Project summary

The value of our long-term future depends largely on two factors:

1. Whether humanity survives the AI transition

2. How good the long-term future is given survival

While there are ~1,000 people working on AI safety, there are closer to ~10-50 people trying to make sure the long-term future is good. I want to change that.

I plan to design and launch a flourishing futures fellowship, similar to BlueDot's AI safety fellowship but focused on making the long-term future as good as it can be.

I believe this is important for reasons similar to those given by William MacAskill in his "Better Futures" series; essentially, getting a really good future requires getting a lot of factors right, a lot of these factors are highly uncertain, and we really don't even know what a "truly good future" fundamentally means. Most importantly, advanced AI will accelerate multiple forces that could push toward path dependence and lock-in not long after AI becomes dangerous (e.g. AGI may be structurally power-concentrating and deep space may be defense-dominant.)

AGI is almost here and we have no idea what a post-AGI future should look like; I want to build a field of people thinking about this and developing high impact interventions to make flourishing futures more likely starting immediately.

This fellowship will develop researchers with a cutting-edge understanding of the challenges involved in achieving flourishing futures, who can then go onto work on researching effective interventions for achieving flourishing futures.

This feeds into my next project, developing a Charity Entrepreneurship-like intervention pipeline that repeatedly generates hundreds of interventions within a flourishing futures focus area (such as AI for epistemics, post-AGI governance, or space governance), systematically evaluates them to surface the top ~5 interventions, and develops comprehensive 2-year charity launch plans which are then incubated by entrepreneurs selected through a highly selective process (Charity Entrepreneurship accepts about 1% of applicants as charity founders).

Note: this is funding only for the fellowship. I mention the intervention pipeline/incubator because an important benefit of the fellowship is developing research talent for the intervention research pipeline.

More detail on this second step here if desired: https://docs.google.com/document/d/1ms4tiEXJHr3ivV4asHK8qeF5LwxltJoUU2NBaQFa4bo/edit?usp=drivesdk

Project goals and theory of change

Right now, two main bottlenecks for ensuring a good long-term future are 1. the scarcity of well-trained talent within the flourishing futures field and 2. the lack of research on tractable interventions that could meaningfully shift trajectories toward better futures.

The main goal of this project is to develop a repeatable, scalable fellowship which gives fellows a deep understanding of flourishing futures theory and strategy (Phase I, weeks 1-6) plus either a broad survey of interventions or hands-on intervention research within the research focus track (Phase II, weeks 7-12).

I have already developed an initial version of the curriculum. The broad outline is as follows:

1. Phase I (weeks 1–6) builds shared strategic foundations: Better Futures and why eutopia is a narrow target; moral uncertainty & cluelessness; convergence & moral trade; lock-in & path dependence; post-AGI existential risk; and strategic targets like post-AGI equilibria & viatopia.

2. Phase II (weeks 7–12) focuses on interventions: fellows either survey all six intervention focus areas or do an independent research project in one.  Focus areas include AI for epistemics/coordination/macrostategy, extreme power concentration & coups, space governance, artificial sentience & AI rights, post-AGI governance, and flourishing futures field-building.

I will consult with top experts on the readings for each week to ensure I'm getting the cutting-edge material for each topic. I've received feedback from leading researchers in several of the above fields and found their insights extremely valuable for my own research. This will help fellows develop an up-to-date understanding and prepare them to meaningfully contribute at the bleeding edge of the field.

Fellows who graduate from the program will have a deep understanding of flourishing futures and be placed to either 1. continue flourishing futures research, independently or at an existing organization, 2. turn their intervention into a project or organization, or 3. contribute to further flourishing futures field-building work. This directly helps us get more research on high-leverage interventions, start moving toward launching the best of these interventions, and continue scaling the field.

How will this funding be used?

Minimum: $10,000

Personal runway: I am in Berkeley but extremely frugal; if needed this will allow me to cover my living expenses for at least the 7 months I estimate it will take me to develop the curriculum, launch five or so (simultaneous) 12 week initial cohorts, and prepare for scaling the fellowship: $10,000

Ideal: $25,000

Buying acceleration: I would like to pause my small business and pay for basic meal prep (I have a health condition which requires a special diet, and I currently prep all my own food.) By focusing entirely on the fellowship, I could reduce expected project time from 7 months to 5 months, helping this tiny field grow significantly faster: $8,000

Important tools and subscriptions: Zoom; Airtable; website hosting/domain; AI compute for curriculum research/website development/various ops work: $1,000

Website development and operations support: I can do, but not my comparative advantage, contracting out would allow me to focus on what's important. $2,000

Targeted fellowship recruitment: e.g. travel costs for an extra 1-2 EAG conferences, promoted posts; important as fellow quality & fit are a key determinant of success: $2,000

~10% buffer: Unexpected mishaps or opportunities, things I'm forgetting. Possible examples: stipend for an extra facilitator if I have an overwhelming number of good fellowship applications, bounties for miscellaneous project tasks, legal & incorporation costs if things go well and I need to transition to a non-profit, etc.: $2,000

Who is on your team? What's your track record on similar projects?

Curriculum development and fellowship facilitator: Myself, Jordan Arel.

I independently became a die-hard committed longtermist in childhood, wrote a book draft in 2021 with the intention of launching a longtermist/x-risk/flourishing futures movement shortly before discovering the EA/longtermist/AI safety movement already existed, and spent most of last year researching flourishing futures, with relatively positive feedback from field-leading experts on my research. When performing this research (a 35,000 word essay) I developed very in-depth models about what flourishing futures involves, and to a large degree already know what will go into the fellowship.

I’ve won prizes in Superlinear’s “Most Important Century” writing competition, AI Impacts/Cotton-Barratt’s “Automation of Wisdom and Phiolosphy” contest, and been featured in Foresight’s RFPs.

As a community builder, I hosted several EA and longtermist fellowships online through CEA and in-person at UC Berkeley, which helped recruit EA UC Berkeley's previous president and at least two AI safety researchers, one now working at AISI. I also trained other facilitators for CEA as a "meta-moderator." I achieved CEA's "highly rated facilitator" distinction based on participant ratings.

Ultimately, I believe I am a leading candidate for developing this fellowship; there are few with deep expertise on this topic, and almost no one focused on scalable flourishing futures field-building. I already have one offer from Sentient Futures to host the fellowship after I design it (note: I have not accepted this offer yet, but find it highly encouraging that I've received an unsolicited offer.) 

Other data-points:

I scored 790 on the GMAT (math, verbal, & reasoning skills) on one of my three practice tests, the 99.98th percentile among MBA applicants.  

In high-school, I won a math competition spanning seven states.

What are the most likely causes and outcomes if this project fails?

The project is very doable, but if not enough participants apply, it could fail to serve the targeted number of cohorts. It will be important to double-check best practices on fellowship recruitment and leverage my network to find interested participants. An analysis from my research seems to show there is a lot of interest in flourishing futures among the EA community, but nonetheless it will worth putting out an expression of interest form to ensure demand.

How much money have you raised in the last 12 months, and from where?

I have been self-funding my research for the past year and a half, and have done a lot of self-funding in many previous years. This project would be the first time that I receive funding for independent charitable work. It would be highly transformative, as I suffer from a disability that makes my life quite difficult.