Superhumanism is a new ideology that is founded on the belief that successive advancements in AI - from right now all the way through the singularity - should be leveraged towards radically empowering as many individuals as possible, at each step of the way. We believe that this is the key to ensuring a free and flourishing future for humanity, in both the short-term and the long-term, as AI continues to advance.
Superhumanism is founded on the following 6 tenets:
1. Successfully directing AGI is all that matters for humanity’s future.
We believe that successfully directing AGI is THE all-deciding event for humanity's future, that we will reach a "point of no return" for this in the next 1-2 decades, and that this is inescapable (and therefore we NEED to go on offense to get it right - playing defense is not an option).
2. Successfully directing AGI is as much (or more) about WHAT we want to direct it towards (ie “value alignment”) as it is about HOW to get it to do what we ask (ie “technical alignment”).
We believe that AI, or technology, is inherently neutral (but immense) power that can be directed towards either good or bad outcomes - ones that are unfathomably utopian or dystopian for humanity's future. Therefore, it is up to humans to point it in the right direction - AI systems are indifferent to which direction we point them in.
Successfully doing this involves solving TWO separate problems: the technical alignment problem (ie "HOW do we get AGI to do what we instruct it to do," which is up to AI researchers to solve) AND the value alignment problem (ie "WHAT do we want to ask AGI to do," which is a cultural problem). The latter problem is an even bigger problem in our opinion, and this is the central problem that our ideology solves.
3. The best thing to direct AI towards - from today to AGI and beyond - is giving as many individuals the ability to radically level-up and co-evolve with advancements in AI, if they wish to do so.
The extrapolation of this principle over time is not only central to successfully handling the extreme power inequality that is likely to arise between the government and the governed in an AI-enabled post-scarcity society, but it is the also the key to ensuring that humanity survives and thrives beyond the advent of human-level AGI. We see this as the only viable way to eliminate the risk posed by AGI - to co-evolve with it - and we believe this option to do so should at least be extended to as many people as possible reach step of the way.
In more concrete terms, below is what “Superhumanism” prevailing might look like for society in the near-term.
Millions of people being radically empowered via many AI-enabled tools and applications:
- Therapists & life coaches.
- Work & productivity assistants.
- Robotics that can automate all manual functions.
- New drug discovery & cognitive enhancers.
- Many breakthrough inventions.
- And more.
And in the long-term, the further-along implementation of “Superhumanism” might look something like libertarian-transhumanism, whereby all individuals have total control over directing their own evolution at their own pace, are completely free from all limitations and constraints (barring the immutable laws of the physics), and can colonize throughout the universe to create utopias to their own liking, ad infinitum. And all the while, the rights of those who would not like to embark on such a journey will be respected and preserved.
4. Adherents of Superhumanism should do everything they can to push for the success of tenet #3 - not just for the sake of helping it win the ideological war over AI and benefitting humanity as a whole, but because it ultimately comes back around to directly benefit them given its aims.
Superhumanism has positive-sum, win-win dynamics built into it. By pushing for larger adoption of the ideology in the world and recruiting more adherents, members directly increase their own chance of personally benefitting from advances in AI, in addition to benefitting humanity as a whole. This stands in stark contrast to the negative-sum, win-lose dynamics of the status quo, which aims to direct the gains in AI towards the benefit of a few (ie a handful of traditional governments and elites) at the expense of everyone else. This feature of the ideology will be indispensable in helping it gain massive popularity (and ultimately win) as AI advances and the contrast between it and the status quo becomes increasingly starker from the viewpoint of the masses.
5. Adherents of Superhumanism should coordinate to temper the development AI should they deem (per the process outlined below) that it is being developed in an especially unsafe or unfair manner.
If adherents deem (per a specific governance structure outlined in the following section) that tenent #3 is executed in a manner that is grossly reckless or unfair - either because AI is advancing too quickly or because the world's power structures are unable or unwilling to handle this transition in the manner described in tenent #3 - then they must take it upon themselves to coordinate with each other via our ideology's "network state"(aka "The Superhuman Diaspora" - also described in the following section) to see to it that this happens. This is entirely consistent with the aim of directing advancements in AI towards helping level-up as many individuals as possible, but we believe that it so important to emphasize that it is deserving of its own point.
6. Adherents of Superhumanism should also do everything they can to bring people back to life (who wish to do this), in the event that AI becomes advanced enough to achieve this.
This is also consistent with tenet #3 of giving everyone (including both the living and the deceased) the option to co-evolve with AI and direct their own evolution - should this become possible. We should take this possibility serious, especially with cryonics now being available plus the upper bound of AI capabilities being unknown.
Last but not least, while this is not a tenent of Superhumanism, it is well worth emphasizing that the ideology of Superhumanism on its own is not enough. In order to succeed, it needs a container that can enable it to blossom and thrive. That container is Balaji Srinivisan’s network state model. Without it, this ideology will never see the light of day, because it will either become corrupted or destroyed by the prevailing ideologies that have captured today's ecosystems. It is simply not feasible to work within the current ecosystems - an entirely new, independent, and decentralized society + governance structure is needed to give new ideologies like Superhumanism a hope of prevailing, and that is exactly what the network state model allows for.
And so with that, here is how we will actually build that network state...
The Superhuman Diaspora.
THE NETWORK STATE
The Superhuman Diaspora is a global community and network state that is dedicated to "Superhumanism" - a new ideology that emphasizes leveraging gains in AI towards the radical empowerment of as many individuals as possible, rather than governments.
Below are the 4 tenets that underpin why building The Superhuman Disapora is critical for enabling Superhumanism to thrive and successfully directing advanced AGI towards a free and flourishing future for humanity, building on top of the 6 tenets of Superhumanism that were outlined in the previous section:
1. The default forces that will get to decide WHAT to direct AGI towards are malevolent "egregores" - emergent monsters that are comprised of a distinct set of widespread values, beliefs, psychological dispositions, institutions, and incentive structures that all work together in a seemingly coordinated fashion to repeatedly further an agenda throughout society. Some examples of these forces are “Wokeism” and “The Deep State.”
While it can be difficult and somewhat subjective to clearly define the boundaries of an egregore, they are just as real and intentional as an individual in their aims, and a thousand times more powerful. As it stands, there are many villain egregores (such as the ones listed above) that have overtly anti-human and anti-freedom agendas, but there are currently no hero egregores.
2. To defeat the malevolent egregores of today, we need to build a new egregore that can take on the existing ones, stand in stark contrast to their agendas in being unapologetically pro-human and pro-freedom, and offer a “definite optimist” vision for how to direct advanced AI towards a truly free and flourishing future for humanity.
This point is mostly self-explanatory and doesn't need further explanation, but here's a tweet from Elon expressing a similar sentiment.
3. The best model for building this egregore is "the network state."
This is the best model for building the hero egregore, because:
A. It is fully decoupled and independent from the unfixable and cancerous society.
The network state model proposes building the new egregore OUTSIDE OF the confines of ecosystems that have been captured by the villain egregores, rather than from within. If we attempt to build the hero egregore within ecosystems that have already been captured by the villain egregores, then our new egregore will either be corrupted or destroyed. Successfully doing this entails building an entirely new society, infrastructure, and governance structure - all of which is ideally fully decoupled (or as decoupled as possible) from the corrupted ecosystems, which is exactly what the network state model allows for.
B. It's decentralized and pseudonymous nature makes it difficult to attack and stop.
The decentralized nature of the network state model, much like the nature of the villain egregores, make it much harder for it to be attacked and stopped. Having many different settlements all over the world that are a part of the network state in different host countries, doing business deals with many different governments to inoculate ourselves from attacks from various countries, and having global citizens who are either pseudonymous or anonymous - all contribute towards making it extremely difficult for anyone to attack and stop us compared to a typical organization or country. As an aside, we expect that our network state will attract much of the top tech and entrepreneurial talent in the world, which will be our primary asset in striking deals with various countries whereby we receive their military backing in exchange for our tech talent in order to inoculate ourselves from hostile nation-states.
4. Finally, the specific ideology that this network state should embrace, which is central to it defeating the villain egregores and aligning AGI towards a free and flourishing future, is Superhumanism - the idea that gains in AI should be applied to the fullest extent possible towards radically empowering individuals.
In addition to the fact that we believe this ideology is the actual north star for helping humanity flourish in a post-AGI world (ie that we should leverage gains in AI to help us become superhuman and ultimately co-evolve with AGI), its stance of using AI to empower individuals INSTEAD OF governments strategically positions TSD to capture a windfall of new citizens as progress in AI accelerates and people from all over the world increasingly realize that traditional governments are either too incompetent or too totalitarian to be trusted with all-powerful AI. As that happens, people will desperately seek out an alternative governance structure that can both competently adapt to the accelerating change and has a post-AGI vision for the future that is fundamentally pro-human. Our network state will be positioned perfectly to take in and coordinate all of these people in service of our shared aim, enabling us to grow large and powerful enough to supplant the old regime. This is how we will overtake the egregores of today, and will ultimately direct AGI towards a flourishing future.
We believe that all of this can be achieved within a decade if we move quickly, that this is the most important battle that will ever be fought for humanity’s future, and that this is how it will be won!
In terms of thinking about how to actually implement the above (the ideology and the network state model) into a specific governance structure that actually delivers on the mission, we think it is helpful to break this into three parts:
1. What the goals of the network state should be.
2. How to create incentive structures for the network state that successfully drive it and its members towards the goals in #1 on a day-to-day basis.
3. How to create a governance structure that is aligned with #1 and #2.
*Note that points #2 and #3 above do have some overlap, but also have enough differences to warrant being separate points.
Now let's cover each section at slightly zoomed in, but still fairly high-level view of things (Please note that for all of the nitty-gritty details on exactly how all aspects of the incentive structures and governance structures of the network state operate, you can contact us and we'll share this with you):
1. The goals of the network state.
The network state's incentive structures and governance structure need to be constructed in a way where members are strongly incentivized and aligned towards driving progress on the following two fronts:
A. Pushing as hard as possible for the implementation of the tenets of Superhumanism into the development and deployment of advanced AI.
B. Building out and improving the structure, functioning, social cohesion, and effectiveness of the The Superhuman Diaspora machine.
2. The incentive structure of the network state (aka "The Points System").
To properly incentive members of The Superhuman Diaspora to drive towards the goals in #1 on a continuous basis, we propose making a “points system” that is the centerpiece the day-to-day infrastructure and functioning of the network state's community - whereby members award a limited amount of points each week to other members based on who they think made the most progress driving towards the goals in #1. We envision there being a proprietary virtual interface that manages this ( similar to Slack or Discord) whereby members can share, measure, track, and award points for each others' progress towards a hierarchy of the stated goals (and many nested sub-goals) of the network state.
How many points a member accumulates over time determines their:
A. Status and ability to shape the sub-goals/tasks of the network state.
B. Position in-line to receive the benefits of advancements in AI.
C. How much weighting their votes/points have that they award to members.
All of the above incentive structures should serve create a very strong positive reinforcement loop that helps the network state continually refine itself over time to become optimally effective in driving towards its aim, assuming that the mechanism design and initial conditions are correct, of course. Getting this incentive design perfect is required for the network state to work, so we are making a massive investment (both in brainpower and capital) in nailing this.
3. The governance structure of the network state.
Here is a more concrete governance structure, that build on top of the incentive design described in #2 via the the points system:
The central feature of the network state's governance structure is a tiering system, whereby all members are attributed to a tier that is determined by the points system and that dictates how much voting power and influence they have in the network state, as well as their position in line to receive benefits from advancements in AI.
Here is a quick illustration for such a tiering system that can accommodate 10B+ people (since we want to build this with scale in mind from the get-go):
Tier 1: Top 10 members (each can award 100 votes per week).
Tier 2: Top 100 members (each can award 10 votes per week).
Tier 3: Top 1,000 members (each can award 1 vote per week).
Tier 4: Top 10,000 members (each can award .1 vote per week).
Tier 10: Top 10B members (each can award .0000001 vote per week).
So to reiterate, which tier of the network state society you are in is determined by how many points you accumulate, and also dictates what your voting power is and your position in line to receive benefits from AI (since it is impractical for the benefits to be distributed to everyone at the exact same time, so there will need to be some kind of ranking). A key feature of this tiering system and points system is that not all members are equal in their skill, contributions, and value-alignment, and therefore their respective voting powers and benefits should be adjusted to reflect that and in a way that is truly meritocratic. And as a reminder, all of the above activity will take place on a proprietary master interface and dashboard, which all members can view, track, and contribute towards on an ongoing basis.
One other feature of the above governance structure that is worth noting, is that the top two tiers are distinctly different from the lower tiers in the following ways:
The top tier (Tier 1 - Top 10 members) are the official leaders of the network state, must all be full-time, and serve a term of 2 years so that they are able to carry out a multi-year agenda (whereas the other tiers are more fluid/dynamic and are constantly changing). The make major decisions for the network state on a day-to-day basis via a majority vote, and can change anything except the immutable laws of the network state (TBD on what the immutable laws are). They may also elect a president of the network state, if they wish to do so.
The 2nd top tier (Tier 2 - Top 100 members) don't run the day-to-day decision-making of the network state, but must be actively involved (attending leadership meetings at least once per month), and have the distinct privilege of voting on who gets elected into the Tier 1 leadership every 2 years (whereas lower levels do not have this ability). In the elections for the Tier 1 council every 2 years, they are also able to write-in 1 recommendation for a non network state member who they believe should be on the council (such as Elon Musk), which is a mechanism that allows the absolute best leaders in the world to get directly pulled into Tier 1 leadership, if the Tier 2 council demands this through enough votes.
All lower tiers do not have these special privileges that Tiers 1 and 2 have, but can earn their way into the top 2 tiers by accruing enough points, and still receive the benefits of having generally more status, priviliges, and influence in the network state, as well as more voting power in awarding points to members and a better standing to receive upgrades from AI advancements.
Of course, there are many, many more details and specifics that go into making this governance structure actually work than can be covered here, so if you would like these please contact us and we are happy to share them.
In the following section, we outline a high-level plan for how to start actually building the entire network state described above.
1. Phase 1 (2023) - Assemble founders, finalize details, launch.
Recruit first 110 core members (through personal network & referrals) who are extremely mission-aligned, dedicated, and competent.
Host regular summits for founding members and potential recruits.
Hash out, iterate, and finalize the ideology, incentive structure / points system, governance structure, and roadmap with the founding members.
Build proprietary tech platform for points system.
Decide on who is in the inaugural class of the Tier 1 council.
2. Phase 2 (2024) - Expand further and launch co-living districts.
Expand to 1,000+ members via founder referrals and points system.
Initiative a massive pipeline building & recruiting effort predominantly via top 20 college campuses to vet and assemble first several hundred members of the network state.
Host 2-4 summits/conferences for all members (officially launch of TSD, member recruitment/vetting, member projects, etc).
Keep building and improving proprietary tech platforms for not only the points system, but also general communication and organization.
Start co-living spaces and co-living districts for hundreds of members in multiple cities.
3. Phase 3 (2025) - Begin launching self-sufficient settlements.
Expand to 10,000+ members.
Keep expanding on all fronts outlined in Phase 2.
Start crowdfunding and opening permanent settlement around the world that are entirely self-sufficient and can sustain 1,000 - 10,000+ members.
4. Phase 4 (2026+) - Orient AGI towards a superhuman future.
Grow citizenry to hundreds of thousands, and eventually millions.
Launch many more settlements and fully automate their operations to make them as independent and self-sufficient as possible.
Gain recognition as independent nation-state from world superpowers.
Become the world's powerhouse for top technical talent.
Obtain protection from militaries of various world superpowers to inoculate self from attacks, in exchange for access to tech talent.
Become the most influential global actor on the development of AGI, and orient AGI towards the ideals of Superhumanism to help ensure that all of humanity has the most free, pro-human, and flourishing future imaginable.
Please fill out the application form below to be considered for The Superhuman Diaspora.
If your application looks promising, we'll follow up with you!