30 chunks · 23 episodes
If your ecosystem has bilateral agreements across peers, then a layer will show up that centralizes. It's better for that to be an entity with structural interest in long-term health of the ecosystem. That is, one that will not over-extract.
What is the system of record for software production? Until recently it was the codebase in GitHub. But only engineers could speak code. Designers had their system of record in Figma. PMs had their system of record in Notion. All three systems have a push and pull. The center of gravity wasn't ever
...to date. Faux systems of record are non-load bearing. If it's load bearing your system record is, by definition up to date, because it must be to be useful.
Most software requires context to be useful. An app either becomes an island or a system of record. The islands get smaller and smaller, focused on a tiny, separate niche that only needs a pinprick of context. The systems of records grow into aggregators. For consumers, there's no real drive for the
For something to be a system of record, everyone who relies on it needs to use the same system. If it's too complex, then no one will keep it up to date. The map won't reflect the territory. If it's too simple, then it won't be useful. The map won't have enough detail to navigate. There can only be
Every enterprise has a system of record.[bx] Every person has... email—a compost heap. What we need: a garden that grows from that compost. Something we tend, prune, curate. An emergent system of record for our personal lives. LLMs should handle the weeding. We should choose what blooms.
One of my least favorite jobs each year is distilling the family holiday card distribution list. Our family "system of record" is spread across Google Contacts, iCloud, AirTable, Minted.com's list, etc. Every time during the year when I get an email from an old friend about how they moved, I make a
More niches lead to more diversity in the ecosystem, which creates more resilience at the level of the ecosystem. Efficiency flattens, makes it so one genotype can dominate all of the others, reducing stores of adaptive ability.
A tree looks beautiful as an outcome, but it's engaged in a brutal, constant struggle for its existence. The generative process of competition in an ecosystem is fundamentally something that causes death. If the context weren't brutal to survive in, then there would be no impetus to change.
Everyone thinks the ecosystem they control is more open than it is. Because everyone is slightly blind to their own power in a context. The omnipresent wind at your back is easy to take for granted; it never changes.
In an ecosystem, which sub-component will be the one where the network effect coheres and start running away? If there's a good enough, sufficiently open system, that's the emergent schelling point. The point that every participant would be OK with. Even if it's not a strong pull for any individual,
An interesting pattern: using LLMs to astroturf content in an ecosystem. The challenge of an ecosystem is not so much the hill climbing of quality, it's the creation of the ecosystem to a critical mass. And now you can astroturf it!
Everyone assumes that the most powerful AI service will be created by a single entity. Creating a single, high quality model makes sense to be best done by a single entity. But if you could figure out a way to use a model, combined with an ecosystem, to produce the overall quality of the service, th
Hype is the emergent kayfabe in an ecosystem. Don't get caught up in the hype. Hold it at arm's length. Surf it.
"Control" and ecosystems don't match well. They're alive! They can be guided but not moved. Your choice as a company is: 1) a very large, wild garden you don't control, or 2) a very small, tidy thing you have control over. The garden takes a very long time to grow, which might be longer than the run
An ecosystem is intelligent. It's a different kind of intelligence: collective intelligence. The hotter it runs, the more the ecosystem can do, the more it can think. Swarm / collective intelligence is kind of alien to us ("where do the decisions happen?") but can be significantly more powerful than
The whole point of an ecosystem is open endedness. The safer the composition of untrusted components, the more open-ended the system: the larger the combinatorial possibility.
Ecosystems that grow fast enough are hard for incumbents to tackle. By the time another bigger player notices and decides to tackle them and decides a coordinated strategy to do it, the ecosystem has already grown to be too big for them to wrestle them to the ground.
To steer an ecosystem you need an intuition of the particular emergent game theory. To have lived through it in multiple iterations, to have been an active player, not a passive participant. To have accumulated the knowhow in your bones.
Creating the seed of an ecosystem is like creating a frankenstein. You assemble a lot of different components and hope that when stitched together in just the right way they'll have the spark of life. What matters is not the quality or coherence of the individual components; it's whether the whole c