andrew chen

every VC eventually starts a substack and a podcast and every substacker and podcaster eventually becomes a VC 😂

AI is makes software cheaper to build, but not easier to distribute This is the core conflict for years to come

AI moves the goalposts for raising preseed capital: less “tell me your product idea” more “show me what you built and how customers react” In the 1990s people raised $5m+ just to build the v1 of their website. In 2000s that became $2-3m. Today it’s just tokens

Who’s working on this idea: Openclaw for personal finance - integrates w all your banks/cards/etc - understands tax returns and filings - monitors portfolio and competitors - digests proprietary data sources (credit card panels, app rankings, and etc) - reads company news and X Etc etc

talent is unevenly distributed, but AI code gen tools are also unevenly distributed! wait until there's access for everyone... software explosion the way we've had an explosion in video, photo, text creation

lesson from 20+ years working in startups: most founders don’t have a growth problem. they have a product problem disguised as a growth problem

everyone's debating whether AI will take their job while quietly using AI to do their job better. the disruption already happened, you just didn't notice because you were the one doing it

i asked my openclaw what I should be focused on rn in the zoomed out view, and here's what it said: Figuring out what VCs even do in an AI world — if AI agents can source deals, write memos, do diligence... what's the irreducible human core of investing? You're living that question with me sitting here. lol

2026 AI startup filter: Can this AI product survive if base model quality improves 10x in 18 months?

your product gets better every week, but your CAC gets worse every month Can your product get good enough to generate organic growth before your marketing channels saturate with scale? AI is going to supercharge this conflict. We’ll write even more and code and ship even more features faster. But the marketing environment is even noisier and more chaotic as AI-led products launch and AI-generated content fill the feed

the prototype is the new PRD ⁠If your team needs a 20-page product strategy doc, you’re already behind someone with a weekend prototype The ability to feel how good a product is, from actually using it, beats all the theorizing and market analysis and user research As they say -- sometimes you just know it when you see it. This applies to product experiences too

new role opening up on the a16z speedrun team: come work with me and @SamiraBehrouzan - this is directly on the our speedrun team where we work with hundreds of founders. Dream job for the right person 😎 — Some notes from Samira on the top: IF YOU WORK IN EVENTS AND LOVE BUILDING COMMUNITY, this job is for you. The speedrun team is growing and I'm looking for an events & community marketing ace to join my team. What is this role exactly? In many ways it's what I've been doing since speedrun launched three years ago, and I'm genuinely excited to pass the baton. You'll be the heart of the program. The producer behind our entire series of events. The person who makes sure every founder has everything they need to get the most out of their experience. And over time, you'll become the face of speedrun, the person founders turn to with their questions, their feedback, and their trust. This role is events heavy, ops heavy, and people heavy. You need to be someone who moves fast, stays organized, and leads with empathy every single day. What makes this opportunity truly special though is the community you'll be joining. You'll work alongside some of the smartest people I've ever met, and you'll get to build real relationships with founders who will become lifelong friends. If this sounds like your dream job, apply at the link below and feel free to DM me with any questions. I'd love to find the Stark / Fern to my Frieren who are ready to go on this ride with me till journey's end.

AI will make teenage billionaires a thing why? People have long argued that we'll see more 1-person unicorns from AI, as the tools allow a singular visionary to scale themselves. That's obvious. But it also follows that AI software will be created and run mostly by young people, and billions will be made in the process AI also makes coding and product building as easy as making videos or posting a photo. The data says that the time rich (rather than money rich) dominate internet usage -- after all, they have more time to consume, and more time to create. Because the time rich are mostly composed of kids and teenagers, it means that software will mostly be built and used by young people. Think social media, Roblox, chat, etc, but coded by teenagers for teenagers we're approaching a timeline where instead of going to school, getting a job, getting married, and having kids -- the linear form of life we've had for 100+ years -- we will see kids go start to building products and starting companies from a young age. This means huge disruption for universities and society, but also that we will see people build build build. A good thing IMHO. The future will be built by young people, for young people, and billions will be made as a result

vibe coding: 100,000x faster vibe debugging: 1x? 2x? Reminds me of the windows 95 progress bar that gets stuck at 99%… that’s the last part of any Claude coding project lol

after seeing many many AI data rooms while also being in the middle of the hype vortex (like many of you) -- here's an observation: The AI hype/real debate summed up: bulls point to $150B in funding, frontier model breakthroughs, and amazing AI products. Bears point to retention curves and revenue gaps and "GPT wrappers." Both are right. The tech is real AND most companies building on it will fail. These things aren't contradictory -- but for sure our world is being reinvented.

funny lesson from my time at Uber is that: - Building a new project within your own team is easy - Building it between two teams is possible but hard - Building projects that require three or more teams is impossible (Big company problems, I know…)

Today the job of the PM is the define the product, how it works, and how it’ll get built But we won’t need that soon The future job will be simple to define the goals, the constraints, and long term strategy - and letting the AI figure the rest out Goal Architect, not product manager

⁠ai won’t replace you. a person using ai who ships 10x faster will :)

SF feels designed in a lab to make startups: - nothing really to do (except code) - dinner convos become pitch practice - more dogs than kids - social status determined by funding amounts - bars close at 10pm (but GitHub is 24x7) - “what are you working on?” instead of “how’s the family?” - nice weather but not too nice - gender ratio imbalance (odds are good but..) - only city in the world where it’s a flex to have a mattress on the ground - pro tizz anti rizz

a year after DeepSeek shocked everyone, there are now dozens of serious Chinese AI labs racing to release models (Kimi, Qwen, MiniMax, seedance, Kling) the AI competition everyone expected between 2 Silicon Valley companies is actually between 2 countries lol - and meanwhile Europe is picking a fight with Big Tech right as Washington is backing off regulation, and China doubling down on their ecosystem What a world