Long-Term Future Fund (LTFF)

Theory of Change

LTFF's stated theory of change has multiple layers. At the highest level, the fund aims to "improve the long-term trajectory of humanity" by reducing global catastrophic risks, primarily from AI. In practice, roughly two-thirds of grants go to AI safety work.

The fund articulates its value-add relative to larger funders (Open Philanthropy, SFF) through several mechanisms:

• Early-stage access: LTFF specializes in small grants (median $25K vs. OP's $257K) to individuals and early-stage projects, serving as a "first funder" that validates people for later, larger grants.
• Speed and openness: LTFF has the largest always-open application form among longtermist funders. Anyone can apply at any time (though it returned to round-based applications in Q1 2025).
• Epistemic diversity: Linch Zhang, in the 2023 fundraising post: "Supporting a set of worldviews that we find plausible and that are not currently well represented among grantmakers."
• Contact with reality: Fund managers spend most of their time doing direct work (running LessWrong, doing AI evals at METR, etc.), not just grantmaking, which allegedly improves their judgment.
• Hits-based giving: When fund managers disagree, LTFF "lean[s] towards funding when one fund manager is very excited." One enthusiastic manager can drive approval.

The fund explicitly does not try to save money for future years. If donations surge, the bar drops. If donations dry up, the bar rises.

What They Do

LTFF has disbursed approximately $20M since its 2017 founding, with roughly $10M going to AI safety. The trajectory tells a story:

The scaling from $1.3M to $12M and back to $5M tracks the FTX-era funding boom and bust.

Notable funded projects include SERI MATS ($316K talent pipeline), Alexander Turner shard theory research ($220K plus earlier grants), Robert Miles AI safety YouTube ($60K + $121.5K), Manifold Markets ($200K), Existential-Risk Persuasion Tournament ($572K to Tetlock/Karger), and Vanessa Kosoy ($100K). The portfolio is heavily weighted toward individual researchers and small teams.

A retrospective of 2018-2019 grants found 26% exceeded expectations, 22% met expectations, 13% fell short, 13% failed, and 26% had too little public information to evaluate. The retrospective author described the 30-40% success rate as "intuitively disappointing" but potentially appropriate for hits-based giving.

LTFF returned to round-based applications in Q1 2025 after years of rolling applications. The fund acknowledged this reversed a key earlier innovation but was necessary because rolling applications created unsustainable workload for part-time managers.

Key People

Caleb Parikh -- Project Lead of EA Funds and interim LTFF Chair since late 2023. Previously worked on epistemic community health at CEA. Despite a job posting for a full-time permanent chair in October 2023, the position does not appear to have been permanently filled as of early 2026.

Linchuan (Linch) Zhang -- The only full-time fund manager at LTFF, formerly a senior researcher at Rethink Priorities. His candid post "Some unfun lessons I learned as a junior grantmaker" revealed that grantmakers spend very little time on any given grant, conflicts of interest are unavoidable, and "people are grateful to you for granting them money -- this is a mistake." He is the fund's most public-facing figure.

Oliver Habryka -- Fund manager since ~2018, founder of Lightcone Infrastructure (LessWrong). Notable for funding researchers whose directions he "pretty strongly" disagrees with if the work quality is high. Has publicly expressed "reservations about the LTFF" while remaining a fund manager.

Notable departures: Asya Bergal (chair Feb 2021-Oct 2023, stepped down after joining OP), Helen Toner (fund manager ~2018-2021, later served on OpenAI board), Adam Gleave (fund manager, left to become Manifund regrantor), Evan Hubinger (fund manager, left for Anthropic alignment research). The pattern of departures is consistently: talented people prioritizing their primary work over part-time grantmaking.

The fund has 2 full-time employees managing eight-figure annual turnover -- an extreme staff-to-budget ratio.

Money and Incentives

Revenue and budget: LTFF-specific financials are not independently verifiable. The fund operates under the Effective Ventures Foundation umbrella (EIN 47-1988398), whose 2023 990 shows $75M revenue and $65.8M expenses across all projects. LTFF's own payout reports indicate spending of $5-12M/year in recent years.

Funding sources: Historically ~40% from Open Philanthropy, ~50% from individual donors, ~10% other institutional. OP has progressively distanced:

• Pre-2023: OP provided direct grants to LTFF grantmaking
• Aug 2023: OP shifted to 2:1 donation matching ($3.5M cap), which was fully filled by Dec 2023
• 2025: OP funds only EA Funds operations ($602K grant), not grantmaking

Business model: 100% of publicly donated money goes to grantees. Operational expenses covered separately, primarily by OP. The fund was considering moving to a model where donors directly pay operational expenses.

Key financial dynamics:

• Monthly spend target: $1M; actual capacity (Aug 2023): ~$550K/month
• Funding bar varies directly with available money (scored -5 to +5; threshold rose from 2.0 to 2.9 as funding tightened)
• FTX collapse forced LTFF to put bottom 40% of above-bar grants on hold
• The 2:1 OP match mobilized ~$1.1M from individual donors, suggesting the donor base ceiling is around $1M when motivated

Structural conflict: Until Aug 2023, both the LTFF chair and the EAIF chair simultaneously worked at OP. Claire Zabel (OP Senior Program Officer) sat on the EV board and regularly met with EA Funds leadership. Caleb Parikh acknowledged that "fund managers believed they 'should be weighing Open Phil's views more heavily' in funding decisions" than was warranted. The distancing reforms (Aug 2023) were real but personnel dependencies persist: most fund managers' primary employers receive OP funding.

Institutional transition: EA Funds and CEA merging (July 2025) and spinning out from Effective Ventures. A Jun-Jul 2025 grantmaking pause occurred during the transition. Post-merger, CEA pledges funding remains restricted per-fund, but the merged entity both runs programs and funds competitors to those programs.

What Others Say

The case against LTFF's independence: Nuno Sempere argued that EA funding structurally revolves around OP, with organizations chasing OP funding and individuals seeking work at OP-funded orgs. LTFF fits this pattern: ~70% of EA funding originates from OP, LTFF managers work at OP-funded orgs, and even after the Aug 2023 reforms, the personnel networks remain entangled. The "epistemic diversity" claim is weakened if the people providing that diversity are themselves embedded in OP's ecosystem.

Single-person evaluation: Multiple critics noted that most LTFF grants are investigated by a single fund manager with limited time. The "centralized grantmaking" critique argues that LTFF's managers share similar backgrounds (technical, Bay Area, rationalist community) and thus share blind spots. The policy funding gap is one documented example.

The transparency tradeoff: LTFF made public reports optional in 2021 and ~11% of funding went to non-public grants. Linch Zhang later expressed "update slightly against" anonymous grants. The fund remains more transparent than most funders but has retreated from its original standard.

Self-criticism from insiders: Bergal described the fund as shifting from "lively epistemic forum" to "solitary grantmaking machine." Linch acknowledged "grantmakers don't spend much time on any given grant." Habryka expressed reservations about LTFF's future trajectory. Fund managers publicly stated they cannot "provide strong assurances" their work hasn't been net-negative. This level of institutional self-doubt is unusual and arguably healthy.

The Beckstead failure: The fund's first 18 months under Nick Beckstead were essentially non-functional. $1M+ sat unallocated, promises of transparency were broken, and the fund marketed itself as supporting emerging organizations while the manager only funded existing ones he knew. This failure was publicly documented and led to complete management overhaul.

What's Absent

• No independent external evaluation of LTFF's grant portfolio exists
• No LTFF-specific financial statements (operates under EVF umbrella)
• No permanent full-time chair has been hired despite a 2023 job posting
• No Q1 2025 or later payout report has been published
• No public donor evaluations of LTFF effectiveness
• No systematic tracking of what happened to LTFF grantees after funding
• No public whistleblower policy or formal grantee complaint mechanism
• No analysis of grants' geographic distribution
• No comparison between LTFF outcomes and competing funders (Manifund, Lightspeed, SFF)

Recommended Reading

1. Asya Bergal, "Reflections on my time on the Long-Term Future Fund" (Aug 2023) -- The single most candid insider account. Bergal, as outgoing chair, describes scaling failures, structural weaknesses, the shift from "lively epistemic forum" to "solitary grantmaking machine," and her personal responsibility for bottlenecks. https://forum.effectivealtruism.org/posts/9vazTE4nTCEivYSC6/reflections-on-my-time-on-the-long-term-future-fund

2. "The EA Community and Long-Term Future Funds Lack Transparency and Accountability" (Jul 2018) -- The strongest critique, documenting 18 months of near-zero grantmaking under Beckstead. Essential for understanding why governance matters. https://forum.effectivealtruism.org/posts/wkLjwiGQ8fRdTDqou/the-ea-community-and-long-term-future-funds-lack

3. Linch Zhang, "Some unfun lessons I learned as a junior grantmaker" (May 2022) -- Honest about the tradeoffs and incentive dynamics of LTFF-style grantmaking. The most revealing look at how decisions actually get made. https://forum.effectivealtruism.org/posts/vPMo5dRrgubTQGj9g/some-unfun-lessons-i-learned-as-a-junior-grantmaker

4. "What Does a Marginal Grant at LTFF Look Like?" (Aug 2023) -- Shows exactly what LTFF would fund at $100K, $1M, $5M, and $10M budget levels. The most transparent funding bar disclosure in AI safety. https://forum.effectivealtruism.org/posts/7RrjXQhGgAJiDLWYR/what-does-a-marginal-grant-at-ltff-look-like-funding