What are you looking for in a founder?
We’re looking for people who can solve a piece of one of the most important problems faced by humanity today.
AI could have a transformative impact for better or worse in many ways, but this call is specifically for founders to take a direct shot at reducing the global catastrophic risks posed by the technology.
Moving the needle will be hard. You’ll need to articulate a compelling theory of change, and make progress with limited feedback loops – nonprofits don’t have “increasing revenue” to tell you if something’s working, and even if you start a for-profit, revenue won’t necessarily track the impact you care about. On top of that, you’ll have to navigate a field where experts disagree on fundamental assumptions. And you’ll need to manage uncertainty, pouring years into achieving an outcome that’s nearly impossible to measure: avoiding a catastrophe.
And those are just the challenges of working on AI safety. You’ll still have to confront the standard challenges of launching a startup: hiring, fundraising, managing, communicating your work, and engaging with stakeholders. We think the stakes are high enough to be worth trying anyway.
This means we’re looking for founders who are ambitious and move quickly, who can ruthlessly prioritize impact, and who can build excellent teams as they scale.
We are particularly excited about individuals with any of the following: strong technical ability; experience building complex hardware, software, or security-critical systems; deep research expertise in machine learning or related fields; auditing, assurance, or red-teaming experience; experience founding or scaling companies, non-profits, programs, or events; exceptional communication skills with technical and non-technical audiences; or an excellent track record identifying talent.
Who shouldn’t launch something new?
Founding something isn’t for everyone. It is, among other things, extremely time-intensive, so if you aren’t ready to dedicate considerable time to this work over the coming years you might want to consider another option. You can still contribute meaningfully to AI safety by finding an impactful role on an important team.
Sometimes joining forces is better. Some initiatives might be particularly good fits for an existing organization because of returns to scale, preexisting expertise, relationships and reputation, or access to limited resources. Before starting something new, we strongly recommend talking to the people who might already be working on it (a good use of a pre-seed exploration window). You may end up concluding that it makes more sense to pitch us on starting a new team at an existing org, and we want to hear about those opportunities too. But if you conclude it still makes sense to launch something new, you will have built the most relevant connections and gathered some credible supporters for your work.
What are the available funding options?
Pre-seed funding: $200k – $2m, usually structured as research or planning grants to develop your idea, build a team, and validate your approach over 9-18 months.
Note you don’t have to wait until the end of your initial funding window to apply for more funding – if your idea is going well after 3 months, we want you to apply for seed funding to scale!
Seed funding: $2m – $20m, to allow your founding team to launch and scale a leading AI safety organization over 18-36 months.
Scale funding: $20m – $200m+ range. We’re excited to explore larger funding options to establish new organizations with a world-class founding team, or allow organizations with a strong track record to continue scaling or launch a new program. We expect these opportunities to directly target a key problem in AI safety and have a meaningful likelihood of success.
Can I pay myself and my team competitive salaries?
Yes – we think talent is the most important constraint to impact and we want budgets that (a) pay founders properly so you can focus on the work, and (b) offer salaries that let your initiative compete for top talent. If pay is your main hesitation, we’d still like to hear from you: it’s usually a solvable problem.
What does the process look like?
Promising candidates will be invited to submit a proposal with your track record, relevant capabilities, and vision for the initiative. Especially strong candidates may be fast-tracked directly to the next step: a short pitch meeting with members of our team to discuss their plans and potential impact. We aim to make a rapid decision and disburse funds shortly after that conversation.
Please note, this is our first call for founders, and we don’t know how much interest it will get. As a result, we do not expect to respond to every submission we receive.
In order to manage volume, our policy is to automatically filter out submissions that propose initiatives not directly related to catastrophic risks from AI, or that our systems indicate were generated by AI. It is also possible we will stop accepting submissions if we receive a large volume, in order to ensure we can continue to move quickly on the most promising submissions.
When will I be able to apply for more funding?
Whenever you’re ready! If you receive pre-seed or seed funding and the work is going well, you don’t need to wait until the end of your initial funding window to apply for a renewal or expansion of our support. For instance, you might apply for seed funding within three months of receiving your pre-seed funds if you’ve tested a few ideas, gotten traction with one, and are ready to scale.
Are you only funding the ideas on your list?
No – if you have a better idea, pitch it to us. We expect the very best founders will have strong opinions on the best approach to reduce catastrophic AI risks, and we want to hear them.
Do you support work outside the U.S.?
Yes – we can generally support work internationally, but non-U.S. individuals and entities can sometimes take longer to receive funding or be subject to additional requirements depending on the specific location.
Do you ever fund for-profit organizations?
Yes – we’re open to it if the work is clearly aligned with our mission and demonstrates a charitable purpose. If a public-benefit corporation structure would likely make your organization more effective, we are happy to consider these proposals (but we don’t expect to fund non-PBC for-profit initiatives). Even if they aren’t a fit for us, we have strong relationships with several safety-motivated for-profit investors and (with your consent) can forward the best opportunities to them.
That said, we are best positioned to fund nonprofits, and we expect funding for-profits to add a bit of additional complexity and time to the grantmaking process.
Do I need to have a legal entity set up?
No – we’re open to giving pre-seed funding to individuals or founding teams who haven’t yet set up an entity. If you are applying for seed funding, you should have (or be ready to quickly set up) a legal entity, either directly or through a 501(c)(3) fiscal sponsor.
Are you hiring?
We just wrapped up a hiring round for our AI programs, but you can express interest here or check for other open roles at Coefficient Giving here. We also think the 80,000 Hours job board is an excellent resource for finding impactful roles in AI safety.