Dear Apart Community,
Welcome to another special edition of our newsletter - Apart News!
We've received another $462,276 since our last newsletter, making the total $597k of our $955k goal! 🤯 We cannot thank all of you enough for your contributions - here are a few highlights:
We've been incredibly grateful for all the support we've seen over the last three weeks and we're excited to announce that we're extending the campaign by 25 days, making our campaign end date July the 13th.
– so now is your chance to share it with friends / family and help support crucial AI safety efforts during this important time!
New Paths To Impact
Since our last newsletter, we've heard hundreds of new thoughts on how Apart has impacted your journey into AI safety. We already knew that our Sprints → Studio → Lab…
Makes it possible for anyone, only based on hackathon projects, to start a career change even during a full-time job and from anywhere in the world,
Creates a Hacker-First culture of thoughtful researchers who are used to sharing their work,
And helps people skip the “upskilling loop” by engaging practically with deep research fields through published work
But beyond these, we've been excited to find new ways we've had a real impact on the community as well:
Apart's culture accelerates self-starters
At every step, we emphasize learning through direct action.
Researchers are invited to the Lab through our hackathons, the Lab isn't done before a finished product exists, and people lead their own agendas with help from us and our mentors.
When this is the baseline for participation, our fellows end up shaping new AI safety organizations, startups, and communities by themselves as well!
Here's a few that were founded by fellows after they joined us (among many others founded both before and after!):
Accessibility is even more important than we expected
We have always had a focus on accessibility and inclusion, but the remote-first nature of Apart's Studio and Lab Fellowship removes deal-breaking barriers to entry in the AI safety space to an even larger degree than expected.
Here's what Amir Abdullah (now leading an AI safety team at Thoughtworks) says:
“I did my Ph.D in Theoretical CS from 2010 to 2016, but found myself out of research positions for the next several years due to location, time and family constraints […] As a consequence of Apart's support and the 3 papers [I published with them,] I received my first full time offer as an AI safety researcher at Thoughtworks as an early team member working in AI safety and understanding […]”
Amir (left) presenting at NeurIPS 2024
Academic engagement creates great experiences
With every Lab team comes a conference or workshop publication, a unique output-first approach that we have pursued since day one to give everyone the chance to become career researchers.
In the past year, we've had (at least) 12 first-timers at ML conferences (of 28+ total conference attendants among our fellows), and all of them are still involved in AI safety today.
An example is Christian Caruba, who says:
“Apart’s guidelines and structure helped me achieve a publishable result, which is something I couldn’t have done on my own, even with a university supervisor. Attending NeurIPS to share our results was an eye-opening experience, particularly valuable in my early stage as a researcher, that would have been impossible for me without Apart’s support.”
Cristian (right) presenting at the Sys2Reasoning workshop at NeurIPS 2024
Never Stop Never Stopping
"We'd like to get to the point where [AI safety] is everywhere, like oxygen or gravity or clinical depression." - Paula (lightly edited) from ‘Never Stop Never Stopping’
Since the last newsletter, we haven't stopped executing on what's important. The work you're making happen with your donations:
The 1-day Policy Sprint for ControlAI ran last Friday, getting 107 signups and 20 entries after just over a week online! Stay tuned to hear about the winning entries that will hopefully shape future AI legislation.
We've published the upcoming (July 25 to 27) AI Safety X Physics Grand Challengein collaboration with PIBBSS, Timaeus, and others - Sign up now!
The reward ceremony for the 18 entries at the Martian Mech-Int Sprint has finished and we've already sent Studio invitations to the top teams (with mentorship from Martian's mech-int team!).
The Studio (our 2-week fast-track from hackathon report to blog post and 4-week journey to a research proposal) has already received three blog posts and accepted two fellows from our Economics of Transformative AI Grand Challenge held 1.5 months ago.
And two other blog posts have been published from the Studio:Schelling Coordination via Agentic Loops — Apart Research AI Control Hackathon and SafeRLHub: An Interactive Resource for RL Safety and Interpretability.
And you will see Apart play a continued role in the AI safety community throughout the next years through direct research, incubation of new agendas, and new ways of exciting community engagement.
The Fundraiser Is Still On
As mentioned earlier, thanks to all of your support, we are able to extend our fundraising efforts!
Our next goal is $954k, which will secure sustainable funding for 12 months as we make room to grow with institutional funders.
That means we can still use your help
Spread the Word!
While it may not seem like much, it’s supervaluable to share this campaign in all the ways you can:
Follow us on X and LinkedIn and like and share our posts whenever you see them (for example this one!)
Forward our newsletter to people you know who would be interested in supporting independent human-first AI research
Head over to our donation page or Manifund and leave a comment about how Apart has impacted your AI safety travels
Donate
Every contribution makes a difference:
$10 gets your name on our donation wall (a real difference!)
$200 Enables the participation of 3 hackathon participants
$5,000 supports a Lab Fellow producing publication-quality research and publishing at a conference workshop (including conference attendance)
$50,000 enables 3 global research Grand Challenges sprints identifying new safety approaches and accelerating sub-fields of AI safety
$100,000 enables the publication of 4 peer-reviewed main conference AI Safety publications
Help us make AI safer!
For those who can give comfortably, becoming a monthly supporter helps us plan with confidence:
Keep Apart sustainable! (click "Give Monthly" when you donate)
Large donors or sponsors are welcome to reach out to us directly at hi@apartresearch.com for collaboration opportunities or sponsorships.
Remember: All donations are tax-deductible through our fiscal sponsor Ashgro Inc, a 501(c)(3) organization.
Thank You
Again, we want to emphasize that we could not have done this without your help. In 3 years we’ve been able to accomplish awesome things. And we cannot wait to see what we can all accomplish together in the next 3.
Jason and Esben
