks, perspectives like this seem to miss the point. It’s not about neobanks having better underwriting or technology per se: qz.com/1728419/point-72-vent…

Interesting to see @AmericanExpress bundling fintech features into their core experience:

🎈 I'm super excited to welcome @sumeet724 to the A16Z fintech fam! Prior to joining he was at @hellobrigitapp + @nyca, founder @ creature / of, and angel invested in many rad companies like @reserveprotocol, @hello_orum, @getmati, @listentopodz, @streamclubhq, @niradotcom. 🎈

Fintech fiends: This month I write about the uncanny valley of self driving money - the vision is great, but building a bundle of products that actually work better together requires finding product market fit over and over, making this a particularly hard product mgmt problem.

For our 2022 predictions I covered the need for partnerships / product to drive growth in consumer, as the alpha in marketing becomes super hard to achieve: a16z.com/2021/12/20/the-big-… I'm very excited to see new thinking in product! Multiplayer / web3 / commerce are underexplored..

.@wabi is the most delightful consumer product I’ve used in years here is a spooky meditation tracker I generated last night the design language is beautiful, playful, intentional, all in one shot 🎃

Fintech feels like it’s all “fin” and no “tech” - who is building the most interesting product experiences for your $$? I’ll go first: Joy: rate your spending by how it made you feel, the goal is to only spend money on stuff that makes you feels good:

Speaking of multiplayer fintech apps, here is the beginnings of Apple’s approach to finance for families.

1/n Financial services are often though of as a zero sum game - there's a fixed pie and therefore “winning” is finding clever ways to take a bigger piece of it. But the intersection of tech and finance has actually been characterized by positive sum outcomes - a bigger pie!

This week, Deel announced an $800M run rate. Congratulations to @Bouazizalex and the entire team on building the products and culture to unlock this milestone! A fun throwback from when we invested in 2020: a16z.com/announcement/invest…

Introducing...gen AI native workflows - our latest thinking below w/ @omooretweets and @venturetwins .. Our consumer team @a16z is looking for products that uniquely combine generative tech with a deep understanding of user needs. These platforms will move us from "single shot" to full workflow - ushering in a new era of prosumer AI. Gen AI workflow products will be built from the "ground up" around a new primitive - and will combine generation, editing, and composition to let ANYONE do professional-grade work. This includes traditional office workers but also SMB owners, freelancers, artists, consultants, and more!

What are we looking for? Weird x working. We seek products that may look extreme or fringe - many legendary consumer companies start this way. And we believe consumer can't be predicted, only observed. We look for rare cases where magic meets momentum, and invest behind it.

This showcases a strength of Google’s - they have spent 20 years developing a sophisticated IP framework around YouTube where rights holders are given a choice of monetizing or blocking offending content. Almost all (except Prince!) choose the big bag of money.

I’ve been playing records and making music for almost 30 years and the AI x music wave feels as important as the invention of samplers, synths, turntables or even recorded music. I expect the upcoming creative wave means a golden era for AI music that matches 90s hip hop, 40s jazz, 50s/60s rock n roll..

Excited to see @packyM chronicle the @uuuntitleddd story in Not Boring. Read on to learn: - How [untitled] was used to make a #1 hit (Million Dollar Baby) - The app's viral spread among the next gen of artists - Why our team @a16z led the Series A, after originally passing!

I agree with this assessment of the past 10 years but the remedies are probably more nuanced than "have network effects" and "be great at product + growth". If I was building in consumer today I'd think about: Models are not products: So far most "products" have been webpages in front of models, there is tons of alpha in building opinionated + sophisticated products. Transcription, text to media etc aren't sufficient any more. AI native product design: Model/Product != Server/Client ... the properties of models are emergent so PMs have to spend as much time "interviewing" their models as they as they do their customers. Extreme products at extreme prices: The capabilities are so crazy that you can go WAY deeper on a single customer than ever before, and the products should be ambitious enough to support a $1000/mo price point. Emerging moats, channels: Obv moats + channels still matter, and network effects are the gold standard, but consumer AI companies should understand new moats (i.e. LoRAs) and new channels (i.e. voice/phone) work.

What stood out to me from this report is that we’ve moved past a model race. What matters now is habit and distribution. The winners are the products that own default surfaces, earn trust, win distribution, and become part of daily life. Here are my key takeaways:

AI has ignited a consumer renaissance. My friend @kevinrose is a legend in consumer tech. He founded a number of early internet startups, including Digg, where he invented an early predecessor to the now ubiquitous like button. He’s now a partner at @trueventures and recently relaunched @digg. We had a wide-ranging conversation on all things consumer tech, including why I think AI code will be an entire industry, why Kevin looks for “weird and working” products, Kevin’s dinners with Mark Zuckerberg, and how AI products will affect the human experience. 00:00 Intro 02:10 How we met 07:50 AI’s Renaissance for Consumer Products 11:48 Companionship apps and the future of human connection 14:01 The optimistic and pessimistic views on AI relationships 19:18 Poke and emotional primitives 21:05 Weird and working: how to spot great founders 30:05 AI companions and AI in relationships 40:08 The modern AI dev stack and building apps solo 47:25 AI music, creativity, and the next cultural wave 53:50 Curiosity, risk, and finding the next big thing 01:15:00 Always-on recording and social norms in tech

This is what it means (to me) to be Canadian.. 🇨🇦 Happy #CanadaDay2020 🇨🇦

1/ We’ve been living in a software shortage for 40 years. Only a small fraction of people could code, so society underproduced the most elastic good in the modern economy. Now AI is collapsing the cost of software creation. This article was posted four months after chatGPT launched and it’s criminally underrated. It opens the door to a bunch of ideas that I think are important and rarely talked about. (more below)