Applications for the first round open today! To be eligible for this grant round applications have to come in before August 23rd, and we will announce grant recommendations by October 23rd. From then on we aim for one funding round every 3 months with rolling applications (so the same application still works if you apply after the 23rd of August, you’ll just get a response 3 months later).

For the first round, I am expecting participating funders to disburse around $15M–25M. Hopefully similar amounts every 3 months (and growing over time).

Funders can withdraw and join the round at any time, so this is in-flux, but our current tentative list of funding for the first round is:

• $10M from Jaan Tallinn (conditional on us getting $10M from other funders)

• ~$5M from Dustin Moskovitz for the first round (and $10M over the first year if the first round goes well)

• $3M+ from an anonymous donor

• $2M+ from a different anonymous donor

• ~$2M from the Long Term Future Fund

• ~$2M from the ARM Fund

• ~$200k from Timothy Telleen-Lawton

• ~$100k from Andreas Stuhlmüller

Please book an onboarding call with me right here. While we place no strict requirements on who can access the platform, I want to make sure we are on the same page about treating applicant information as appropriately sensitive, and that you assign some substantial probability to distributing at least $50,000 in the coming round.

Evaluators receive 2% of any funding directed through them, and we additionally charge a 3% platform fee (for a total of 5% on top of the grant amounts). We lower platform and evaluator fees for funders who distribute large amounts of money through the platform.

The funding process works on two simple principles:

1. You fund what you value the most

2. When several people want to fund the same thing, they split its cost in proportion to how much money they are giving away

There are two main classes of participants in the Lightcone Commons: funders and evaluators.

All funders and evaluators enter a funding order, i.e. the ideal order in which they would distribute money in the round. For instance:

“Project Alpha is most effective, so I want to fund that until marginal money given to it is as good as funding project Beta, at which point I want to fund them both until they are as good as project Gamma, etc. until eventually I run out of money”

Evaluators do so for projects, and funders have the extra option of doing the same for evaluators. For instance:

“Alice seems to have the best judgement, so I want to fund her until her recommendations are only as good as evaluator Bob’s recommendations, at which point I want to fund them both until their recommendations are only as good as Carol’s, etc.”

When two funders want to fund the same evaluator, they split the bill in proportion to their total budgets. When two evaluators/funders want to fund the same organization, they too split the bill in proportion to their total budgets. If one funder wants to fund a project for up to $100k, and the other one wants to fund it to $150k, then they split the bill on the first $100k, and the second funder who is more excited about the project and wants to fund it more, covers the remaining $50k.

A virtue of this system is that it tells us immediately how to reallocate funds if anyone’s funding order changes, or a funder’s total budget changes, while following a simple rule for ensuring that costs get split fairly between participating funders.

If you want to build a more intuitive understanding, you can play around with the widget at the top of this page.

The key thing I recommend funders spend their time on is to figure out which evaluators they trust to produce good grant recommendations, and when the recommendations by an evaluator hit diminishing marginal returns. If you brought your own evaluator to a grant round, then you could make a final funding decision in as little as 2-3 hours of getting yourself familiarized with the process, the UI, plus your onboarding call.

If you do not have a strong pre-established trust relationship with any evaluators in the round, I recommend that you budget on the order of 10 hours to read evaluations and meeting notes or watch recordings of debates and discussions between the evaluators, and to read other content written by the evaluators that helps you understand their worldviews.

You can also participate by funding organizations directly, without deferring to any evaluators. In that case, I recommend more like 50 hours to have enough time to read through applications, ask organizations questions, and make a proper grantmaking portfolio.

But ultimately, if you end up finding that you don’t have the time to come to a sufficiently robust assessment of any evaluator or organization to make you think giving to them is a good idea, you can just decide to reduce your giving. You are not committing to any time investment by signing up as a funder.

We actively recruit people with expertise and competence to evaluate grants on our platform. Some evaluators currently confirmed are:

• Zvi Mowshowitz, author of widely read AI-analysis Substack "Don't Worry About the Vase"

• Yafah Edelman, chief strategy officer at Epoch AI

• Eliezer Yudkowsky, co-founder of MIRI and early AI alignment researcher

• Nate Soares, president of the Machine Intelligence Research Institute

• Caleb Parikh, managing director of the AI Risk Mitigation Fund and head of the Long Term Future Fund

• Elizabeth Van Nostrand, widely read author of AcesoUnderGlass.com

• Me! (Oliver Habryka, CEO of Lightcone Infrastructure)

• Katja Grace, lead researcher at AI Impacts, is confirmed to join for future rounds

And I am actively working on getting more evaluators on the platform, with examples of who I would love to have on board being:

• Scott Alexander, widely known author of SlateStarCodex and AstralCodexTen

• Ryan Greenblatt, Chief Scientist at Redwood Research

• Ajeya Cotra, former Program Director in AI Safety at Coefficient Giving (formerly known as Open Philanthropy)

• Ben Hoskin, managing director of VARA, a highly successful hedge-fund, and author of a number of very high quality historical AI charity reviews

• Eric Neyman, former researcher at ARC, now focused on identifying high-leverage giving opportunities in policy and advocacy

• Richard Ngo, former policy researcher at OpenAI, and widely read author on AI safety and AI policy

• Kelsey Piper, staff writer at The Argument and author of many great essays on altruism, ambition, and philanthropy

• Daniel Kokotajlo, head of AI Futures Project, author of AI 2027 and ex-OpenAI governance researcher

• Gavin Leech, co-founder of Arb and co-author of The Scaling Era with Dwarkesh Patel and author of 1,188 book reviews

• Rohin Shah, lead of the AGI Safety & Alignment team at Google Deepmind

(Just to avoid any ambiguity, the people in this second list have not agreed to participate as evaluators! They have not endorsed this process! But it seemed useful to give a realistic sampling of people I am aiming to get on board.)

No. We do not require applicants to have 501(c)3 status, and coordinate grants and investments to for-profits, individuals, educational institutions, 501(c)4 organizations, non-US organizations and individuals. Funders can specify arbitrary restrictions on what kind of organizations or individuals they want to fund, depending on what makes sense from their ethical, legal or tax perspective.

No. Any evaluator can recommend grants to any organization (or private individual) that they think is a good funding target. We only ask the evaluator to fill out basic information about the organization, such as what kind of legal entity they are, and to perform some basic due diligence to avoid fraud.

We have also imported recent live applications from the Long Term Future Fund, Manifund.org and grantmaking.ai, and we are hoping to share more applications with other funding bodies in the future, to make life easier for applicants.

Yes! Many of the funders participating in our first round are funds or foundations, not just individuals (such as the Long Term Future Fund and the ARM Fund).

You can have your staff join as evaluators, and use your existing decision-making processes to decide your preferred funding order. You can import any existing grant proposals you are evaluating into the round, which means you get to potentially split the bill with other funders on those grants. You also get access to other institution’s funding applications, increasing your pool of potential funding targets, and get to read and learn from the grant evaluation practices of other evaluators and institutions.

When a grant round gets finalized, every funder is asked to lock in their final amount of funding and allocation order. At that point each funder can see what their final grant recommendations will be. When they are locked in, the funder commits to us and the other participating funders and evaluators that they will send out the funds for these grants in the next 2 months. If they fail to do so, we bar them from ever participating in any Lightcone Commons round again (and will try our best to find replacement funders for the affected organizations).

When any funder makes a change to their final budgets and allocations, all other funders get a chance to do the same in response. Only if everyone explicitly approves their final grant allocation do we move forward with the end of the round. If this somehow ends up going in cycles, I will coordinate some kind of compromise, or make an executive decision to remove a funder from the round (though this seems to me very unlikely to occur).

No. For charitable donations, we have a nonprofit that can receive your donation and handle sending out the grants, final due diligence, and logistics. For grants to non-501(c)3 entities, we have an LLC that creates and signs contracts with target organizations. The costs for these services are covered by the 3% platform fee on all recommendations (which to be clear is charged regardless of whether you make use of these services).

Both of these are optional and you can also make whatever donations or grants directly from your own DAF, LLC or private bank account (though this does not reduce the platform fee).

We pose no restrictions on organizations that evaluators want to recommend grants to, and provide no mechanism for evaluators or funders to block grants by other funders. I aim for Lightcone Commons to be a platform that facilitates coordination between funders, evaluators and organizations with a wide variety of worldviews, many of which will strongly disagree with each other. I think the world will be a lot better if we can nevertheless successfully coordinate, critique each other’s plans, and split the bill whenever many people want to see the same thing funded, even if they diverge strongly on other issues.

(If you wish to use the nonprofit and LLC grantmaking services, I will do my best to make whatever grants you want to make, but do maintain the option to refuse to make any specific grants, and will occasionally face legal restrictions on the grants we can make)

In addition to facilitating funding rounds every 3 months, we also make it easy for funders to run a "direct grantor program" where invited grantors get a special budget that allows them to make grants immediately and unilaterally. This enables more active grantmaking (where the grantors are able to take the time to talk with and get to know a grantee), and allows grantors to use this unilateral grant budget to insure projects against the risk of not getting funding in the next grant round (which enables more risk-taking).

When a grantor funds a project like this, let’s say for $10,000, then in the next grant round the first $10,000 allocated to that project will, instead of going to the grantee, be used to reimburse the grant budget of the grantor.

Beyond that, funders are encouraged to increase the grantor budget of those who made grants that ended up being evaluated as highly impactful, and partially reimburse the budget of those who made grants that almost made the cut of the grant round. We provide UI on our platform to make this easy.

Funders can specify what level of review they want such unilateral grants to undergo before they go out. If they use our nonprofit and LLC grantmaking services we can accommodate both a manual review by the funder, a simple checklist process run by our staff, or a full unilateral system where grants go out as soon as possible.

You can invite anyone you want as a direct grantor onto the platform in the "Direct Grants" section of the app.

Any grantor who participates in another funder’s direct grants program can also choose to be listed as available for being a grantor for other funders. One of the easiest ways to get started with a direct grantor program is to look through the existing list of direct grantors, and give the ones you think will make good grants by your lights a direct granting budget.

There will probably be some allowance for this; see the Grantor Program FAQ addressed at SFF speculation grantors here.

It is up to the evaluators to decide who to share their evaluations with. The platform encourages evaluators to make their evaluations visible and transparent to all participating funders and other evaluators, which makes it more likely that funders will decide to defer to their grantmaking portfolio, and more likely that other evaluators will engage critically with their evaluations. But we provide tools to make evaluations only visible to subsets of funders and other evaluators, and if some evaluations involve sensitive information, or we grow to have a large enough number of participating funders such that expectations of privacy become an issue, I expect more evaluators to limit their evaluations to a narrower audience.

Yes. Evaluators receive the 2% fee charged on any grants given based on their recommendations.

Sometimes a funder wants to bring in an evaluator who is already on their payroll in some fashion; in that case it’s between the evaluator and the funder to determine what to do with the 2% evaluator fee (which should plausibly go to whatever organization employs the evaluator instead of the evaluator directly, or should just be waived for the case of the funder directing money through their own employed evaluator).

That’s not up to us! Since any participating funders maintain full discretion about where their funding goes, there is no super short and simple description of what projects will get funding. That said, I am happy to say a few things about what the funders who are currently participating in the platform have historically funded and have publicly said about their giving priorities.

One of the biggest funders in the upcoming round is Jaan Tallinn who has conveniently written publicly about his philanthropic priorities. He has also been the primary funder behind the Survival and Flourishing Fund, which has made all of its grant recommendations over the last 6+ years public.

At a high level, I would summarize Jaan’s giving as "a very intellectually diverse approach to trying to shape the long-term future of humanity". His giving has historically involved recruiting people he respects with quite different perspectives on the world, and giving them lots of latitude on what they can recommend funding to.

This has produced grants ranging from "advocacy to give legal rights to rivers" to "longevity research" to funding for pronatalist.org, and AI x-risk reduction work at Palisade Research. The bulk of his funding goes to work aimed at aligning, controlling or otherwise shaping superintelligent AI systems, but the tail of his giving is very diverse and it’s hard to rule out much ambitious charity.

One of the other big funders for the first round is Dustin Moskovitz. His most recent essay about his giving is probably most helpful for understanding his perspective. Coefficient Giving (previously known as Open Philanthropy) has also facilitated over a billion dollars of grants of his, many of which are publicly available on their fund pages.

The Long Term Future Fund and the ARM Fund are both contributing ~$2M each for this round. They both focus almost exclusively on how we can shape the development of AI in a positive direction, but are also both open to infrastructure-level interventions to achieve that aim, such as by funding Effective Altruist, Rationality or AI-safety community building, supporting U.S. voting reform, and occasionally the LTFF might also make a grant to avoid other existential risks, like engineered pandemics.

The rest of the funding comes from smaller donors or anonymous donors whose giving I know little about! My guess is they will also skew pretty focused on AI existential risk topics, but I really have a lot of uncertainty here.

No! We can make grants to 501c3s, individuals, and for-profit organizations. For individuals, we will determine whether your work qualifies as charitable, and provide you with fiscal sponsorship if so. Some of our funders can only make charitable grants (but others aren’t limited to charitable grants, so you will be considered in either case).

No! We accept applications from anyone internationally. However, there are some countries where it is in my experience approximately impossible to send money to (in particular Russia and India), so it’s pretty likely we can’t make a grant to you if you are located in one of those places.

We accept rolling applications, with a deadline two months in advance of when grants are recommended. The current round closes on August 23, 2026, for grant recommendations made by (roughly) October 23, 2026. Any applications after that will get a response by (roughly) January 23, 2027.

We try hard to make the application process quick and straightforward. The application has only four required sections:

1. What are you working on?

2. What would you do with funding?

3. Who is involved?

4. Logistical details about your project

You can structure your answers to the first three questions however you want, and the application interface is a live-editable collaborative document.

We might also already have an application from you, if you applied to the Long Term Future Fund, Manifund, or grantmaking.ai in the past few months. If so, you should have received an email with instructions on how to get edit access to your imported applications.

You can update and edit your application any time after you submit it, so don’t worry about getting it all perfect immediately. If your plans have changed, we encourage you to revise your proposal accordingly.

After the application deadline, evaluators and funders using the platform will consider your proposal and make recommendations for how much funding they think it should receive. You can read more about our evaluators and how recommendations work in the Funder FAQ. Any applications received after the deadline will be evaluated in the next round, three months later.

Most projects will hear back when we announce recommendations, usually two months after the application deadline (October 23rd for the first round). You should receive money within a month after that (and technically you could fail due diligence as we try to send you money, though I expect that to happen very rarely, and it can take longer if you are located outside of the US).

Some projects may hear back shortly after submitting, if your project appeals to a direct grantor (individuals who were given a budget by a funder to unilaterally make grants). Direct grants will be limited for the first round, but our goal is that in future rounds most successful applicants will receive at least partial funding within 1-2 weeks of applying.

You can submit multiple applications for your organization, but please try to avoid them being mutually exclusive or overlapping too much. We might consolidate applications ourselves on the backend if you submit multiple, and leave it up to you how to allocate funding between your projects if you do get funding.

If you submitted to grantmaking.ai and Manifund, or the LTFF, we may have imported your application. If your application was imported, you should have received an email with a sign-in link. You can also sign in with the same email address to claim your account.

You’ll likely need at least 40-50 hours of up-front work when you first start evaluating, plus 20–30 hours each round (quarterly) to update your recommendations. This is a lower bound! Many evaluators will spend much more time than that.

The up-front work consists of getting familiar with the applicant pool and writing about your taste and strategy, to help funders decide whether they want to defer to you. After that, you’ll have new applications to read each round, but you’ll be making updates to an existing set of priorities, not starting fresh.

Yes, evaluators are paid 2% of any funding directed through them.

In short: read applications, and rank them in a priority order for funding.

You’ll produce three kinds of outputs:

1. Your funding priority order (most important)

Your funding order is the answer to the question "in what priority order would I recommend distributing funding if I was the only funder in this round".

This kind of ranking allows your priorities to be compared directly to other evaluators’, and used to allocate a variable amount of funds. Funders defer to evaluators by placing you in their own funding order, so the amount flowing through your recommendations is not known in advance and shifts as funders join, leave, and adjust their allocations. For the details of our algorithm see "How does this whole process work on a technical level?".

Your funding order can and should be considerably longer than the amount funders allocate to you, given the way collective funding is allocated. When several evaluators want to recommend the same organization, they split the bill in proportion to the budget they’re working with, so your recommendation tends to go further than expected.

2. Explanation of your funding priorities

You are strongly encouraged, but not required, to explain your funding order. Explanations will frequently consist of writeups about specific organizations for other funders and evaluators to read, based on each organization’s application. But high-level, general material is also useful — your views on cause prioritization, what you look for as signs of strong teams or projects, and your predictions about the future will all help funders understand the thinking behind your recommendation, and feel more confident in deferring to you.

3. Engagement with other evaluators

You are strongly encouraged to engage with other evaluators — reading and commenting on their explanations, answering questions on your own, and debating points of disagreement. Applications and evaluations live in documents that support top-level and inline comments, with configurable visibility, and the platform includes features for notifications and direct messages to help evaluators manage many parallel threads. We also encourage recording and posting synchronous debates, that other funders and evaluators can watch.

Technically, yes. You can sign up with either your real name or a pseudonym, which others on the platform will use to track your priorities and commentary across applications and rounds.

We expect pseudonymity to be most useful when specific funders bring in evaluators who wish to remain anonymous. Currently, the only other way to become an evaluator is for me (Oliver Habryka) to personally select you to participate, which is less likely for evaluators who don’t want their identity revealed. Pseudonymous evaluators are also less likely to be given additional funding to allocate by new funders, since it makes it difficult to find other information about them outside the platform.

By default, your identity on the platform will be visible to all other evaluators and funders — and we aren’t asking anyone to commit to secrecy about which evaluators are part of the program.

Your funding priorities, explanations, and discussions have a few different options for visibility. By default, your funding order will be visible to all funders and evaluators on the platform. Other documents you write — including explanations of your reasoning, and evaluations of specific organizations/applications — can have more granular visibility settings. Options for those documents are:

• Private, visible only to you

• Visible only to a specific set of people you’ve shared it with

• Visible to all funders and evaluators, but not to direct grantors or observers

• Visible to everyone on the platform besides applicants: funders, evaluators, grantors, and observers

Your comments on other evaluators’ work will be visible to everyone who can see the document you commented on — usually everyone on the platform, but sometimes only the specific set of people they’ve shared it with.

All applications are read by a screening team as soon as they come in, so by the time you read an application, it will be tagged with:

• Multiple initial 1–5 star ratings from the Lightcone Commons staff

• A Pangram score

• Tags related to cause area and problem domain

In future rounds you’ll also be able to see past comments about organizations that have applied, from evaluators who made their comments broadly visible.

We give all evaluators a generous LLM token budget and a set of AI tools we’ve developed to help evaluate applicants. We don’t require you to use a specific rubric for evaluation, so you can build whatever workflow works for you.

You also don’t need to read every application. Your funding priorities will be more useful to funders in proportion to how many applications they cover, but you’re free to stick to your areas of expertise and ignore applications outside of it.

Yes! A fund, foundation, or research group can participate as an evaluator, and receive the 2% evaluator compensation as an organization. You’ll need one point of contact to sign up as an evaluator on the platform to submit the group’s recommendations.

The basic process is the same: evaluators submit funding priorities and comment on other evaluators’ priorities, and projects are funded in proportion to recommended budgets.

The biggest change is ongoing engagement: participation in Lightcone Commons is intended as an ongoing engagement, as opposed to the round-based engagement in SFF rounds. Evaluators are welcome to only participate for a single round and then become inactive, but they are also welcome to keep submitting evaluations as an ongoing long-term activity.

The other differences are:

• No required meetings. You can technically participate without ever talking to another evaluator. But discussion with other evaluators helps funders calibrate on how closely your priorities match theirs, so evaluators who don’t engage in discussion will likely see less money routed through them.

• No requirement to review all applications. You can make recommendations that only cover the applications you feel confident in evaluating, though your funding priorities will be more useful to funders if they include more applications.

• No speculation-grant filter. Grants no longer need to have received a speculation grant to be considered in a round.

• Matching pledges get no special treatment. Applicants can mention them; you should treat them as ordinary evidence about cost-effectiveness.

• No special tracks or sub-area rounds

• No required rubric for evaluating applications

Lightcone Commons’ direct grantor program lets funders delegate to grantors who can direct funds immediately, without waiting for a round to close. As a direct grantor, you get a budget under a funder, and can make grants to any funding targets in the system using your granting budget. You get some multiple of your budget refilled after funding rounds, with the multiple depending on how well the projects you funded get evaluated in the funding round.

You can even add your own funding targets, even if they haven’t filled out an application at all! (Applications are generally optional in the Lightcone Commons system).

You start out with a budget of $0. To receive more, you need to be allocated a budget by a funder.

Once you have a budget, you can make a grant using this form in the app:

This triggers a due diligence process by Lightcone Commons staff (and a potential review by the funder), after which we will send out the grant. This will happen as close to immediately as possible.

Since Lightcone Commons has multiple large funders participate in its rounds, there are some basic ground rules to be aware of that come into play when you make direct grants:

1. Your grantor budget is associated with one funder (and other grantors might be associated with different funders).

2. The basic algorithm of the process only enforces that when you make the first direct grant of size $X to an organization that later gets funded for at least $X, then you get $X reimbursed into your granting budget. This is true even when a different funder — not your funder — ends up funding the organization, in which case they "reimburse" your funder for the direct grant your funder’s grantor program sent out.

3. The choice of how much to potentially increase your budget beyond any direct reimbursement is up to the funder under whose granting program you are operating. We provide tools to help them figure out which granting budgets to increase, but the final choice is up to them.

4. Any funder can reduce or increase the granting budgets of their direct grantors at any point in time, for any reason (though they might want to make commitments to their grantors to not do so, but such commitments are not enforced by Lightcone Commons and is considered a private agreement between the funder and the grantors)

Over time we will add more features for facilitating discussion, debate and idea-sharing between grant evaluators. Please let me know what kind of features you would be most excited about, as this is a domain I am particularly uncertain about. I also actively encourage evaluators, direct grantors, and funders to make their own Slack channels, Discord servers, email mailing lists, with any subset of people participating in Lightcone Commons that they want to talk more to (potentially with their own application processes).

Most of the app is built around a wiki-like-system in which evaluators build out personal wikis with their grant evaluations and broader thoughts on their grantmaking principles, which funders can then read and comment on to help figure out which evaluators to defer to in their grantmaking.

As a grantor, you will have access to a subset of that wiki system. Evaluators have fine-grained control over who to share their evaluations and grantmaking essays with. One of the template privacy settings for those evaluations includes direct grantors, and one of the other template privacy settings excludes them, so it will ultimately be up to the evaluators what subset of the wiki you have access to.

If you want, you can also make your own wiki pages, either just for your own private note-keeping privately, or to share with evaluators, funders and other grantors.

As a grantor you do not have access to the numerical grant evaluations and funding orders of evaluators and funders in the system.

Evaluators are distinct from Direct Grantors, and participate in the funding rounds (that make the decisions that determine, among other things, the degree to which direct grantor budgets get refilled).

There are currently two ways to become an evaluator for Lightcone Commons:

1. A funder who is planning on distributing at least $50k per round is bringing you on board as an evaluator

2. I personally select you as an evaluator to participate in the platform

For the latter, please feel free to send me an email with a short pitch about why you want to join as an evaluator.

As soon as you fill out the "Direct grants" form, we’ll sanity-check that there aren’t any obvious typos or mistakes in how you filled out the form, and then move forward on doing some basic due diligence of the recipient (mostly confirming that they satisfy any restrictions the funder wants to put on grants going out from this direct grantor program).

After that, we send the recipient a grant agreement and confirm they are willing to receive the funds. This grant agreement will either be with our DAF entity, or our LLC entity, depending on the recipient. After the grant agreement is signed, we ask the funder to send us the money to make this grant, at which point he also has a final opportunity to veto the grant, and then the money goes out.

Yes! Anyone who signs up as a direct grantor can be allocated granting budgets from other funders. If you are part of multiple direct granting programs, you can choose which budget you want to initiate a grant from in the grant form on the app.

Note that different funders will use different principles for grant budget replenishment, so it can matter quite a bit from what budget you initiate a grant.

"Speculation Grants" are called "Direct Grants" in the Lightcone Commons platform.

Direct grants are not used for any kind of pre-filtering in the Lightcone Commons process, so don’t worry about comprehensively trying to stay on top of all incoming grants.

You will also no longer get an email notification for every single application that enters the system! Instead, we will send you weekly or daily digests (depending on the number of applications coming in) that link in a condensed fashion to the full applications in the Lightcone Commons platform. If you want to receive emails for every application that comes in, or adjust your email notification settings, you can do that in the app under your account settings.

Discussion is no longer by default happening across thousands of email threads. Every application has a Google-Docs-like document with all of the application info, that any evaluator can leave both top-level and inline comments on (with a setting on whether a thread is visible to only grantors, evaluators and funders, or the applicant as well). When you comment on an application you will get email notifications (or text notifications if you set that up) for each new comment on that application.

Multi-sigs are de-emphasized compared to SFF. You are encouraged to just go and make a grant to an organization if you think it’s a good idea. You are not encouraged to wait on other grantors to join you in making a grant. Indeed, by making a grant earlier, you are more likely to get your grant budget reimbursed or increased (since the first grants that go out, are also the first to be reimbursed and score the highest on any proxies of quality, since we assume diminishing marginal returns in the algorithms behind Lightcone Commons). This is experimental and maybe the social dynamics around just encouraging people to make grants immediately without coordinating with other funders on multi-sigs will turn out badly, but I currently believe it will be better.

I am planning to use the 2023 Lightspeed Grants codebase as the basis for this project, with some major modifications. We will migrate from using Airtable as a backend to using a standard Postgres database hosted on Neon, and will migrate a lot of the frontend code to somewhat more modern patterns (e.g. using Tailwind instead of styled components).

I aim to reach almost complete feature parity with the SFF S-Process, but am planning to substantially redesign a few non-core parts of the process where I am dissatisfied with the current implementation.

First, I want to remove special code and special treatment around matching pledges. Grantees are welcome to indicate in their application that they are seeking matching pledge funding, but evaluators should evaluate whether those matching pledges increase the cost-effectiveness of a grant the same way they evaluate all other evidence about the cost-effectiveness of a grant.

Second, I intend to remove requirements for every grant to first receive a speculation grant in order to be considered in a grant round. Furthermore, I want to remove any requirement by evaluators to look comprehensively at all applications submitted to a grant round. Evaluators are encouraged to build grant portfolios that limit themselves to domains where they have expertise.

Third, I expect to rebuild the tools for getting AI systems to assist with evaluations from the ground up, mostly just because I don’t like the current rubric-based system and want to give evaluators more freedom on how to leverage AI systems in evaluating grants.

Fourth, I want to remove requirements for evaluators to attend meetings and engage with other evaluators in highly specific ways. It is OK for an evaluator to be on the platform and never talk to anyone else (though doing so will likely result in funders trusting them less and distributing less funding through them). I want to create affordances and incentives for engaging with the evaluations of other funders both live and asynchronously, but not make it a requirement.

Fifth, I am not planning to have any kind of special tracks, or special grant rounds focused on sub-areas, which feel more opinionated than makes sense for neutral infrastructure like this (though individual funders and congregations of funders are welcome to do special promotions or requests for proposals, and we’ll likely help at least with the basic logistics of those).

Eventually yes, but I am not guaranteeing that we will have made it open source by the first grant round. A project being open source introduces a few more security risks (such as accidentally leaking credentials or tokens in the code) that I would like to not have to think about early on, but I am aiming to make it possible for someone to fork the code behind this project and run their own copy of it within a few months of getting started.

Ideally we simply import all applications from SFF into the Lightcone Commons grant round, and make it easy for applicants (via simple email confirmation) to link any additional information they want to provide to Lightcone Commons evaluators to their SFF application.

If that isn’t possible (because e.g. SFF doesn’t want to share application details with us), we can run a post-hoc process where we privately determine overlap between a given SFF application and Lightcone Commons application, and share that information with any funders participating in the SFF.