To aid coordination, technology trees help decentralize science and provide an overview of the field, existing work and open challenges.
Have you ever played the game series Civilization, created by designer Sid Meier? Through the years, much has changed, but one of the unchanging hallmarks of the series has been the technology tree. Why has it been such a stable component of this game? Because it allows you, on one look, to get a bird’s-eye view of the technological capabilities necessary to make progress on your audacious civilizational goals.
Compare this with our real civilization. If we wanted to, we could probably map the many technological capability paths that got us to where we are today. After all, our current tech stack is what the Civilization tech tree is modeled after. What if we could build a tech tree that was future-facing, starting now? Reality is, arguably, more complex than a computer game. So, rather than mapping civilization at large, perhaps we could start with individual technology areas and map those out, one by one. Within technology domains, one could break down the main goals for the field into future capabilities required to get there and recursively work ourselves backward to the present capability stack.
Even if it’s possible, what’s the point? The point is that, apart from being an intellectually interesting endeavor, it may well dramatically speed up progress. Imagine you’re a funder, or talented postdoc, an entrepreneur in residence, or an advocacy leader looking to advance your technology area of choice. Currently, it’s pretty difficult to figure out how to plug in. Even after graduating in that field, digesting much of its literature, drawing on interviews, and online courses, it’s not very intuitive to see how to connect the dots within an area in a way that would advance the field. There is plenty of information out there, but without a scaffold to map the context and dependencies of different opportunities, one can only guess that the one you’re zooming in on is actually a crucial bottleneck in the field rather than an irrelevant detail that stands to be solved by an approaching technological innovation upstream from that area.
A dynamic overview of a field would make it easier to coordinate efforts, find and fund undervalued areas and determine how, together, they unlock novel capabilities and applications.
Tech trees: The reality
So far so good, in theory. Could this work in practice? At Foresight Institute, we’re trying to find this out. Foresight operates five technical programs:
Decentralized computing, focused on secure cooperation, chaired by Mark S. Miller, chief scientist at Agoric.Molecular machines, focused on atomic precision, chaired by Ben Reinhardt, PARPA.Biotech and health extension, focused on rejuvenation, sponsored by 100 Plus Capital.Neurotech, focused on brain-computer interfaces and whole-brain emulations, chaired by Randal Koene, CarbonCopies.Spacetech, focused on space exploration tech, chaired by Creon Levit, Planet Labs.
These programs come with expert groups of around 200 scientists, entrepreneurs and funders per group cooperating to drive long-term progress, supported by workshops, fellowships and prizes. To face the problem of onboarding the growing number of new coming enthusiasts into these fields, in early 2022, we decided to create tech trees to map each area.
Led by domain expert interviews, this pioneer team is now building tech trees of each field, starting at the state of the art, mapping each to long-term goals with conditional nodes, one branch at a time. At the end of Q1, we completed the first tech tree prototypes.
Rather than engaging in armchair philosophy, our tech tree architects are developing the tech trees through discussions with domain experts working on each node. Cycles of feedback will lead to iterations of the tree until we get a clear picture of the field. Once v1 is complete, we’ll open up the trees to crowd-sourcing.
Each node will be clickable, allowing people to zoom in on any particular node to see relevant companies, advocacy groups, labs and independent projects. Others will want to know which open challenges need incentivization through funding. Researchers can submit challenges to make progress in their domain. We can set bounties and prizes on bottlenecks to incentivize progress.
Tech trees: The potential
Membranes separating the trees may prove to be pretty permeable. For instance, the computing tree, with tools such as privacy-preserving machine learning, will have something to say about the longevity tech tree. The molecular machines tech tree, with tools such as unclonable polymers, will be relevant for the encryption technology stack in the computing tree. All of them will inform our future in space from material and energy advances through molecular machines to human capabilities strengthened by longevity and neurotechnology.
As the branches of different tech trees start nestling with each other, risks will also become more apparent. Advanced artificial intelligence will be a major revolution and risk vector throughout all trees. But technologies to mitigate risks, such as computer security, will also become more visible and, thus, fundable. This could increase funding for “differential technology development” — i.e., the development of civilizational safety-enhancing technologies over those that are risky.
Some pioneers may want to coordinate on desirable paths through the forest of trees, such as this civilizational map proposed by Trent McConaghy. Others will want to specialize in advancing the frontiers of their local domain, company or project, such as Balaji Srinivasan.
What comes after VC thesis statements and portfolio pages?
The tech tree. A constantly updated open source map of everything you want to fund, how it interrelates, and what remains to be built. A vision of the definite future that still accommodates uncertainty and creativity. pic.twitter.com/s4yyRvcV6Z
— Balaji Srinivasan (@balajis) November 4, 2021
Tech trees allow a variety of pioneers to compare notes and accelerate progress across the boards.
Such a long-term project may sound naive from where we are today. One reason is that we have suboptimal tools. To solve this, we co-hosted a hackathon to build an app for better crowdsourcing and crowdfunding of such maps, together with Srinivasan of 1729.com; Evan Miyazono of Protocol Labs; McConaghy of Ocean Protocol; Amir Banifatemi of XPrize; and Seda and Matthias Röder, and Andy Smolek of Sonophilia. The top submissions are now collaborating on future road-mapping efforts through MapsDAO.
Finally, trees take time to grow. But the earlier we seed them, the earlier we start the many cycles of iteration required to harvest their fruits.
This article does not contain investment advice or recommendations. Every investment and trading move involves risk, and readers should conduct their own research when making a decision.
The views, thoughts and opinions expressed here are the author’s alone and do not necessarily reflect or represent the views and opinions of Cointelegraph.