Verent is a research lab and consumer product company working toward the capability to model and predict complex behavior at scale. Our end goal is hyper-realistic simulation that can render judgments about what happens next.
We believe AGI will arrive as a neuro-symbolic orchestra capable of predicting the next moments of the world at macro and micro scale, rather than as a hyper-advanced pure language model. Our engine is trained by working backwards from the beginning of recorded history to the present, and by building a universal base representation of a human from detailed conversations that we then specialize cheaply into agents.
Our first product is Pluto
Pluto is a CAD for social media. Influencers, brand agencies, startups, and enterprise teams submit a post (or K variations of a post), and Pluto runs large-scale simulations of millions of agents across each major platform to identify the strongest-performing version. Pluto autonomously discovers the right variant, the right time of day, and the right targeting for the campaign goal.
The simulation engine plugs into our reimplementations of platform ranking algorithms, with open-source approximations where the originals are closed. Results are grounded in how the real platforms actually behave. Pluto also surfaces the dispersion of information across the network, including the specific accounts and segments to reach for a given goal: a tech audience, people in a particular city, buyers of a category. The accuracy is well beyond what current tools deliver.
A sequence of products, named after planets
Pluto is the first product in a sequence. Each product takes the name of a planet, and each goes after a domain where people pay to know what happens next. Future products will move into finance, govtech, healthcare, defense, admin, and any field where prediction has clear economic value. We figure we will not run out of planets until simulation has taken over everything.
Research and product, side by side
A hybrid by design. Verent runs as a research lab and a consumer product company at the same time. The two halves feed each other. Research keeps the simulation engine grounded in real behavior. Product gives us the feedback signal and the data to make the engine sharper. Our team includes scientists shaping the engine and builders shipping Pluto this year.
Working backwards from history. We train the engine on human behavior as recorded across history, then bring it forward to the present. Patterns of attention, persuasion, organization, and panic repeat. Capturing those patterns at the scale of recorded history gives the engine a base it can specialize from.
A universal base representation of a human. From detailed specific conversations, we build a representation of a person that generalizes across contexts. Cheap specialization on top of that base produces agents that act like real users in real platforms, which is what makes the simulations accurate enough to trust.
Why simulation
Most of the value created by AI in the next decade will come from answering one question well: what happens next. Markets price it, governments plan around it, founders bet on it, brands buy attention based on it. The current state of the art is a mix of polling, dashboards, and gut. We think a faithful simulation of the relevant system, run cheaply and at scale, is the better answer in most domains.
Pluto is the first place we are proving that out. Social media is a domain with fast feedback, public ground truth, and obvious economic value. The engine we build to win there is the same engine that goes after the rest.
Join us
We are building the prediction layer for the next decade of software. If you want to help model the world at scale, in any of research, infrastructure, applied modeling, or product, we want to hear from you.
Follow us on X at @verent for updates, or see our open roles below.