The “Rocketship” Era — Going Big on Startup v2
🚀

The “Rocketship” Era — Going Big on Startup v2

Those alive today will be among the first to settle space and live there permanently. This is a step change that will happen exactly once from now until the end of life. Once we have footholds beyond spaceship Earth we can expect growth to be geometric, if not in human lives, then in our reach and resources.

Therefore, those alive today have a literal once in a universe opportunity to have impact before the seed of humanity sprouts and grows into a forest. Being part of that initial seed means we get to shape everything downstream; every action taken, what we buy, the words we say, the business we do — it will echo for the next X trillion years in the gestalt of humanity.

So, the next time you hear a grumble about unused or misused privilege consider the privilege of our generation from our ancestors and our descendants. Consider this unique opportunity, consider that we are living in the most objectively abundant time that has ever existed for any life in the last 3.7 billion years. Then focus your energy on how you are going to exploit this once in a universe opportunity.

If the upside doesn’t motivate you then consider that Earth is default dead and so life is default dead. If we do nothing all life will end.

Given these perspectives, I view it as our duty to contribute with our maximum potential to the long term future of life. How? There are ↗️Three Ways to Contribute to Humanity and, for myself, I’ve decided that development of space is underserved, so I’ll work with the domestication of the solar system as my north star.

For more see 🪖The Duty of the 1%, 🦃Don’t Be a Turkey — The Future is Space and 🤯The Meaning of Life in 5 Minutes.

The Ingredients for Self-Reproducing Systems

In 🌍A Better Way to Work on Space Today I discuss what the real bottleneck to terraforming space is — self-reproducing systems that solve labor chains end-to-end.

It boils down to these ingredients:

As Kurzweil said, the toughest part about invention is the timing. So, I dedicated the last decade to becoming a reliable founder and waited for the soil to grow fertile. Now, foundation models have reached a tipping point. They’ve become useful for “good enough” use cases.

These foundational models will enable the next generation of FAANG companies to build useful applications but the companies that do this with a broader vision in mind will dominate the long term future. I’m looking ahead to the real opportunity — self-learning.

Enter Von Automata

I’m code-naming my next venture VonAutomata in honor of Von Neumann, who theorized universal constructors, automata, and Von Neumann probes. Here’s the overview:

  • Mission: domesticate space with profitable terrestrial intermediaries
  • Evolutionary Path:
    • enable indie developers to build AAA experiences by leveraging genAI with open-source tools
    • research and test self-learning techniques
    • continue until building games is like working with an AI consultancy
    • generalize to other labor chains
    • translate to the physical world
    • translate to the space environment

Tools that talk back and use natural human interfaces like demonstration, example, and language will be the next wave of software — universally accessible software. The world is already heading that direction my goal is to leverage that momentum, accelerate it, and turn it towards important problems (like space development) down the line.

The Immediate Opportunity

🫡
collapse AAA-like game development to a single-person

Long term plans aside, there is a clear opportunity ahead of us enabled for the first time by technological advancements in mid 2023 in open source foundational models.

The landscape everywhere is changing and it is heading towards indie. Why? Because better tools democratize access. As part of this mega trend game dev is naturally collapsing into indie shops.

While the big players are scrambling to adjust their production pipelines no one is properly serving the indie market. Indie developers are growing in number, looking for simpler tools (evidenced by the rise of Roblox and simpler open editors like Godot) but still they are underserved. Meanwhile the game market is growing rapidly.

So, I’m going build an AI-first editor to enable indie studios to create AAA quality games because:

  1. I have a clear go-to-market path
  2. the market opportunity is huge — the broad game market is growing at 20%+ CAGR and will only accelerate as the world comes online
  3. the indie dev market is underserved — new game developers are overwhelmed, AI tools for open-source tools are asked for
  4. indie dev market share is growing disproportionally — indie is the inevitable future; traditional game dev shops are the newspapers, book publishers, and TV networks — this is underscored by the rise of simpler tools like Roblox and Godot
  5. the opportunity to 10x productivity have been enabled by tectonic shifts in open source gen-AI in mid-2023
  6. the stakes and cycle time are low which means fast iteration speed is high
  7. games are a great test bed for next-gen no-data ML methods like imitation learning, self-learning, etc. through autonomous agents at every level from NPC to human roles like play testing, animating, art, writing, to game design and balance
  8. optionality is preserved:
    1. it’s a rich idea space with plenty of pivot opportunity
    2. most of what I do in the next six-months will translate
    3. it’s adjacent to practical applications in AR and robotics
    4. games have 🍲The Ingredients for Limited Machine “Consciousness”
  9. I am positioned to execute via a fusion of
    1. relevant AI/ML chops
    2. software chops and
    3. prior founding experience
    4. background in game design
    5. some exposure to game development

Connecting the Dots

If I had a magic wand and could control the future the evolutionary path of the venture would look like this:

  • Lifecycle 1: build the kernel of self-learning digital employees by collapsing game development (5-10 years)
  • Lifecycle 2: translate to physical robotics beachhead (5-10 years) — deep sea mining is a great analog to space development
  • Lifecycle 3: encompass the entire chain of physical labor from mining, manufacture, to labor (10-30 years)
  • Lifecycle 4: translate to space (X years)

The great thing about this is while this moves us in the direction of settling space it create prosperity in the near term as a general capability.

This path seems clear, all that’s left is to walk it and see how it shakes out.

Update 4/2024: decided to skip games and head directly to robotics.