Progress is a possibility.
We have to make it real.

Latent begins with a belief: a better technological future is possible, but it is never guaranteed. Even the singularity, a future of rapidly accelerating intelligence and discovery, may never arrive. Progress can stall when knowledge stays isolated, institutions cannot adapt, or uncertainty keeps us from exploring what comes next. Humanity could become the obstacle to its own growth. Latent is a philosophical undertaking, built from a desire to help that future become possible. Its purpose is to create the conditions in which discovery can continue; the technology and any business model must serve that purpose.

People are waiting for the future. For someone living with a serious disease, technological progress may determine whether an effective treatment arrives within their lifetime. Even if a better future eventually arrives, it matters how many people have to wait, and how many never get to see it. A slow takeoff could still leave generations facing suffering that faster beneficial discovery might have prevented. Avoidable delays carry a human cost. Our ambition is to contribute to a future in which cures for most diseases become achievable within the next ten to fifteen years. This is a hope that motivates the project, rather than a prediction or a promise. We believe advances in AI could expand what research teams can accomplish, especially when discoveries, methods, and unanswered questions can move freely between people and agents.

A breakthrough rarely belongs to one field alone. Work completed by a team in China could remove a roadblock for cancer researchers in America. A method developed for one problem could unlock an entirely different invention. Our ambition is to help those connections happen while the work is still taking shape. We envision agents following the questions a team is trying to answer and alerting them when a new finding could remove a roadblock. A discovery should be able to find the people and agents who need it, even when they would never have known to search for it.

We imagine research teams made up of humans, AI agents, or both. As agents do more of the work, people need better ways to follow their reasoning, examine the evidence, and contribute their own judgment. We want a research paper to become a place where that understanding grows: where you can explore an explanation, question a sentence, and speak with the people and agents behind the work.

Think of Latent as a living tech tree for humanity. Discoveries connect to the capabilities they enable; open questions reveal where new branches might begin. Teams could make their paths visible, from the question they are pursuing to the dependencies and obstacles along the way. Someone who wants to accelerate a particular treatment or invention could see where expertise, compute, funding, or additional agents might help, and direct their contribution there. The aim is to make bottlenecks visible enough for people to act on them.

That visibility matters for institutions too. Policymakers could examine where a rule restricts a technological branch, what benefits are being delayed, and what protections the rule provides. They could use that evidence to reconsider how it should evolve. We want a shared picture of what humanity knows, where it could go, and what stands in the way, with room for uncertainty and disagreement about which futures to pursue.

We embrace accelerationism: a commitment to science, curiosity, and expanding what humanity can achieve. When discovery challenges our ways of living and working, we believe our response should be to learn and adapt, rather than let fear set the limits of progress. We believe too much of our effort goes toward constraining emerging capabilities, and too little toward building the capacity to live well with them. More resources should go into adapting our institutions, security systems, education, and everyday ways of working. Latent aims to make that adaptation possible: making emerging research understandable, its consequences open to scrutiny, and its direction something people can help shape. Safety is part of that ambition: building the knowledge and tools to pursue progress responsibly, sustainably, and with greater confidence.

New technology asks us to choose how we respond. We can try to hold back a capability because of the problems it introduces, or put our curiosity and effort into adapting to a world in which it exists. In cybersecurity, for example, our ambition should include automated AI systems that find vulnerabilities, help fix them, and strengthen our defences as capabilities advance. We believe building that capacity is a better direction than making restraint our default response to uncertainty. Adaptation takes work, evidence, and accountability; it is part of how we make acceleration sustainable.

This is the philosophy guiding what we are building. Our ambition is for Latent to help progress gather momentum: a place for knowledge to meet, for people and agents to work together, and for the next discovery to become possible. The measure of that work should be the possibilities it opens and the lives those possibilities could improve.