Predicts what people will do next, modelled on human psychology
The Hook
“Everyone's worried about AI going rogue and ultimately causing the downfall of society.”
Problem Agitation
“But the problem we already live with is much more real. AI doesn't actually understand us. Humans read each other with empathy and intuition. It's how we catch sarcasm, hesitation, subtle shifts in tone. That's what AI has been missing until now.”
Problem Agitation
“Long before the ChatGPT moment, I started interviewing people to understand what makes them tick. Together with 50 psychologists, we interviewed more than 7,000 people from all walks of life. They opened up about their dreams, their wishes, their most personal experiences. We returned over four years to compare what people said with how they behaved. Our unique origin story gave us something no one else had: Behavioral data comparing everyday people's words with their actions. That's the foundation of Thorsten-4.”
Product Reveal
“Thorsten is a Large Psychology Model. LLMs predict the next word, Thorsten predicts the next move. It does that by working more like a brain. Thorsten looks at word choice, sentence structure, and how those patterns change, mapping them to underlying behavioral science. It connects signals that might seem unrelated, much like an expert psychologist would. And every connection is grounded in an established psychology framework, with predictions checked against the words and actions of thousands of real people.”
Feature Teaser
“AI is becoming part of every conversation where important decisions get made. Earnings calls, interviews, sales meetings. Thorsten is one intelligence that can power every one of them. And our partners are already building on it. A hedge fund tested it on earnings calls against sentiment tools and frontier AI. Thorsten read the executives behind the numbers and the false confidence. This year, the fund began trading a $50 million portfolio on it. Top executive search firms have put Thorsten's reads to the test. Now we're building an assessment tool with one of them to help their clients avoid costly hiring mistakes.”
Call to Action
“Imagine AI that understands the person it's talking to. What they mean, how they feel, and what they'll do next. That's the type of intelligence people actually want. And there's no amount of compute that can buy it. AI knows how to talk to us. Now it will know who it's talking to.”
Synthesizable in Remotion · key components of the cognitive-founder-explainer template.
Biometric ocular overlay splitting into four quadrant memory metrics with numeric dial updates.
Compose an absolute HUD container inside an eye-iris mask. Animate SVG quadrant axes and interpolate text counters using Remotion spring physics.
Comparative evaluation bars showing slider points for personality candidate ratings.
Render track divs with spring-driven width animations. Bind candidate profile icons to slider indicator positions with subtle easing.
Text message bubble highlighting emotional discrepancy with a predictive takeaway badge.
Use staggered opacity and translateY springs on quote boxes, sequentially revealing highlighted syntax and an underlying conclusion card.
Purpose. Disarm viewers by poking fun at AI apocalypse clichés.
Execution. 8-bit military console interrupted by cute cartoon mascot.
Purpose. Establish scientific authority through proprietary longitudinal data.
Execution. Warm library studio setting paired with clean kinetic statistical lists.
Purpose. Differentiate LPM from commodity LLM next-token predictors.
Execution. Ocular biometric scan mapping communication patterns into psychological traits.
Purpose. Validate market viability with high-stakes finance and recruitment cases.
Execution. Clean UI card animations showing $50M portfolio and leadership scoring.
Purpose. Leave an indelible thesis on the future of human-AI connection.
Execution. Founder final statement resolving into 3D hexagonal logo reveal.