a16z Show

The 10 best a16z Show episodes

The episodes that are not about this quarter. a16z publishes close to daily and most of it is commentary on the AI news cycle; ranked here are the ten that teach a mechanism, a history, or an operator's method instead.

New to a16z Show?

Half of these assume you already follow the firm's house arguments — the entry-point list is a cleaner first hour.

→
  1. 1
    The computer science one

    How Bitcoin Rewired a Classic Computer Science Problem

    In 2007 a workshop on Byzantine fault tolerance largely concluded the technology was unnecessary and far too slow. Every major chain now runs a version of it. Tim Roughgarden and Ittai Abraham trace how Bitcoin attacked the Byzantine Agreement problem from an unexpected direction and how decades of academic consensus theory converged with what shipped. Twenty-two minutes, and the densest of them on this list.

  2. 2
    The one with the number that lands

    Inside Palantir: Building Software That Matters with Shyam Sankar

    In 1989, 6% of major US weapons-system spending went to dedicated defence contractors. Today it is 86%. Shyam Sankar builds his whole argument on that shift — Chrysler once made minivans and Minuteman missiles, consumer demand subsidising the industrial base — and on the "Last Supper" consolidation that took the primes from 51 to 5 and turned growth into buybacks.

  3. 3
    The financial history one

    Building Blackstone, Backing Costco, with Tony James

    Tony James joined DLJ in 1975 when it had no recent deals, spotted the leveraged buyout after KKR's Houdaille in 1980, and built a merchant bank whose first fund returned a 90% IRR. He also led the Series A into Costco. He then sold to Credit Suisse in 2000, and he is specific about why he read the top correctly. Practitioner history rather than reminiscence.

  4. 4
    The case study

    Inside Cursor: The Anatomy of a Generational Startup

    In early 2024 the consensus was that Microsoft would take AI coding on platform power alone. Cursor made two non-obvious calls — fork VS Code rather than build a plug-in, and refuse to train a foundation model — and the investors who backed it reconstruct the reasoning. It works as a case study in what a contrarian bet actually looks like from the inside before it is obviously right.

  5. 5
    The one about risk, from someone who ran it

    Lloyd Blankfein on Risk, Crisis, and Leadership

    Blankfein's account of risk management is not a framework but a bifurcation: push the desk to take more risk exactly when fear has set in after losses, and pull back roughly a third of the time when the book is overcommitted. His argument against prediction and for contingency planning has aged well through several crises he was actually in the chair for.

  6. 6
    The operations one

    The SpaceX and Tesla Playbook for Hard Tech Startups

    A former Starship propulsion lead and a former Tesla battery supply chain lead, both now running hardware companies, on what actually transferred. Their answer is unglamorous: flat structure to keep information moving, junior engineers with direct access to executives, and high-conviction leaders who decide fast — with the silo problem arriving reliably at about a hundred people.

  7. 7
    The reframe

    Tyler Cowen & Alex Tabarrok on AI, Jobs, and Economic Growth

    Tabarrok's framing carries the episode: AI creating 50% unemployment and AI halving the working week are close to the same fact, and which one you get is a question of distribution and timing rather than of technology. He and Cowen then go to the historical record on how much control societies have actually had over that split.

  8. 8
    The odd one out

    Rick Rubin on AI, Creativity, and The Way of Code

    Rubin became the accidental mascot of vibe coding, so he wrote a version of the Tao Te Ching for programmers, knowing nothing about programming. The joke is the setup; the argument is that a tool has no point of view, and that people designing the future might use a three-thousand-year-old text about balance. Unlike anything else in this feed.

  9. 9
    The nineteen-minute one

    Don’t Follow Your Passion | Ben Horowitz’s Advice for New Graduates

    Horowitz's 2015 Columbia commencement address, preserved here: follow your passion is probably wrong, because passion tends to be a product of mastery and contribution rather than an input to them. The version of it he lived is the story of announcing a computer science major after reading about Turing's universal machine, to general discouragement.

  10. 10
    The market-structure one

    Why $1B Exits are Dead

    The top 1% venture exit went from $10 billion in 2020 to $32 billion, which changes what a fund has to own to return capital at all. David George and VenCap's David Clark work through the arithmetic, including the observation that Anthropic and OpenAI are adding revenue per month faster than Meta, Google or Microsoft while real-economy diffusion is still under 5%. The dated numbers do not spoil the structural point.

What's missing, and why

The Marc Andreessen conversations. There are a lot of them, they are the reason many people subscribe, and they are the definition of a thing you listen to the week it lands. They are all on the show page in date order; none of them is what you go back to a year later.

Read one before you commit an evening to it

Every episode above links to its summary here — the argument, the chapter outline and the transcript, free and without an account. PodSized has 215 a16z Show episodes summarized; the full episodes stay in whatever podcast app you already use.

Frequently asked questions

What is the best a16z Show episode? ↓
How Bitcoin Rewired a Classic Computer Science Problem. In 2007 a workshop on Byzantine fault tolerance largely concluded the technology was unnecessary and far too slow. Every major chain now runs a version of it. Tim Roughgarden and Ittai Abraham trace how Bitcoin attacked the Byzantine Agreement problem from an unexpected direction and how decades of academic consensus theory converged with what shipped. Twenty-two minutes, and the densest of them on this list.
How were these a16z Show episodes chosen? ↓
The episodes that are not about this quarter. a16z publishes close to daily and most of it is commentary on the AI news cycle; ranked here are the ten that teach a mechanism, a history, or an operator's method instead. The ranking was made by reading the 215 a16z Show episodes PodSized has summarized — a large slice of the catalogue, not every episode ever released — rather than by download counts, which we do not have and which measure release-week promotion more than quality.
I have never listened to a16z Show. Should I start here? ↓
Half of these assume you already follow the firm's house arguments — the entry-point list is a cleaner first hour. There is a separate page of five entry points for a16z Show, in the order a newcomer should hear them, with an answer to whether episode 1 is one of them.
Why isn't "The Marc Andreessen conversations" on this list? ↓
There are a lot of them, they are the reason many people subscribe, and they are the definition of a thing you listen to the week it lands. They are all on the show page in date order; none of them is what you go back to a year later.
Can I read these instead of listening? ↓
Yes. Every episode above links to its summary on this site — a structured text summary with a timestamped chapter outline, free and without an account. The full episode stays in whatever podcast app you already use; PodSized summarizes episodes, it does not play them.

Keep going

Where to start with a16z Show

Five entry points in order, and whether episode 1 is one of them.

→