This project builds upon the successful ACX 2024 proposal to create a minimal viable product (MVP) of a software platform called Spartacus.app designed to solve collective action dilemmas and coordination problems. Tetraspace and I were awarded $17K each to collaborate on it, and we succeeded! The MVP went live in early June. We're currently conducting closed beta testing in search of high-impact use cases.
I'm seeking additional funding to continue developing this project full-time and to extend our runway for another 9-12 months. Funds would be allocated to operational costs, compensating collaborators, and financing growth strategies (described below) to expand the user base, create a sustainable revenue model, develop case studies, and explore non- and for-profit commercialization options.
Unlike crowdfunding platforms, our primary value proposition isn't raising money but facilitating the formation of purpose-driven groups, organizations, and associations through a novel combination of coordination mechanisms. From there, a group can set goals that can only be accomplished through synchronous behavior. Our design principles focus on lowering the risk and raising the expected value of participation in the group formation process, especially in an adversarial environment.
Despite dissatisfaction with a state of affairs, a stable equilibrium can persist if each person in a system perceives the cost of action to outweigh the benefit to themselves. If successful, Spartacus can transform dispersed and/or suppressed preferences into a consolidated public entity capable of pursuing its interests with power and leverage available only through collective action, inverting the cost/benefit calculus. Where the isolated individual has no incentive to act, the individual within a sufficiently coordinated group does.
Our principle innovation combines an assurance contract framework (e.g., person A commits to an action but only if B, C, D…also agree) with temporary anonymity, only revealing identities if and when the campaign reaches its goal. From a risk mitigation standpoint, this allows for a seamless, almost magical transition from the safety of anonymous, provisional commitments to “safety in numbers.” This phased anonymity also directly addresses the inertia from the disproportionate risk burden incurred by first/early movers.
Here’s a sample use case:
Jeremy is a software engineer at a big AI tech company. He’s concerned that the company is developing some potentially dangerous features in its next release. He could face retaliation or even get fired if he speaks up about it. However, he’s pretty sure he’s not the only one who feels this way.
Jeremy can use Spartacus.app to create a plan of action, covertly gather provisional commitments from other coworkers, and, when they have enough support to make a sufficient impact, reveal their identities to follow through on the plan together.
Jeremy can use settings in spartacus to make sure only people who meet certain criteria can sign up for the campaign, such as ID verification, domain exclusions, or unique invite codes.
Title: Speak up against Project Enigma
Description: Colleagues, I'm deeply concerned about potentially dangerous features in our next release. I know I'm not the only one. We can't stay silent. Let's demand a halt and a thorough ethical review. We have the power to shape a responsible future. If you agree, join the Discord group chat "EthicalAINow" to discuss further action. Together, we can make a difference.
Call to Action: Join the discord group chat "EthicalAINow”
(See a demo of this use case HERE)
Collective action problems have three main characteristics:
Individuals would benefit from cooperating around a common goal.
They don’t because of conflicting interests and incentives.
They lack a coordination mechanism to agree on, transition to a different incentive structure, and/or prevent others from benefiting without contributing (“free riding”).
These kinds of problems are particularly frustrating because everyone following their rational interests, the basis of our economic system, is the source of catastrophe.
Some examples:
Climate: Overfishing, Water Scarcity, Deforestation, Littering, Conservation
Public Goods: Research funding, Open Source, Philanthropy
Public Health: Antibiotic Resistance, Biosafety, Organ Donation, Pollution, Phones in Schools
GroupThink: Institutional Capture, Echo Chambers, Peer Pressure
Political Polarization: Misinformation, Voting Rights, Extremism
Abuse of Power: Data Privacy, Corruption, #MeToo, Class Action Litigation, Whistleblowing
Feedback Loops: Arms Control, "Race to the Bottom" competitive dynamics, AI Safety, Panic Hoarding
Solving collective action problems requires stakeholders to override their immediate interests to factor in future and/or collective benefits. This generally entails everyone agreeing to sacrifice for the “greater good," i.e., their future collective well-being. But if only some people changed their behavior and others didn’t, those who did would be comparatively worse off, tolerating “cheaters” would destroy incentives to cooperate, and the dilemma would remain unsolved. Knowing this, a rational person has no reason to adjust their behavior unless there’s a way to coordinate action with other stakeholders and enforce compliance. This shifts the challenge from acknowledging the collective action dilemma to the process of coordination itself.
Coordination is not straightforward. Any attempt is often hindered by the lack of a clear or dominant outcome. There are several impediments to organizing a process in which everyone is incentivized to participate. There may be various ways to achieve a common goal, but the mechanism to ensure they all make a compatible choice may not exist.
At a more fundamental level, there is also the matter of Manifestation - the bootstrapping of change through the power of belief itself. Not in a Woo-Woo sense, but for many cases of human organization around plans, goals, norms, and values, if enough people believe something is possible and act accordingly, that alone is enough to make it possible. It is a self-fulfilling outcome manifested by the decision itself. Inversely, something that hypothetically could happen never will if not enough people believe it can.
Relevance to Effective Altruism
A keyword search for Collective Action Problem(s) and "Coordination Problem(s) returns over 200 results in EU forum posts alone. Some examples:
How can information be gathered to generate the best opportunities for donors and recipients to find each other and meet each others’ expectations?
How can knowledge gathering, dispersal, and specialization all be optimized at the same time?
How can EA harness collective action to affect the behavior of large-scale institutions?
Novel fundraising mechanisms like the one used to fund this series of projects!
Exploring different ways of raising and dispersing funds at scale requires building support for new experiments and norms.
Gain of Function research and AI Safety
How can we enable insiders to mobilize in opposition to institutional incentive structures surrounding endeavors that carry extreme tail risks to humanity?
One can delve deeply into Game Theory to better understand coordination problems and equilibrium states. Spartacus.app is designed to take those insights, instrumentalize them, and productize them as a mechanism of coordination where the need exists but the means are inadequate or nonexistent. For the purposes of this proposal, we can break down some discrete impediments to coordination and respective solutions we’re building into the Spartacus platform:
Challenge 1: Multiple equilibria: Multiple stable outcomes can be achieved.
Challenge 2: No dominant strategy: No single choice is always the best, regardless of what others do.
Challenge 3: Heterogeneity: If individuals have diverse interests or preferences, finding a solution that satisfies everyone may be difficult.
Challenge 4: Asymmetric Information: When individuals have different levels of information, it can lead to misunderstandings, mistrust, and a failure to coordinate efforts.
Solution: Spartacus can be a vehicle for the revealed preferences of large segments of a community, organization, or institution and weaken the forces of preference falsification. This enables stakeholders to better understand what others are thinking, what they want and to calibrate their positions accordingly, which increases the likelihood of convergence.
The default campaign creation process allows for the evaluation of specific proposals. Other campaign types may involve surveying a community directly for various potential solutions to known problems that can be selected by rank-choice. There are options for both campaign creators and participants to designate “tipping points” or action thresholds based on strong signals of other peoples' preferences and risk tolerances. Surfacing a range of alternative equilibria and providing credible indications of what other people want adds meaningful data to the information landscape, leading to more rational and effective decision-making.
Challenge 5: Interdependence: Each individual's outcome depends on others' choices.
Solution: Spartacus makes numerical support for a given proposal visible before identities are revealed, similar to an anonymous poll, but with much stronger commitment signals. The information a Spartacus campaign reveals goes beyond sentiment analysis by binding participants through mutual assurances of action.
Challenge 6: Lack of trust: Individuals may be reluctant to cooperate or contribute if they do not trust each other or the mechanisms involved.
Solution: We use (or plan to implement) several mechanisms to ensure data privacy and security, including enterprise-grade ID and liveness verification, 2-factor authentication, encryption of PII, domain exclusion, financial staking, and unique invite codes. We're aware of the threat of Sybil attacks, “honeypot” entrapment, brigading, and other malicious tactics bad actors might employ to sabotage campaigns or harm users. We will utilize all feasible best practices to mitigate these threats.
Challenge 7: Uncertainty: If individuals are uncertain about the benefits of cooperation or the costs of not cooperating, they may be less likely to participate.
Challenge 8: Transaction Costs: The costs associated with organizing and coordinating collective action can be significant. These costs may include time, money, and effort. Individuals may be less likely to participate if the costs outweigh the benefits.
Solution: The assurance contract framework addresses uncertainty by holding all commitments provisional until a high-impact threshold is reached. This dramatically raises the expected value of participation and reduces the risk that efforts will go to waste or fail to achieve viability. This general principle has been well validated by the success of companies like Groupon, Kickstarter, and dozens of other crowdfunding platforms. We can further reduce uncertainty by adding a financial staking component to filter for even stronger commitment signals and allowing each participant to choose a personalized participation threshold.
Challenge 9 Free-riding: Even if individuals understand the benefits of cooperation, they may still be tempted to free-ride by benefiting from the efforts of others without contributing themselves.
Solution: One of the side effects of transitioning from anonymity to real identities is the introduction of reputation management dynamics, which add social pressure to keeping one's commitments. The structure of Spartacus campaigns ought to filter out people with low follow-through intent. We are not seeking to replicate the lowest friction methods of registering the most superficial indication of preference and sentiment, as many petition platforms do. While we cannot coercively enforce follow-through and expect some amount of leakage, we’ll experiment with gamified accountability mechanisms and novel "skin in the game" elements that preemptively select against those most likely to flake and deter unserious participation.
Challenge 10: Inefficient institutions: Weak or ineffective institutions may be unable to provide the necessary incentives or coordination to overcome collective action problems.
Challenge 11: Path dependence: Past choices or historical factors may create barriers to cooperation or limit the effectiveness of solutions.
Solution: The immense value of creating brand-new associations and organizations that could be viable alternatives to a dysfunctional, unresponsive status quo is precisely what Spartacus is designed to do.
Challenge 12: Time horizon mismatch: Individuals may prioritize short-term gains over long-term benefits, leading to suboptimal outcomes for the group as a whole.
Solution: If people weighted their short-term and long-term interests equally, many collective action dilemmas would disappear. Studies and observations show that people generally have high time preferences, meaning they value immediate gratification and prioritize present needs over future needs. Spartacus can create a playing field where long-term interests, which many acknowledge but cannot effectuate, find a venue to compete with default behavioral patterns.
Challenge 13: Intimidation / Retaliation: In some cases, individuals may be afraid to participate in collective action because they fear retaliation or sanction.
Challenge 14: Lack of Leadership: Strong and effective leadership can be crucial for overcoming collective action problems. Without good leadership, individuals may be unable to coordinate their efforts or achieve a common goal.
Solution: Using phased anonymity to protect first/early movers and delaying public exposure unless and until a campaign success threshold is crossed addresses critical reasons why people are reluctant to initiate action or voice dissent, especially if they’re unsure how much support there is for their position. Few people are comfortable being the lone voice of dissent.
Risk aversion can be modeled on a curve. If the risk of an action is very high, only outliers with extraordinarily high tolerance for risk will act. The problem with waiting for heroes is that they rarely appear; if they do, they’re often not the people best equipped to deliver productive change. Unfortunately, there’s scant evidence that bravery and competence are correlated. On the contrary, reckless disregard for personal risk is associated with incompetence, irrationality, and extremism.
People in the normal distribution of risk aversion - most reasonable people - become paralyzed if they do not think the benefits of action outweigh the risks. They will certainly be disinclined to test that assumption without assurances. They will acquiesce to dysfunction, corruption, abuse, insanity, and even institutional collapse, all the while justifying their inaction in perfectly rational terms. The parable of the naked emperor comes to mind.
Empowering those of average bravery and risk aversion to take action by decreasing the level of risk below their action threshold would open the door to many more people willing to act on matters of collective importance.
As mentioned in the intro, after about six months of spec and design work, we debuted the MVP in early June at NYCTechWeek. It's a web app built on Next.js, with APIs to services like Supabase, Twilio, Plaid, and Auth.o.
Right now, anyone can sign up and test the core functionality of real-time and private campaigns. Feel free to check it out. (We recommend the mobile interface and using a phone# to sign in). Create, Join, and Share user flows, and some basic user settings are all working.
You can find more details on our progress journey in previous monthly posts to our substack.
We're continuing to focus on user acquisition, use case discovery, testing, identifying and publishing success stories and case studies, and using feedback to determine how to shape our medium- to long-term product roadmap.
Extending our runway for another 9-12 months.
Here's a rough breakdown of the annual project OpEx at current levels of spending:
Continued compensation for co-grantee TetraSpace at $50/hr for 10-20 hrs of work per week (variable).
Software & subscriptions: $5,000 | Essential dev tools, service providers, APIs, CRM, lead generation tools, user support, and product management software. (may fluctuate based on metered services).
Hosting & infrastructure: $3,000 | Cloud hosting, website, and basic IT infrastructure.
Marketing: $5,000 | Social media, content marketing, and SEO.
Public relations & events: $3,000 | conferences, exhibitions, presentations, demos, and speaking opportunities.
Legal and accounting: $3,000 | Basic legal services and accounting for tax compliance and financial management.
Outsourcing, consulting, and interns: $6,000 | Paid internships, compensation for consultants and subject matter experts, guest speaker fees.
I'm covering my own expenses with personal savings while working on the project full-time.
I’m Jordan Braunstein (He/Him). I've been an avid reader of ACX/SSC for eight years and actively participate in the LA rationalist community. I'm also connected to the EA community via EA Los Angeles and Lighthaven in Berkeley. I can be found on LinkedIn and Twitter.
I have 15+ years of experience as a business development lead and product manager at various startups, ranging from pre-seed to Series B, including employee #1 at a VR startup called Vivid Vision. I have a B.S. in Economics and Political Science from SUNY Purchase, NY. I've lived in NYC, Chicago, San Francisco, and now Santa Monica, LA.
My primary technical collaborator is Tetraspace (She/Her). We independently submitted similar applications to the ACX/Manifund grant program and were told we could receive a more significant joint grant if we agreed to work together. I had a robust framework for the project, and Tetra had the technical skills I lacked. As her contribution, Tetra joined the Spartacus.app project as a freelance engineer, using her half of the grant as compensation.
Market-Related:
We vastly overestimate demand for a platform like this.
Users don't see enough value in the platform to justify paying for it.
We fail to market to the most receptive audiences because of incorrect assumptions about our ideal use cases.
A competitor sees what we're doing, copies our functionality, and poaches our users.
Association with controversial subjects causes the project to become polarized and anathematized by a large segment of potential users.
Product-Related:
The platform is too complex, complicated to use, or unreliable.
We can't scale the platform effectively to meet growing demand.
Making collective coordination easier and less risky has a double edge. We don't sufficiently limit how the platform can be used, allowing activities that cause real-world harm.
Financial:
We run out of money before achieving profitability or securing additional funding.
We vastly underestimate the costs of user acquisition and retention.
We can't find a viable revenue model and/or business model.
Team-Related:
Conflicts and disagreements within the team.
Key team members leave the project.
We don't have the means to recruit necessary additions to the team.
If we ever reach a point of irreversible failure or resource exhaustion, hopefully, we can salvage the valuable assets and find a new home for them.
This project received $34,000 in February 2024 from the ACX 2024 grant series, split between two projects (which were functionally combined).
If we continue progressing, our tentative plan is to transition into a public benefit corporation. This will allow us to grow faster and diversify our funding pool. If we can't source sufficient philanthropic funding, we'd be very open to securing private funding from value-aligned angel investors contingent on validating a viable business model.
App status across various funders