the VC almanac

> be 2009 > two guys decide to start a venture capital firm > problem: every top VC has decades of reputation > Sequoia backed Apple, Cisco, Google > how do you compete with that starting from zero? > one guy co-invented the web browser > the other took a company public with $2M in revenue > press called it "The IPO from hell" > he sold it for $1.6 billion > they raise $300M year one > $650M year two > $1.5B by year three > start backing unknowns like Airbnb, Coinbase, Instagram, Slack > every single one becomes generational > Coinbase alone returned $7 billion to their investors > by 2025 they captured 18% of ALL venture capital dollars in America > January 2026: raised another $15 billion in one sweep > now managing over $90 billion > 1,076 portfolio companies > more unicorns backed since 2020 than any other firm on earth > oh and the CEO guy? > started his career as a rapper > runs a nonprofit giving pensions to overlooked hip-hop pioneers > wrote the most famous product management memo in tech history > out of frustration with his own team > his life philosophy? "Life isn't fair" > his firm? Andreessen Horowitz > a16z tl;dr two guys with zero VC reputation in 2009 built the most dominant venture firm in history by backing founders everyone else ignored. 16 years later they manage $90B, back 1,076 companies, and captured nearly 1 in 5 venture dollars in America.

@acenario and @abhijay_cloaked just raised $375M for @keepitcloaked to fix the internet's biggest privacy problem. In this episode, they take us back to the beginning and share the untold stories about everything it took to get here. Episode 04 drops tomorrow.

@saranormous cc: @steipete

More about "Dobby the Elf" with @saranormous. "I used the agents to find all of the smart home subsystems on the local area network. I just told it I think I have Sonos at home, can you try to find it? It logged in and said oh yeah, you have these Sonos systems installed." "It did some web searches, found the API endpoints, and asked me if I wanted to try it. I'm like, whoa, you just did that. Can you try to play something in the study?" "It did the same for lights. It hacked in, figured out the whole thing, created APIs, created a dashboard so I could see the command center of all my lights in the home. Then it was switching lights on and off." "It controls all of my lights, my HVAC, my shades, the pool and the spa, and also my security system." Video: @NoPriorsPod

Alex Karp just warned Silicon Valley: you're about to get nationalized. "If Silicon Valley believes we are going to take away everyone's white collar job — highly educated people who went to elite schools, who vote for one party — and you're gonna screw the military, if you don't think that's gonna lead to nationalization of our technology, you're retarded." "And you might be particularly retarded because you have a 160 IQ. But this is where that path is going." "You cannot have technologies that simultaneously take away everyone's job and then be perceived as screwing the military. The danger for our industry is that you get a famous horseshoe effect where there's only one thing people agree on — and that's that this is not paying the bills and our industry should be nationalized." The $18B CEO of the company that built the Pentagon's intelligence systems is telling Silicon Valley point blank: work with the government or lose your company.

Palantir CEO Alex Karp on the best outcome for AI in America: “We show our adversaries: you can’t fuck with us.” Via @tbpn

Alex Karp: "We should publicly humiliate people who don't support our war fighters." "We have American war fighters on the battlefield, willing to sacrifice their lives. Some of whom have sacrificed their lives. They have families and kids. The kids and families don't know if their loved ones are coming home." "People who are not aware or somehow so spoiled that they don't realize what these people do for us, we should publicly humiliate them. Everyone has a role in that." "Especially those of us who are riding a crest of intergenerational cultural, intellectual, courage advantage, which it is to be an American nowadays, should not forget the war fighters who are disproportionately from the middle of our country and disproportionately have gotten screwed." "I'm proud, very proud, that at Palantir we get arrows. Half the people attacking us, it'd be good if they spent two minutes on Wikipedia at least learning the talking points they're regurgitating." video: @a16z

The difference between "similar" and "relevant" is why most enterprise RAG deployments quietly fail at scale.

We've raised $6.5M to kill vector databases. Every system today retrieves context the same way: vector search that stores everything as flat embeddings and returns whatever "feels" closest. Similar, sure. Relevant? Almost never. Embeddings can’t tell a Q3 renewal clause from a Q1 termination notice if the language is close enough. A friend of mine asked his AI about a contract last week, and it returned a detailed, perfectly crafted answer pulled from a completely different client’s file. Once you’re dealing with 10M+ documents, these mix-ups happen all the time. VectorDB accuracy goes to shit. We built @hydra_db for exactly this. HydraDB builds an ontology-first context graph over your data, maps relationships between entities, understands the 'why' behind documents, and tracks how information evolves over time. So when you ask about 'Apple,' it knows you mean the company you're serving as a customer. Not the fruit. Even when a vector DB's similarity score says 0.94. More below ⬇️

"Artificial intelligence would be the ultimate version of Google." Larry Page in 2000. "So if we had the ultimate search engine, it would understand everything on the web. It would understand exactly what you wanted. And it would give you the right thing." "It would be able to answer any question, basically, because almost everything is on the web. So we're nowhere near doing that now. However, we can get incrementally closer to that. And that's basically what we work on." "If you printed out our index, it would be 70 miles high now. We have about 6,000 computers. We have enough disk space to store 100 copies of the whole web." "From an engineering, scientific standpoint, building things to make use of this is a really interesting intellectual exercise."

What is a VC? "Most of these people we call VCs, they're not actually rich people with a lot of money that want to make some money. They're vehicles that raise money from institutional investors that have to return a certain amount of capital in a given amount of time." "If I'm a professional VC and I have a $300 million fund, my mandate is to 2x or 3x that fund over the course of seven years." "As an entrepreneur, it's very important to understand what the person on the other side of you is actually looking for. And what they're looking for is to return a fund of a certain size."

Anduril’s @PalmerLuckey on AI and defense: It’s not up to unelected Silicon Valley executives to decide US policy. “Don’t try to control policy by refusing to give the government the best tools.” “You’ll see a lot of companies saying, ‘Here’s where we draw the line. This is what policy should be. This is what we’re going to allow the government to do and not do." “That’s not the place of Silicon Valley companies. They shouldn’t be telling what their policy is.” “Build the technology, advocate for policy.” “The US government and military make better decisions when they have better information.” “It’s not up to us.”

Palantir CEO Alex Karp on controversial uses of AI: “Do you really think a warfighter is going to trust a software company that pulls the plug because something becomes controversial, with their life?” “The small island of Silicon Valley that would love to decide what you eat, how you eat, and monetize all your data should not also decide who lives in a country and under what conditions.” “The core issue is who decides?”

JPMorgan CEO Jamie Dimon on AI and the future of work: "We use AI for risk, fraud, marketing, underwriting, note-taking, idea generation, error reporting, reducing errors. There are 600 use cases. 50 I'd put in the important category." "It's no different than the past. If we can use it to do something better, faster, quicker, cheaper, higher satisfaction from the customer, we are going to do it." "People talk about the negatives. My guess is, in 30 or 40 years, your kids are going to be working four days a week. Maybe three and a half days a week. Living to 120." "A lot of cancers will be cured. A lot of disease will be cured. Food will be safer. Cars will be safer. It will be a wonderful thing. And I think that's legitimate." JPMorgan has 600 AI use cases in production. Video: @BloombergTV

Parker Conrad on what people get wrong about software: “People think great software comes from focusing very narrowly on one specific domain and crafting an artisanal product around it.” “But companies don’t just use one application. They use many. And that creates a different problem.” “The real power in business software comes from the underlying capabilities that keep showing up everywhere: permissions, reports, analytics, workflow automations, approvals.” “If you’re building across many applications, you can afford to invest deeply in those foundations and create a much better experience for customers. If you’re only building one thing, you usually can’t make that R&D lift.” “This actually isn’t a new idea. It’s how Oracle, SAP, Salesforce, and Microsoft were built as platforms first, with shared capabilities that could power many different applications.” video: @NoPriorsPod @saranormous @parkerconrad

Marc Andreessen on why NVIDIA's dominance is an accident of history: "Intel had a monopoly on CPUs. NVIDIA fought the GPU wars for 30 years and came out the winner. It was a hyper-competitive market for graphics processors. Actually not that high margin. Actually not that big." "Then it turned out there were two other forms of computation that were incredibly valuable, that happened to be massively parallel, which happened to be very good fits for GPU architecture." "Those two highly lucrative applications were cryptocurrency starting 15 years ago, and AI starting about four years ago." "If you were designing AI chips from scratch today, you wouldn't build a full GPU. You would build dedicated AI chips that were much more specifically adapted to AI. Much more economically efficient." NVIDIA didn't design chips for AI. AI just happened to fit. The next chip company might not get lucky. They might build it on purpose. video: @pmarca @a16z

Zach Yadegari on why you should build apps in 2026 and why it's the best thing to build right now. "100%, it's the best time ever to build. It's becoming easier and easier, even if you don't know how to code." "These vibe coding platforms were really bad when they started. Now they're actually great." "My friend Rourke, who I went to high school with, built an app on Replit called Fight AI. He released it two months ago and it's now doing $30,000 a month." "He's just a normal kid in college. Very hardworking. Didn't know how to code. Learned how to market on TikTok organically. And now he's killing it."

Chamath on Mamdani canceling gifted education: "Prove to me you're an adversary of America without saying you're an adversary of America. In what world does that make any sense?" "If you're a right tail outlier intellectually, you have the ability to take gifted classes. It's for the smart kids. I wasn't a gifted kid, but I went to school with them. They're really smart. We should have a world that embraces those kids. Instead, they canceled it all." "You're gonna let them sit and stew in the same classes as the rest of us. We know how that goes. They get frustrated, they misbehave, they don't achieve anything. That's not what we should be doing. We should be unlocking these kids to achieve their potential." "You don't take an extremely good athlete and make them play two grades beneath. It's the opposite. You take the best athlete and say, play two grades above. Test yourself. Go to the limit." "You don't take an incredible musician and say, only play these four notes. None of it makes sense. Yet we're doing that with core structural intellectual development." @chamath

Inside ASML: The $350 million machine that makes every modern chip. "This is the most advanced machine humans have ever built. Decades of development. Billions of dollars. All to get this humongous beauty." "If you saw pictures on the internet of the first chip printed at 8 nanometers, that was this machine. The smoothest object on earth. It's all in here." "It's cool to realize that every smartphone nowadays has a chip made with a machine that was actually put together here." "You need such a big machine, so much infrastructure, to make the tiniest things we can make at scale. It's inversely proportional. The smaller you want to go, the larger everything around it becomes." One machine. The entire modern world infra runs on it.

Turing Award winner Leslie Lamport on why learning to code is like learning to type: "If people are trying to learn programming by being taught to code, they're being taught writing by being taught how to type. That doesn't make much sense." "The best way I have for teaching programming, as distinct from coding, is to think about what the program is supposed to do mathematically." "There's a very big practical problem with this. The mathematical education in this country is pretty terrible. Most people wind up being afraid of mathematics. This is even senior programmers." "I've developed a language called TLA+ for writing down the ideas that go into the program before you do any coding. It's a hard thing for engineers to get into, but when they do, it develops their ability to think mathematically." Learn stuff that would challenge you to think differently, as @elonmusk says.