8

Building and maintaining the Alignment Ecosystem

Not fundedGrant
$0raised

Project summary

We will maintain, improve, and create resources in the alignment ecosystem, similar to the ones we’ve already built such as AI Safety Communities, AI Safety World, EA Domains, AI Safety Training (in collaboration with AI Safety Support), and Alignment Ecosystem Development. We want to finish and launch projects that are currently works in progress, such as AI Safety Careers, Grantmakers for AI Alignment, and EigenKarma Network. We will also support our partner projects such as AI Safety Info, AISafety.com, AI Safety Quest, EA Gather Town, and AI Plans.

What are this project's goals and how will you achieve them?

We aim to improve resources in the alignment ecosystem to allow people to find the things they need to work effectively, such as communities, training programs, funding sources, career and research opportunities, etc.

By moving from an entirely volunteer-driven structure led by one self-funded person to a professional team of three, we will have the bandwidth to create higher quality and more reliably maintained infrastructure to support the large number of new people who are trying to join the effort to save the world.

How will this funding be used?

$30,000/y salary for each of  3 employees, accounting to a total of $90,000

$5,000 software and travel budget

$10,000 buffer

5% fiscal sponsorship fee

Total: $110,250 (or £89,000)

--

At lower levels of funding, we would run the project for less time, possibly with a smaller team. At higher levels, we would have flexibility to take on more staff or pay contractors to do specialized work.

At the minimum funding bar of $36,750 (total), we would likely dedicate 2.0 FTEs to the project for 6 months. We estimate this would allow us to make significant progress on new projects, while maintaining and coordinating volunteer efforts on the current projects. 

At the target funding bar of $110,250, we would dedicate 3.0 FTEs to the project for 12 months. We estimate this would allow us to bring many projects to a high quality, and have spare capacity to pick up new projects as they arise, while also allocating efforts towards directly improving the current projects.

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

The team came together in September 2023 on a hackathon aimed at designing and building a prototype of AISafety.com.

plex has been a full-time self-funding free agent building resources for the past three years. He’s stepping back from management of AED in order to pursue research directly and will not be receiving any funding form this, but will stay available for the pass-over and as an advisor for the new team, and is keen to see others step up to keep the resources well looked after.

Bart Jaworski started a small company making websites when he was 18 and shut it down after around 1.5 years due to the market saturation by companies such as Wix.com and Squarespace. Since then, he has done theoretical and technical alignment research in Oxford. Currently, he's helping with operations for DirectEd Development Foundation.

Rika Warton has been volunteering with AED for around 6 months, helping maintain various resources like the map of AI safety.

Siao Si Looi currently works at AISafety.info, writing and editing on AI safety, coordinating hackathons, and contributing to UI/UX design. She has experience in community infrastructure web projects, and runs the EA Gather project, which has been used by over 20 EA organizations.

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

  • Team members might find other things which are of higher impact to work on. If this is because other things are higher impact but this is still high impact, we would look for replacements.

  • It may be more difficult than we think to create high-quality resources, such as because it requires more specialized knowledge than we have, or involves more time-consuming processes than we envision. This would mean we produce fewer outputs than we expect.

  • We may find that creating high-quality resources in all the most important places is completed early with several people focusing on them. If this happened, we would focus on supporting the partner projects with most of our time.

  • There could be difficult team dynamics, though the team has worked together before and this seems unlikely.

  • Significant changes to the AI Safety landscape could mean that the projects are less helpful.

  • The resources might not reach the people who need them, if we fail to share the resources well.

What other funding are you or your project getting?

AED has received zero funding so far, and has until now been entirely volunteer driven. We have no other open applications as of the date of submission.

Siao is currently a funded AISafety.Info fellow, and received accommodation and board from CEEALAR from January to September.  Rika received accommodation and board from CEEALAR for six months from September.  

KabirKumar avatar

Kabir Kumar

about 1 year ago

From what I know, AI Safety Careers isn't funding constrained- how would the funding help with this?

plex avatar

plex

about 1 year ago

AI safety careers mostly needs content writing and editing. We've looked for volunteers, and there has been some progress, but things would move much faster if we could take on a contractor who could focus on this full time, perhaps by sponsoring an AI safety info distillation fellow as the content is automatically used by both sites.