Former OpenAI researcher Daniel Kokotajlo's nonprofit wants to slow the AI race down, and it has a surprisingly detailed plan for how to do it
The guy who quit OpenAI over safety concerns now has a plan to keep superintelligent AI from arriving too soon. Daniel Kokotajlo’s AI Futures Project just dropped “AI 2040: Plan A,” a report that argues the world’s best move is to deliberately pump the brakes on artificial intelligence development and push the arrival of superintelligence out to 2040.
The core pitch: get the US and China to agree, by 2029, on a framework for full transparency in AI research and development, backed by verification protocols and what the report calls “mutually assured compute destruction.” Think nuclear deterrence logic, but for data centers.
From doom forecast to deliberate optimism #
The AI Futures Project previously published “AI 2027,” a forecast that painted a faster and considerably more alarming trajectory for AI progress. “AI 2040: Plan A” takes the opposite approach, mapping out what could go right if the major players choose cooperation over competition.
The report proposes a phased development plan. Rather than racing toward the most powerful systems as fast as possible, it envisions a gradual scaling of AI capabilities, with expert-level systems arriving around 2035. The final stretch to superintelligence would come five years later, in 2040, with enough safety infrastructure in place to manage it.
The authors behind the plan include Kokotajlo, Thomas Larsen, Eli Lifland, Romeo Dean, Brendan Halstead, and Ryan Greenblatt.
Why Kokotajlo matters in this conversation #
Kokotajlo resigned from OpenAI in April 2024, publicly citing safety concerns about the company’s direction. He founded the AI Futures Project in 2025, positioning it as an independent organization focused on AI forecasting and policy research. The nonprofit runs on donations and grants, deliberately keeping itself outside the orbit of any major AI company.
The “AI 2040” report draws on insights from company experts and policymakers, though the AI Futures Project has been clear that the plan represents its own vision rather than any consensus among AI developers. So far, the report has generated significant discussion on platforms like LessWrong and X, though mainstream media coverage has been limited.
What this means for investors and the broader market #
The emphasis on transparency, verification protocols, and standardized safety benchmarks aligns with the direction that both US and European regulators have been moving. Reports like this one give policymakers intellectual ammunition and concrete frameworks to point to when crafting legislation.
The “mutually assured compute destruction” concept is particularly interesting from a hardware perspective. If any version of that idea gained traction in policy circles, it could reshape the economics of large-scale compute infrastructure, potentially affecting the valuations of both chip manufacturers and cloud providers.
The counterargument is that slowing down AI development unilaterally, without genuine buy-in from China, simply hands the lead to whoever doesn’t cooperate. That’s the fundamental tension the report acknowledges but doesn’t fully resolve.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our