LLenny's Podcast
← All frameworks
Strategy

Determinate vs Indeterminate Optimism (Betting Under Deep Uncertainty)

Founders need one specific plan; allocators need many bets and a discount on their own forecasts

Difficulty
Advanced
Time to result
~ongoing to results
Steps
4
Confidence
82%

A decision framework for acting during a technological transition when the structural outcomes are genuinely unknowable. It separates two valid postures: the determinate optimist executes one specific, concrete plan, while the indeterminate optimist runs a broad portfolio of bets and stays flexible. Use it to decide whether your situation calls for single-minded conviction or diversified experimentation.

Origin

Andreessen invokes Peter Thiel's 2x2 of optimism/pessimism × determinate/indeterminate. Thiel favored determinate optimists (like Elon: 'I'm going to build the electric car'); Andreessen defends indeterminate optimism as a stronger phenomenon than Thiel credits — not wishful thinking but a deliberate strategy of backing many determinate-optimist founders at once. He pairs it with an explicit instruction to discount your own forecasting ability during the first 10–15 years of a platform shift.

Core principles

  • 01A determinate optimist believes the world gets better because they will do one specific thing
  • 02An indeterminate optimist backs many determinate optimists rather than predicting the single winner
  • 03During big transitions, put a heavy discount on your forecasting ability — early confident predictions are usually badly wrong
  • 04The way to handle unknowable outcomes is to run as many experiments as possible, not to prejudge the structure

How to run it

  1. 1

    Diagnose your role

    Decide whether you're the operator who runs one company (founder) or the allocator who can hold a portfolio (investor). The right posture depends on which hand is on the steering wheel.

  2. 2

    If you're the founder, be a determinate optimist

    Commit to one specific, concrete plan and execute single-mindedly against it — you only get one bet and you actually get to run the company.

    Pro tip History remembers Henry Ford, not the seed investor who also funded ten failed car companies.

    Watch out You don't get to hedge; conviction and focus are the price of controlling the outcome.

  3. 3

    If you're the allocator, be an indeterminate optimist

    Rather than predicting the single winner, back as many strong determinate-optimist founders as possible and make bets along every plausible strategy.

    Pro tip Optimize by having as many good bets as possible, each running as hard as possible.

  4. 4

    Discount your forecasts and stay adaptable

    During a platform shift, refuse to prejudge which company, product category, or moat wins. Treat it as a complex adaptive system, stay flexible, and be open to structural surprises.

    Pro tip Re-read confident predictions from the internet's first decade — almost all were wrong — before trusting today's.

    Watch out Media rewards definitive answers over open questions, so confident-sounding forecasts are systematically over-represented and unreliable.

In the wild

The a16z largest-fund strategy

Andreessen places himself firmly in the indeterminate-optimist camp: back the many bright determinate-optimist founders (Elon and thousands of others), run as many experiments as possible, and don't try to name the single winner in advance.

A deliberate portfolio strategy for a moment when the structural outcomes of AI are unknowable.

Moats you can't yet call

Andreessen runs the moat question as a thought experiment — AI models could become a defensible oligopoly (billions to build, scarce talent) or could commoditize fast (open-source GPT-3 clones, DeepSeek, Claude Code cloned in a week and a half). He concludes you can squint either way and should discount your forecasting.

A worked example of refusing to prejudge structure and instead betting across strategies.

Common mistakes

Indeterminate optimism as wishful thinking

Thiel's critique is that 'things will get better somehow' with no mechanism risks being delusion. Andreessen's defense only works if the portfolio is built from concrete determinate-optimist founders, not vague hope.

Prejudging the winner too early

Stating structural outcomes with false confidence in the first years of a transition — which company wins, where the moat is — is how nearly all early internet predictions went badly wrong.

Is it for you?

Best for

Investors and operators deciding how to allocate conviction vs diversification during the early, unknowable phase of a technology platform shift

Not ideal for

Founders who need single-minded focus — spreading across many bets undermines the determinate optimism their own company requires

From the transcript

determinant optimists which are people who are like no the world is going to be better because I'm going to do this specific thing

1:19:00

the right way to deal with that is to run as many experiments as possible and have as many smart people try to do as…

1:20:30

I need to put like a big discount on my forecasting ability on this one

1:17:30

From the episode

Marc Andreessen: The real AI boom hasn’t even started yet