AI Brains

A special Brains cohort to accelerate for AI-focused research programs

Picture of purple material

Speculative Technologies is running a special cohort of the Brains Accelerator focused on helping spin up ambitious AI research programs, with a special focus on security and governance capabilities.

You might have noticed that the world of AI research is moving blisteringly fast. The world needs ambitious, high-leverage AI programs whose scope is bigger than a single company or lab. This program is meant to help talented researchers with experience in AI hardware and software build skills, refine ideas, and make the connections to spin up coordinated research programs in governments and nonprofits.

The program will run from August to November 2025, with applications due June 16, 2025.

To learn more about how the Brains Accelerator works, check out the main Brains page.

THE FELLOWSHIP

Brains provides training, mentorship, and connections to help scientists and technologists design and execute on coordinated research programs. If they already have a home for their program, Brains helps them take their idea and their ability to execute on it to the next level. Think YCombinator for coordinated research programs.

Brains supports fellows with:

Training

The accelerator focuses on calibrating risk and reward, thinking at the level of coordinated programs, and using best practices for program design and management.

Mentorship

Experienced mentors from DARPA and beyond help rapidly refine fellows’ ideas for execution and impact.

Networks

The team and our community connects fellows to funders, partners, and peers; helping them find the organizations and people they need to execute on their idea – whether that is a funder for an FRO or a role in an ARPA organization.

Brains is a remote, 15-week, part-time program, bookended by two in-person events – a 2-day kickoff a concluding symposium. Every week fellows will:

  • Meet with a mentor who has experience in the trenches at DARPA or other coordinated research organizations.

  • Spend several hours on personalized activities that help embed lessons and drive towards running an effective program.

  • Attend a small group meeting to discuss the week’s work with their peers.

  • Attend a panel or closed-door fireside chat with FRO founders, ARPA leaders, and philanthropists.

AI Brains cohort will run from August through November 2025. If you’re someone with a deep technical experience in AI and an ambitious idea that is beyond the scope of a single startup or academic lab, we encourage you to apply!

Key Dates

May 12, 2025: Applications open.

June 16, 2025: Application deadline.

July 8, 2025: Admission decisions finalized.

August 6-7, 2025: In-person kickoff.

We will fly all fellows and mentors together to Washington DC to meet in person and kick off the program.

August — November 2025: 15-weeks of training and mentoring.

November 2025: AI Brains Symposium, an in-person event to showcase fellows’ ideas.

We will fly all fellows and mentors together to a location in Washington DC to meet in person and pitch your programs to interested partners.

FAQ

You can see the FAQ for the main Brains Program here. If you have any questions that aren't answered on the FAQ, please email us at brains@spec.tech

As always, Yes.

Like core Brains, ideal participant:

  • Has a deep technical background.
  • Has done the equivalent of PhD-level research.
  • Has worked both inside and outside of academia.
  • Has a strong bias towards action.
  • Is comfortable with acting under uncertainty.
  • Is interested in pursuing their ideas by getting a new role and/or starting a new (not-vc-funded) organization to pursue their ideas.

For the AI cohort, we're going to bias towards people who:

  • Have experience building (not just using) AI software or hardware (for obvious reasons).
  • Have American citizenship (so that you can work on classified things).
  • Are willing to consider moving to DC and taking a pay cut to work on excellent things for the public good for a few years.

Of course, there are exceptions to every rule, especially on the frontiers of human capabilities.

  • Technical ideas around AI
  • We are open to any technical AI ideas, but will slightly bias towards ideas that focus on AI itself like security, governance, privacy.
  • Work that would benefit from the coordinating power of government organizations.
  • Work that needs to be parallelized across several groups to succeed,
  • NOT products or research that mostly benefits the organization that does it
  • The cohort will be smaller: seven people instead of 16.
  • The program will be especially focused on creating a bridge between Silicon Valley and DC.
© 2025 Speculative Technologies.
Created by And–Now and Anson Yu.