Sydney AI Safety Space - General Funding

3 Little Queen St, Chippendale, NSW, Australia 2008

The Sydney AI Safety Space is a co-working hub in Sydney, Australia, offering free office space for people working in the field of AI safety.

Endorsed by

Ryan

The Sydney AI Safety Space is a co-working hub in Sydney, Australia, offering free office space for people working in the field of AI safety.

Endorsed by

Ryan

Project Details

This request is for the general operation of the Sydney AI Safety Space through to July 2027.

At the low end, the primary expense is for office space. We expect to pay approximately US$56,000 for our office space usage over the next 12 months (an average occupancy of 8 desks between over 20 members), with approximately $14,000 of that accounted for with member fees, and $33,000 from existing cash on hand. A buffer for events, fiscal sponsorship fees, insurance, and amenities is added on top for a minimum viable ask of $20,000 USD for twelve months of operation.

With additional funding, we will be able to hire a part time office manager from within the community to take over some of the voluntary work I have been doing, and make sure the space is clean, tidy, and welcoming all the time. This might cost about $15,000.

We are also interested in running a micro-grants program. Over the last year, we have noticed several opportunities where a small grant (currently intending AU$1000 [US$716] per week for six weeks) would have helped support people at the early stages of projects, and reduced the risk for them to begin working on AI safety.

I think SASS as a co-working space is in a unique position here -- we see new projects appear at the earliest stages, and can make assessments of quality based on co-working with the individuals requiring funding rather solely via traditional paper applications. A low-overhead, rapid, micro-grant system -- similar to BlueDot Impacts Rapid Grants program -- might be able to significantly increase the odds that someone jumps into work full time, or does so sooner than they otherwise would be able to.

I also suspect that being able to offer micro-grants might make it easier for people to find and engage with SASS. While we have been growing (up to 17 members in April 2026, compared with 10 in June 2025), a small grants program increases our legitimacy and quality of our marketing further, and might lead to more high quality people finding and working out of SASS.

Theory of Impact

SASS, as a coworking space, has already grown the AI safety community in Sydney, and helped produce orgs such as Lyptus Research and given researchers a place to work while in town for events like the AI Safety Forum.

Having an office filled with capable people working on important problems is motivating. It makes it easier to get feedback and throw around ideas, it makes it possible for research groups and projects to form with people who meet here. It grounds our ideas and prods us to do better, and is a hub for news and events. It also provides a network where people can get help, mentorship, funding, and so on, outside of the Bay.

It’s a diffuse sort of impact, because many of our members would have been pushing their work forwards regardless, but I think SASS’s existence has been a catalyst to make our work go faster and further. And I think if we are able to use our position to fund promising work very early on, before it makes it into a formal application, we'll be able to do more to help other projects get started.

People

Hunter Jay

Team Member

Funding Details

Start Date: Mar 1, 2025
End Date: -
Expected Duration: -
Funding Raised to Date: -
Annual Budget: -
Monthly Burn Rate: -
Current Runway: -
Funding Goal: -
Funding Stage: Partially funded
Fiscal Sponsor: Good Ancestors Policy Ltd

Track Record

SASS is a coworking space, so our contributions are quite diffuse and mixed with work that our members would have done independently. That said, the most notable things that SASS has supported in FY26:

The founding of Lyptus Research, which received attention recently for replicating the METR time-horizons chart in the domain of offensive cyber. Sean Peters, Jack Payne, Slava Chalnev, as well as myself as a board member (Hunter Jay), were all working out of SASS independently prior to Lyptus being founded, and I think SASS can be assigned some credit for the coming together of this specific group on this work.
The Sydney edition of the Apart AI Control Hackathon, run by Stephen Elliott, and with several other SASS members taking part.
The Sydney AI Safety Fellowship, ran by Chris Leong and Jack Payne, which ran out of the space with support provided for food and amenities.

Grants Received

Add grant

Brayden McLean

from Brayden McLean

$10,000

AI Safety Tactical Opportunities Fund

from AI Safety Tactical Opportunities Fund (AISTOF)

$38,000

Funding Asks

grantmaking.ai Launch Round

Applied

Minimum

$20,000

Ideal

$86,000

What this ask is

This request is for the general operation of the Sydney AI Safety Space through to July 2027.

At the low end, the primary expense is for office space. We expect to pay approximately US$56,000 for our office space usage over the next 12 months (an average occupancy of 8 desks between over 20 members), with approximately $14,000 of that accounted for with member fees, and $33,000 from existing cash on hand. A buffer for events, fiscal sponsorship fees, insurance, and amenities is added on top for a minimum viable ask of $20,000 USD for twelve months of operation.

With additional funding, we will be able to hire a part time office manager from within the community to take over some of the voluntary work I have been doing, and make sure the space is clean, tidy, and welcoming all the time. This might cost about $15,000.

We are also interested in running a micro-grants program. Over the last year, we have noticed several opportunities where a small grant (currently intending AU$1000 [US$716] per week for six weeks) would have helped support people at the early stages of projects, and reduced the risk for them to begin working on AI safety.

Had we been running this micro-grant program in FY26, potential candidates would have included Slava Chalnev's Cycle-Consistent Activation Oracle work (which was later referenced by Anthropic), Sean Peter's career transition and early cyber work -- which later became Lyptus Research -- and Stephen Elliott's multiagent systems work.

I think SASS as a co-working space is in a unique position here -- we see new projects appear at the earliest stages, and can make assessments of quality based on co-working with the individuals requiring funding rather solely via traditional paper applications. A low-overhead, rapid, micro-grant system -- similar to BlueDot Impacts Rapid Grants program -- might be able to significantly increase the odds that someone jumps into work full time, or does so sooner than they otherwise would be able to.

How the money will be spent

Primarily on office space (~US$55k). If we reach a stretch goal, then on a part time office manager (~US$15k) and on up to twelve $4.3k micro-grants throughout the year (US$52k).

A detailed budget is here: https://docs.google.com/spreadsheets/d/1NcPt4akColO4q51zWMAreXc6E2U7ZQd6axc93Fp62yE/edit?usp=sharing

Discussion

No comments yet. Be the first to share your thoughts.