People
Updated 05/22/26Co-Founder
Chief Executive Officer
Funding Details
Updated 05/22/26- Annual Budget
- $5,300,000
- Current Runway
- -
- Funding Goal
- $3,000,000
- Funding Raised to Date
- -
Org Details
Updated 05/22/26Lightcone Infrastructure is a 501(c)(3) nonprofit organization headquartered in Berkeley, California, dedicated to building technology, community infrastructure, and physical spaces for people working to improve humanity's long-term future. The organization was announced on October 1, 2021, as an expansion of the LessWrong development team that Oliver Habryka had founded in 2017 to revive LessWrong.com and reinvigorate the intellectual culture of the rationality community. Lightcone's flagship project is LessWrong.com, a community forum focused on rational thinking, AI alignment research, and long-term decision-making. LessWrong is widely regarded as upstream of a significant portion of AI safety work, serving as the primary venue where key ideas in alignment research are developed and debated. Lightcone also operates the AI Alignment Forum, a more focused companion site for technical AI safety research. In 2022, Lightcone acquired the Rose Garden Inn, a hotel property in downtown Berkeley, for $16.5 million with a $20 million loan from Jaan Tallinn's Slimrock Investments, and renovated it into Lighthaven, a 30,000 sq. ft. campus with approximately 40 bedrooms and 80 beds. Lighthaven hosts conferences, research fellowships, and intellectual gatherings. Notable events include LessOnline (a festival for bloggers and truthseekers) and Manifest (a prediction markets conference), which together draw over 600 attendees each summer. The campus has also hosted MATS AI safety fellowships, the Progress Conference, SPARC (a program for talented high schoolers), and Inkhaven writing residencies. Lightcone runs Lightspeed Grants, a rapid-turnaround grantmaking program for projects that help humanity flourish, accepting grant requests from $5,000 to $5 million. The organization also co-built the grant evaluation infrastructure used by the Survival and Flourishing Fund, which has distributed over $100 million in donations over the past five years. In 2025, Lightcone contributed significantly to the AI 2027 project, a high-profile scenario analysis created in collaboration with the AI Futures Project. The organization transitioned from an approximately $8 million per year budget to around $3 million per year following the FTX collapse in late 2022 and changes in the funding landscape. In 2025, Lighthaven had approximately $3.6 million in expenses (including a $1 million annual mortgage interest payment) and generated about $3.2 million in program revenue. Lightcone's other projects, including LessWrong, required an additional $1.7 million. The team consists of eight people: Oliver Habryka (CEO), Ruby Bloom, Jacob Lagerros, Ben Pace, Raymond Arnold, Rafe Kennedy, Robert Mushkatblat, and Ronny Fernandez.
Theory of Change
Updated 05/22/26Admin editLightcone Infrastructure believes this century is a critical period for humanity, particularly due to the development of advanced AI, and that building strong epistemic infrastructure is essential to navigating it well. Their theory of change operates across three channels. First, by maintaining LessWrong and the AI Alignment Forum, they provide the primary intellectual commons where AI safety ideas are developed, stress-tested, and disseminated, serving as a key upstream driver of alignment research. Second, through Lighthaven they create physical spaces where researchers, forecasters, and thinkers can collaborate intensively, hosting programs and conferences that catalyze new research directions and build community ties. Third, through grantmaking infrastructure like Lightspeed Grants and their work with the Survival and Flourishing Fund, they help allocate philanthropic capital more effectively toward reducing existential risk. They take an active leadership role in the rationality and AI safety communities, working to identify and fix bad incentives, promote clear thinking among key decision-makers, and take responsibility for the end-to-end effectiveness of these communities.
Grants Received
Updated 05/22/26Projects– no linked projects
Updated 05/22/26Discussion
Key risk: The primary concern is that much marginal funding would underwrite a capital-intensive Berkeley venue and broad community operations whose causal pathway to existential risk reduction is indirect and hard to measure, creating counterfactual and execution risks relative to funding direct alignment research.
I think this project is great!
This is a community endorsement comment
Case for funding: Funding Lightcone buys high-leverage upstream infrastructure—LessWrong/Alignment Forum, the Lighthaven campus, and Lightspeed/SFF grant systems—that has repeatedly seeded alignment ideas, convened top researchers (e.g., MATS, LessOnline/Manifest), and improved capital allocation, making many other AI safety efforts more effective.