When Mira Murati, former OpenAI CTO, unveiled Thinking Machines Lab in February 2025, tech headlines blazed with excitement. Her $2 billion funding target and vision for human-centric, widely understood AI systems captivated investors and tech enthusiasts alike. Yet behind this splashy debut lies an overlooked reality: Solsten has already built this future.
For six years, Solsten has operated as the quiet architect of psychology-driven AI that genuinely understands human motivation. While Murati’s team begins their journey, Solsten’s technology already powers experiences for over 500 million users through partnerships with giants like EA, Peloton, and DraftKings. The race Thinking Machines Lab just entered has a frontrunner they never mentioned.
The Psychology Layer Missing From Today’s AI
Most AI systems process what people do but remain blind to why they do it. They track clicks, analyze text, and generate content without grasping the psychological drivers behind human behavior. Such a fundamental gap explains why even advanced models often feel mechanical and miss the mark on truly human interaction.
Solsten tackles this problem at its root by measuring over 300 psychological traits, including personality, values, and intrinsic motivations. Using patented clustering technology and AI-powered assessments, they’ve built what CEO Joe Schaeppi calls “the emotional DNA layer for AI systems.”
“Surface-level insights won’t cut it anymore,” notes Schaeppi. “If you don’t understand the deeper motivations behind your audience’s actions, you’re essentially flying blind.” This psychological foundation, not raw computing power, marks the true next frontier in human-compatible AI.
The company didn’t stumble upon this concept recently. Their Navigator platform has been developed from millions of psychometric profiles and billions of behavioral data points gathered over six years.
From Patents to Practice: The Quiet Giant’s Head Start
Thinking Machines Lab promises AI systems that adapt to human knowledge and enable broader applications. Meanwhile, Solsten holds multiple patents covering precisely these capabilities – systems that correlate user behavior with psychological attributes, methods to adapt digital environments based on user psychology, and frameworks determining content presentation based on interaction patterns.
More than mere theoretical concepts, the technology is now used across the gaming, fitness, and entertainment industries. Bastian Bergmann, Solsten’s co-founder, explains: “We’re giving companies the ability to understand their customers as human beings, not just numbers on a spreadsheet. This is about creating experiences that truly connect.”
SolstenChat: Tomorrow’s Vision Available Today
While Thinking Machines Lab sketches plans for more comprehensible AI, SolstenChat already exists as a portal into the psychological motivations of any audience worldwide. This platform allows businesses to virtually “interview” consumer groups, test concepts, and develop AI agents with specific personality traits.
What makes SolstenChat so special isn’t just its ability to simulate audience responses. Its true power comes from grounding those simulations in validated psychological science – the NEO-PI-3, Achievement Motivation Inventory, and other established frameworks that have mapped human motivation for decades. When combined with behavioral data from over 20,000 brands and 40,000 interest groups, the result outdoes what even $2 billion of fresh funding could quickly build.
Marketers using the platform gain immediate insights that once required weeks of focus groups and surveys. Product teams can test ideas against psychological profiles before writing a single line of code. Game developers craft NPCs that adapt to player psychology in real-time, creating experiences that feel genuinely alive.
The contrast couldn’t be clearer: while new entrants chase visions of AI that understands people, Solsten quietly built it years ago. And in the race to make artificial intelligence more human, that head start might make all the difference.
