Comments
(outdated version) Shortlist of Longtermist Interventions
Note (added after original post): This essay was ghostwritten by AI and as such has a few significant, sometimes subtle, mistakes. An updated final version can be found here.
Note: after posting a temporary version for the "Essays on Longtermism" competition deadline, this series of posts has been temporarily removed from the front page, and will be reposted in the following days after updates have been made.
Epistemic status: In the spirit of draft amnesty, I am posting this series slightly before it is fully ready or in ideal form.
This represents many years' worth of my thinking and I think the core material here is quite important, but I really wanted to submit them for the "Essays on Longtermism" competition, due today which ended up competing with some other important fellowships and other applications, hence getting far lower priority and attention than it deserved.
Nonetheless, I believe these ideas are fundamentally important, execution may just be closer to middle draft than final draft.
That said, I will likely be updating this series significantly in the following days, especially the last post, “Shortlist of Longtermist Interventions,” and the last two pieces which are unpublished. I will have a running list of when updates have been made, so that readers can easily see when each essay has been updated and finalized, available on the series explainer page here:
(Brief and Comprehensive Series Explainers Here)
This essay, easily readable as a stand-alone piece, is this first essay and part of my submission to the "Essays on Longtermism" competition, based on my unpublished work on Deep Reflection (comprehensive examination of crucial considerations to determine the best achievable future). (Deep Reflection Summary Available Here)
TL;DR
This essay presents a shortlist of concrete high-leverage Better Futures interventions for moving toward viatopia, based on the key principles established in Introduction to Building Cooperative Viatopia and Viatopia and Buy-In.
These principles prioritize interventions that: improve infrastructure for the longtermist community, focus on values and human psychology/flourishing, improve exponentially through compounding feedback loops, enhance strategy, generate more resources and interventions, create ongoing self-improving institutions, and leverage AI in ways that become more effective as compute and capabilities grow.
The interventions span multiple categories: fellowship programs and incubators (like a Charity Entrepreneurship-style Better Futures fellowship), research automation tools (workflows that help researchers be exponentially more productive), coordination platforms (novel systematic collaboration mechanisms for sharing best practices and ideas), value reflection infrastructure (institutions and technologies for systematic moral progress), field-building initiatives (creating the ecosystem needed for viatopia work to flourish), and AI tools designed to compound in effectiveness over time.
This essay addresses the gap between longtermist theory (which the EA community has developed extensively) and practical implementation infrastructure (which remains severely underinvested).
Each intervention includes brief rationale for why it's high-leverage, demonstrating the breadth of concrete work we could be doing right now. The list shows that moving from philosophical arguments about viatopia to actual implementation is achievable via building extensive practical infrastructure.
Interventions
I am generally posting these interventions in priority order. However, I will first start with this list I created a few months ago, which serves as a good brief orienting overall prioritized list. I originally created this list of interventions for Deep Reflection, but it mostly equally applies to Viatopia and serves as a great starting point for highly effective interventions. I will include research as an “intervention” when research is necessary to have the right strategic picture to take effective action. You may notice these are mainly "meta" interventions, and that is because I believe these are especially high impact early on when building a field such as Better Futures/Viatopia/Deep Reflection. For a great list of object level interventions, see MacAskill’s essay on this topic:
(I've included one page on this idea here because it is quite important to this series, even though I'm not sure it is actually the number two most important idea here. Unfortunately I ran out of time so am just including one page on it instead of the twelve pages I had planned. The same will be true of the Children's Movement. I should hopefully, be able to post the full versions in the next few days.)
The Hybrid Market represents one approach to a core challenge in creating better futures: how to aggregate diverse values across society while systematically incentivizing value-creating behavior. As viatopian infrastructure, it addresses the institutional design problem of steering toward better outcomes without requiring universal longtermist adoption or centralized control.
Core Mechanism
Unlike traditional markets measuring only financial performance, the Hybrid Market has traditional financial transaction like but simultaneously prices all externalities—positive and negative—across all timeframes, using Advanced AI (notably, AGI or at least powerful forecasting AI is likely necessary for easy implementation) to model flow-through effects and integrate with the Internet of Things to act as an intelligent oracle. Organizations receive currency ("happies") proportional to total value created, pegged to and individual’s subjective improvements in well-being over a given period of time, allowing them to explicitly make decisions on how they spend money based on how much value they get. Yet, value is measured not only in terms of happiness but in in terms of all important public goods, near-term and long-term, including: reducing existential risk, improving wellbeing, accelerating beneficial research, or any measurable contribution to long-term flourishing. They are penalized for harm caused.
Critically, the currency flows to those actively creating value rather than accumulating with asset holders. Because assets must be continually "rented" (similar to Harberger taxes) and the overall hybrid market return rate is 0% on average, wealth naturally redistributes to value creators. This creates automatic incentives: generating public goods yields currency, while hoarding assets costs currency. The system thus systematically channels resources toward interventions that improve society's long-term trajectory.
Why This Increases Long-Term Value
The mechanism creates a clear causal chain: First, by rewarding all positive externalities (not just profitable ones), it dramatically increases production of public goods—particularly those with large long-term benefits but weak near-term market incentives, like existential risk research, institutional improvement, and moral progress initiatives. Second, this systematic increase in public goods provision makes society function substantially better across multiple dimensions: better epistemics, improved coordination, stronger institutions, and more resources allocated to genuinely important problems. Third, a better-functioning society with robust institutions and wise resource allocation is far more likely to navigate future challenges successfully and realize significant long-term value.
One way to think of this is as a systematized, marketized, universal, effective altruism mechanism which applies to all financial transactions and investment.
Importantly, the system will develop specific proxies for longtermist goods. Just as carbon markets created standardized metrics for emissions, the Hybrid Market would establish proxies and indices for x-risk reduction, institutional quality, better futures trajectories, and other far-future-relevant factors. This makes previously invisible long-term considerations economically legible.
Making Trade-offs Explicit
Perhaps most crucially, the Hybrid Market renders trade-offs transparent. When a harmful product's true social cost appears in its price, individuals directly observe that they're trading potentially vast numbers of future lives for trivial present consumption. This visibility itself shifts behavior by making consequences salient in ways current markets systematically obscure.
Implementation and AI Enhancement
The system functions as decentralized mechanism design—implementing Cotton-Barratt and Hadshar's insight about market-based provision with taxes and subsidies, but without requiring state coordination. AI serves two critical roles: futarchy-style prediction of which interventions will succeed, and modeling complex flow-through effects computationally intractable for humans. However, humans retain control over values—deciding which indices matter most through investment choices.
As AI capabilities improve, measurement accuracy and predictive power scale, making the system increasingly effective at channeling resources toward genuinely valuable interventions. This represents infrastructure that compounds in effectiveness as technology advances—precisely the kind of intervention longtermism should prioritize.
The Children's Movement represents a human-centric approach to viatopia, targeting humanity at arguably its highest point of leverage: early childhood development, when epistemics, values, and collaborative capacities first form. Originally conceived in 2021 as part of my pre-EA work, this intervention maintains human agency while systematically improving humanity's capacity for wisdom and long-term thinking.
The Leverage Argument
Several considerations suggest childhood intervention may be exceptionally high-leverage for longtermist goals. First, effects compound over entire lifetimes and across generations—a single cohort of better-raised children influences society for 70+ years and shapes how the next generation is raised. Second, empirical evidence is encouraging: studies show $4-12 return per dollar invested in quality early childhood programs, measured through later achievement and prevention of negative outcomes. While these studies are somewhat crude and focus on near-term metrics, they suggest how a basic compounding mechanism may be at play. Third, and perhaps most importantly, this intervention addresses the "human substrate" that all other interventions depend on—creating populations with stronger epistemics, better collaborative capacity, and more careful reasoning about values.
Cotton-Barratt and Hadshar emphasize "educating people so they are well equipped to tackle challenging research work" and developing "psychological health and ability to productively work over long periods." Childhood intervention directly addresses both challenges while also tackling the political feasibility problem: creating citizens who genuinely support longtermist priorities, avoiding any need for coercion.
Why This Improves Long-Term Value
The causal pathway is straightforward: By systematically building strong epistemics, collaborative mindsets, moral reflection capacity, and psychological health from earliest ages, we create future generations naturally inclined toward the careful reasoning about values and far futures that longtermism requires. These adults will be better equipped to navigate strategic cluelessness, avoid premature value lock-in, solve coordination challenges, and generally make wiser decisions about humanity's trajectory.
Childhood represents perhaps our most crucial leverage point for systematic value improvement. Following MacAskill's framework on moral uncertainty, we need institutions that help humans reflect on values (rather than optimizing for a single fixed conception of value), experiment with different value frameworks, engage in substantive debate, and systematically move toward better values over time. Childhood is when value formation is most plastic.
This aligns with "keep the future human" frameworks that prefer allowing human values to evolve deliberately rather than immediately optimizing via AI. If we want humans to remain the primary factor determining values into the future—at least until we're confident about AI's role—then improving humans' capacity for wisdom becomes essential.
Evidence-Based Implementation
The following interventions represent a starting framework based on both empirical evidence and first-principles reasoning. These specific proposals may be controversial and should be viewed as proof-of-concept rather than final recommendations. Implementation should prioritize the most effective, high-leverage practices as determined by rigorous evaluation, and incorporate AI tools where they prove as or more effective than human alternatives.
The intervention builds on the existing Convention on the Rights of the Child while extending it significantly, adding positive rights (what children should receive) alongside negative rights (what shouldn't be done to them):
Core Intervention Categories:
This represents institutional design from first principles, demonstrating how systematic attention to human development could serve as powerful viatopian infrastructure for improving humanity's capacity to navigate the challenges ahead.
Charity Entrepreneurship style intervention research pipeline for Better Futures, especially Viatopia/Deep Reflection:
I believe most expected value lies in Better Futures, and information on high-impact interventions is especially valuable. In particular, Viatopia interventions, which ultimately lead to comprehensive reflection (Deep Reflection) in which we’ve analyzed all crucial considerations, & equally importantly, a state of humanity in which we are likely to take action to ~maximize value based on these considerations via “existential compromise.”
The research pipeline would generate 200-300 interventions across Better Futures/Viatopia/Deep Reflection intervention focus areas and then do progressively deeper dives on them to find which is likely to be most effective and robust against downside & failure.
Charity Entrepreneurship style Better Futures Fellowship-Incubator 1-Page Summary:
Following my personal experience developing and enacting the Charity Entrepreneurship style Better Futures research pipeline, I could leverage my facilitation experience leading fellowships and facilitator trainings on EA/longtermism/x-risk for CEA and at UC Berkeley to launch a fellowship with 16-week cohorts.
These cohorts would participate in reading/discussion groups on a core curriculum encompassing
Fellows would then each generate and progressively evaluate 100-200 interventions in whichever Better Futures focus areas they have personal interest or expertise in, with emphasis on interventions most likely to create a state of Viatopia, eventually leading to comprehensive reflection.
This fellowship has the advantage that it upskills fellows who are then in a much better position to, depending on their comparative advantage:
This structure would scalably increase the amount of talent focused on high-impact Better Futures research, while simultaneously increasing the tractability of the field by developing highly effective interventions.
This takes advantage of the massive gap between field interest and a seeming paucity of specialized talent, organizations, and interventions in the field, as evidenced by my analysis of the EA forum "existential choices debate,” showing a slight preference for work on improving the quality futures in which we survive over increasing chances of survival (n=366) (link)
Charity Entrepreneurship's Viatopia/Deep Reflection adaptation.
(very rough draft)
Stage 0 — Process Design (30 hours): Pre-register decision criteria, establish a "bar to beat" (interventions must demonstrate X% estimated increase in likelihood of comprehensive reflection, without increasing x-risk), define evaluation weights, and specify kill-criteria.
Stage 1 — Mass Hypothesis Generation (200-300 ideas, 50 hours): Generate interventions across categories using literature sweeps, brainstorming, and expert consultation. Categories include AI tools, organizations, new institutions, policy advocacy, social mechanisms, governance mechanisms, and field-building. Define major theories of change, such as creating self-reinforcing cycles toward comprehensive reflection.
Stage 2 — Quick Prioritization (~50 ideas, 120-150 hours): Apply four rapid evaluation methods: quick cost-effectiveness estimate, initial weighted factor scoring, evidence quality scan, and informed consideration snapshot.
Stage 4 — Expert Review & Critical Uncertainties (~10 ideas, 100 hours): Interview 1+ experts per intervention. Identify 1-6 "killer" uncertainties and spend 90 minutes each to address them. Red-team assumptions.
Stage 5 — Deep Dives (~5 ideas, 250 hours): Produce comprehensive reports including: detailed theory of change, stakeholder power/interest matrices, scalability assessments, externality analysis, crucial considerations, and expert review.
Stage 6 — Implementation Blueprints (2-4 finalists, 5-30 hours): Create full "charity blueprints"—detailed 2-year launch plans with monthly milestones, team composition, funding roadmaps distinguishing EA vs. mainstream sources, measurable success proxy metrics for increased reflection likelihood, and limiting factor mitigation strategies.
Supporting Infrastructure: I would develop AI research tools to accelerate this work, based on my award-winning work on the automation of research. Tools would be made publicly available for other high-impact researchers.
Documentation and Scale: Every process would be documented to facilitate future work, including potential growth into a full CE-style Better Futures incubator, sustainably improving the field’s neglectedness and tractability.
Personal fit: My entrepreneurial experience, facilitation background (CEA), and research foundation position me uniquely to bridge theory and practice.
Some miscellaneous AI stuff which I need to edit a lot more and improve on this doc. I've put pretty low effort into this so far: (too rough, unposted)
5.
Have a community Dropbox where people can put their best ideas for creating high-leverage infrastructure or systematically improving the longtermist community. Then, people vote on them, at regular intervals or ongoing, and the very best ones will have a prize competition created around them for both designing the idea and then for building the idea.
Also a norm that every day people should take 30 minutes to write down the very best idea they have, either in general or specifically on longtermism. Could be ideas for action, like interventions, or strategy insights, or ideas for self-improvement for longtermists, or community ideas, or important personal practices and values, and anything else that seems high leverage, or achieving longtermist goals.
They should write this up as a 1-5 minute post., Can just speak in natural language about the idea to AI for 15 minutes, and then have AI write it out. Give AI feedback until it gets it right.
Could write out a workflow, which it becomes a norm that everybody in the community just does takes half an hour every day to do this and write out their very best idea that they can think of, which they haven't written before or even have written before but have new insight on.
Then everyone publishes their thing in the same place. There could be a bespoke forum just for this purpose, and then everyone votes on each other's ideas. Could have tagging for categories, but also have qualitative voting where you vote on how good an idea is across various important dimensions, like:
• Strategic Intelligence
• Easy to put into action
• Opportunity cost; how much displacement of other resources required to put into action.
• High leverage
• Scalability
• Novelty
• Robust against failure
• Robust against downside.
Etc.
People could vote on any of these qualities or use come up with their own to rate ideas, as well as just how good the idea is overall.
This can help spread best practices of the most effective longtermists and raise awareness of the most important ideas.
Ideas are very digestible because they are so short. And so, many ideas can be evaluated quickly.
