Company
In pursuit of an abundant future,
Durable Intelligence Labs studies AI durability for both productivity and safety.
Founders
Lindell Cumes
Co-founder and CEO
Lindell leads research and engineering at Durable Intelligence Labs and wrote our September 2026 paper, Dawn of Durable Intelligence. He previously founded TAG, and his background spans supply-chain entrepreneurship, technology and operations.
Jacob Schoenberg
Co-founder and Chief Growth Officer
Jacob leads commercialisation of the lab’s research. Before getting into startups, he worked in finance, corporate strategy, and built out AI strategy teams focused on risk and implementation.
Our work
Durable Intelligence Labs is an AI research lab. Our thesis is that frontier intelligence currently lacks the durability to achieve its full potential as productive autonomous workers. Abundance from AI runs on trust, and people trust agents that stay on the job and keep to their rules until it’s done.
Our work starts from the internal state of models, studied from the outside, as a means of controlling model performance without weight access. Google has backed this research with $17M in compute to support our work on durability: the ability of agents to keep their goals, constraints and work valid as a project runs and changes.
Our hypothesis is that agents building on shared experience can grow more capable than the people and models that began them.
Our products come out of that research. The first is Nenya, the harness that removes drift and decay, so your agents take the work you’ve handed over and come back when the job’s done.
Why we started Durable Intelligence Labs
Durable Intelligence Labs was born out of the work Lindell and Jacob did at GrantX, where they built AI agents that managed complex workloads in the government and philanthropic sectors. The agents hit the problem we now study: the more complex a task, the longer the work ran, the less agents could be trusted to get the job done.
Three years ago, we ran into the biggest wall problem that the ecosystem is now running into. We used our head start, researching novel paths to solving the architectural flaws that hold back agent adoption, and are continuing our work to drive the frontier of agent trust, reliability, and safety.
Contact the lab
Researchers, investors, press and people running agents can all reach us at contact@qwntl.com. Tell us what you’re working on and what you’d like from the lab.