AI Models That Simulate Human Thinking

What if we could simulate human thought—accurately, at scale, and without needing a single human? That’s no longer science fiction. A new foundation model called Centaur, just published in Nature, marks a major leap in cognitive AI. Trained on Psych-101, a dataset of over 10 million real behavioral choices from 60,000 participants across 160 psychological experiments, Centaur doesn’t just match human behavior—it predicts it better than traditional cognitive models. You can read more here: 🔗 https://lnkd.in/dyCN4rkp But this isn't just a technical milestone. It’s a signal. Why it matters now 1. Cognitive simulation becomes programmable Centaur allows us to run human-like experiments in silico. Want to test how people with anxiety respond to stress? Or how teens might react to social pressure? You can now do that virtually—no lab required. 2. A new era for social sciences Behavioral economics, psychology, education, UX testing—every field that studies how humans think and act can now prototype, validate and refine ideas at machine speed. 3. Foundation for future super-agents Centaur isn’t just performant—it’s brain-aligned. Its internal representations mirror neural activity better than any other model to date. That opens the door to agents that don’t just mimic human behavior, but actually understand it. 4. Interpretability meets generalization Where most large models are black boxes, Centaur blends predictive power with explainable mechanisms—critical for AI safety, governance and trust. My Key takeaways: General-purpose cognition models are emerging—and they're fast, scalable, and effective. Behavioral simulation is now part of the AI toolkit. Human-aligned agents are no longer theoretical—they’re arriving. The next generation of AI will think with us, not just for us. This post kicks off a summer series I’ll be publishing on the next generation of AI models, the rise of complex super-agents, and the transformational breakthroughs reshaping our field. Let’s get ready for what’s coming. #AI #CognitiveAI #SuperAgents #FoundationModels #HumanBehavior #SyntheticUsers #FutureOfAI

Yesterday, OpenAI released its latest model, O1, which marks a significant leap from chatbots to sophisticated reasoners. Named to indicate that it is resetting ‘the counter back to 1’, O1 pushes the industry beyond conversational AI to models capable of solving complex, multistep problems with near-human-level reasoning. 🧠 Reason: O1 excels at complex tasks. On a qualifying exam for the International Mathematics Olympiad, it correctly solved 83% of problems compared to GPT-4o’s 13%. In Codeforces programming contests, it ranks in the 89th percentile, with claims it will soon rival PhD students in physics, chemistry, and biology. 💸 Expensive: Quality comes at a cost. At 3-4x the cost of GPT-4o, O1-preview comes with a hefty price tag—$15 per million input tokens (vs $5), $60 per million output tokens (vs $15) 🐢 Slow: O1 mirrors human problem-solving, taking extra time to think through responses. This deliberate "thinking time" may slow performance but significantly enhances accuracy. 🚧 Limitations: O1 struggles with factual knowledge, lacks browsing capabilities, and can’t process images or files. While it reduces hallucinations, they’re not entirely gone. The cost-performance trade-off is real: not all tasks will need O1’s advanced reasoning, and its higher cost and latency might limit its applicability. But for use cases demanding deeper problem-solving, the potential is enormous. For those, the question isn’t how quickly AI can respond—it’s how deeply it can think. As one OpenAI researcher noted, O1 thinks in seconds—future models might think in hours, days, or even weeks, solving world-changing problems like curing cancer or innovating new battery technologies. It’s clear the future of AI isn’t about one model to rule them all. To each use-case, its own solution!

A new open-source Python library called TinyTroupe is here to redefine how we simulate human behavior using LLMs, advancing the field of AI agents. TinyTroupe allows you to create TinyPersons – simulated agents with unique personalities, goals, and interests – capable of interacting within custom TinyWorld environments. Unlike other LLM-based simulation approaches that focus on gaming, this library targets business scenarios, creating and interacting with AI-powered personas to test products, ads, and ideas before spending real money. Think running a focus group with AI-powered physicians, lawyers, or knowledge workers. The library enables diverse applications, from evaluating digital campaigns with simulated audiences and running AI-powered focus groups at scale to generating realistic test inputs for software, collecting requirements from specific personas, and creating domain-specific training datasets. This work could accelerate research in autonomous AI agents by providing a controlled environment to study agent-to-agent and human-to-agent interactions, such as in customer support and sales. Code and examples https://lnkd.in/g9TqYiVZ P.S. I've just open-sourced Voice Lab, a framework to evaluate LLM-powered agents across different models, prompts, and personas https://lnkd.in/gAaZ-tkA

Human conversation is interactive. As others speak you are thinking about what they are saying and identifying the best thread to continue the dialogue. Current LLMs wait for their interlocutor. Getting AI to think during interaction instead of only when prompted can generate more intuitive and engaging Humans + AI interaction and collaboration. Here are some of the key ideas in the paper "Interacting with Thoughtful AI" from a team at UCLA, including some interesting prototypes. 🧠 AI that continuously thinks enhances interaction. Unlike traditional AI, which waits for user input before responding, Thoughtful AI autonomously generates, refines, and shares its thought process during interactions. This enables real-time cognitive alignment, making AI feel more proactive and collaborative rather than just reactive. 🔄 Moving from turn-based to full-duplex AI. Traditional AI follows a rigid turn-taking model: users ask a question, AI responds, then it idles. Thoughtful AI introduces a full-duplex process where AI continuously thinks alongside the user, anticipating needs and evolving its responses dynamically. This shift allows AI to be more adaptive and context-aware. 🚀 AI can initiate actions, not just react. Instead of waiting for prompts, Thoughtful AI has an intrinsic drive to take initiative. It can anticipate user needs, generate ideas independently, and contribute proactively—similar to a human brainstorming partner. This makes AI more useful in tasks requiring ongoing creativity and planning. 🎨 A shared cognitive space between AI and users. Rather than isolated question-answer cycles, Thoughtful AI fosters a collaborative environment where AI and users iteratively build on each other’s ideas. This can manifest as interactive thought previews, real-time updates, or AI-generated annotations in digital workspaces. 💬 Example: Conversational AI with "inner thoughts." A prototype called Inner Thoughts lets AI internally generate and evaluate potential contributions before speaking. Instead of blindly responding, it decides when to engage based on conversational relevance, making AI interactions feel more natural and meaningful. 📝 Example: Interactive AI-generated thoughts. Another project, Interactive Thoughts, allows users to see and refine AI’s reasoning in real-time before a final response is given. This approach reduces miscommunication, enhances trust, and makes AI outputs more useful by aligning them with user intent earlier in the process. 🔮 A shift in human-AI collaboration. If AI continuously thinks and shares thoughts, it may reshape how humans approach problem-solving, creativity, and decision-making. Thoughtful AI could become a cognitive partner, rather than just an information provider, changing the way people work and interact with machines. More from the edge of Humans + AI collaboration and potential coming.

𝗧𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗔𝗜 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 𝗺𝗶𝗴𝗵𝘁 𝗹𝗼𝗼𝗸 𝗹𝗲𝘀𝘀 𝗹𝗶𝗸𝗲 𝗰𝗼𝗹𝗱 𝗰𝗮𝗹𝗰𝘂𝗹𝗮𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗺𝗼𝗿𝗲 𝗹𝗶𝗸𝗲 𝗶𝗻𝘁𝗲𝗿𝗻𝗮𝗹 𝗰𝗵𝗮𝗼𝘀. Not in a bad way. A new paper from Google and the University of Chicago shows that top-performing reasoning models like DeepSeek-R1 don’t just think harder, they simulate something more profound: 𝘈 𝘤𝘰𝘮𝘮𝘪𝘵𝘵𝘦𝘦 𝘰𝘧 𝘥𝘪𝘴𝘢𝘨𝘳𝘦𝘦𝘪𝘯𝘨 𝘦𝘹𝘱𝘦𝘳𝘵𝘴 𝘪𝘯𝘴𝘪𝘥𝘦 𝘵𝘩𝘦𝘮𝘴𝘦𝘭𝘷𝘦𝘴. Here’s what’s wild: When solving a problem, these models: - Question themselves ~7 times - Shift perspectives ~3.5 times - Engage in internal conflict ~3+ times 𝗮𝗹𝗹 𝗽𝗲𝗿 𝗽𝗿𝗼𝗯𝗹𝗲𝗺. Researchers discovered a neural "switch" they call the conversational surprise marker. Turning it up doubled performance, from 27% to 55% on hard reasoning tasks. Even when trained without prompts for dialogue or debate, models spontaneously evolved internal argumentation, because that’s what works. This behavior, what the authors call a "society of thought", mirrors how we humans reason best: by weighing diverse perspectives, entertaining doubt, and challenging our assumptions. 𝗧𝗵𝗲 𝗶𝗺𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻? Scaling reasoning isn’t just about bigger models. It’s about structured internal diversity, simulated voices with distinct expertise, personalities, and even emotional roles. For AI teams, this opens up new design frontiers: - Reward behaviors that mirror high-performing teams. - Think in terms of internal collaboration, not monologues. - Embrace the messiness of thought. 𝗧𝗵𝗲 𝘀𝗺𝗮𝗿𝘁𝗲𝘀𝘁 𝗺𝗼𝗱𝗲𝗹𝘀 𝗮𝗿𝗲𝗻'𝘁 𝗺𝗼𝗻𝗼𝗹𝗶𝘁𝗵𝗶𝗰 𝘁𝗵𝗶𝗻𝗸𝗲𝗿𝘀. 𝗧𝗵𝗲𝘆'𝗿𝗲 𝗺𝗶𝗻𝗶-𝘀𝗼𝗰𝗶𝗲𝘁𝗶𝗲𝘀 𝗶𝗻 𝗱𝗶𝗮𝗹𝗼𝗴𝘂𝗲. #AI #MachineLearning #Reasoning #LLMs #CognitiveScience #ArtificialIntelligence

Researchers recently unveiled Centaur: an AI model trained on 10M+ human choices across 160 psychology experiments. Unlike ChatGPT, it wasn’t fed internet text; it was trained on real experimental data (via the Psych-101 dataset) and fine-tuned on Meta’s Llama model (Nature; arXiv). The result? Centaur predicted human behavior in new experiments — outperforming classic cognitive models. So here’s the question: Are we inching toward a unified theory of cognition — or just building a more sophisticated mimic? For me, the real test isn’t whether AI can reproduce what we choose — but why we choose it: the messy, emotional, often irrational ways we think. That insight matters beyond psychology. It’s how we design better products, craft inclusive experiences, and honor the complexity of being human. What this could mean for industry research: - From description to prediction → Simulate choices before testing - New benchmarks → Validate outcomes against cognition, not just KPIs - Mixed-methods boost → Use AI as another “participant” to stress-test ideas and surface blind spots - Ethical guardrails → Ensuring predictive power is used responsibly,  empowering people rather than undermining their agency - Elevated researcher role → Guide model training, identify failures, and inject human context Could AI like Centaur augment industry research; or will there always be something irreducibly human only people can reveal? 👀 https://lnkd.in/g8b-vsvM https://lnkd.in/ghC6FiJK

Can you stop running A/B tests? MIT may have proved there's a better way. While marketers obsess over attribution, MIT cracked the real problem: prediction. They seem to have found new ways of simulating human behavior with startling accuracy. Here's what they did: Built AI agents grounded in behavioral psychology Trained them on small samples of real human data Tested them on 883,320 different scenarios The AI predicted human choices better than game theory, better than statistical models, sometimes even better than actual human data from similar studies The key insight?  When you ground AI in real behavioral science (think biases like loss aversion, social proof, etc) it doesn't just mimic human responses. It understands them. This is your own marketing holodeck. Right now, the testing bottleneck kills good ideas faster than the Terminator. You brainstorm 50 concepts, test 5, scale 1.  The other 49 die in committee because testing is expensive and slow. But what if you could simulate customer responses to all 50 concepts first? Test everything.  Scale only what works. MIT's agents reduced prediction errors by 53-73% compared to baseline models. In their largest test, the AI was 3.41 times more likely to predict what humans actually did. Imagine your next campaign planning:  Instead of arguing about which concept to test, you simulate all of them.  The winners are obvious.  Spend your budget only on ideas that already proved themselves in simulation. Think your creative lacks creativity? When testing costs drop to near zero, you can afford to test weird stuff.  The campaigns that either bomb spectacularly or become legendary. Right now, playing it safe is like choosing vanilla at Baskin-Robbins—boring and forgettable. Brands that build this capability first won't just be more efficient. They'll out-create everyone. What's the wildest campaign idea sitting in your "too risky" folder?

𝐖𝐡𝐚𝐭 𝐢𝐟 𝐀𝐈 𝐜𝐨𝐮𝐥𝐝 𝐭𝐡𝐢𝐧𝐤 𝐥𝐢𝐤𝐞 𝐮𝐬? 𝐄𝐧𝐭𝐞𝐫 𝐂𝐞𝐧𝐭𝐚𝐮𝐫. Researchers fine-tuned a cutting-edge language model on 10 million human trial responses from 160 𝐜𝐥𝐚𝐬𝐬𝐢𝐜 𝐩𝐬𝐲𝐜𝐡𝐨𝐥𝐨𝐠𝐲 𝐞𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐬, creating a “foundation model of cognition”. 🔥 𝐓𝐡𝐞 𝐖𝐎𝐖 𝐦𝐨𝐦𝐞𝐧𝐭 𝐟𝐨𝐫 𝐦𝐞: Centaur doesn’t just guess what people will do—it predicts real human behavior better than traditional, handcrafted cognitive models and even the original LLM. It generalizes seamlessly—to new task structures, cover stories (bandits, space missions, magic carpets), and domains it’s never seen—all while mirroring population-level psychological variability.   This isn’t just data science—it’s a leap toward modeling the dynamics of human cognition: sequential decisions, exploration strategies, individual differences, even neural signatures align more closely with people (after fine-tuning!). 𝐖𝐡𝐲 𝐢𝐭 𝐦𝐚𝐭𝐭𝐞𝐫𝐬? Centaur claims to transform how we prototype user experiences, cognitive therapies, or business decisions: e.g. run human-like simulations in silico before real trials. You’re not guessing personas—you’re simulating real thought patterns. Wondering how this can impact travel? Let’s imagine how this could power better UX, smarter AI assistants, or even next-gen market research tools.   Shout out to Travel Tech Essentialist for covering this in #179    #traveltech #AI #cognition #personalization https://lnkd.in/eNXAaJaC

What if one AI model could predict how people think and behave across hundreds of different psychological tasks? That’s essentially what a new study just published in Nature has achieved. The researchers trained a large language model, fine-tuned on over 10 million choices from 160 behavioral experiments. The result: Centaur, a foundation model of human cognition. It outperforms classic cognitive models, generalizes to new tasks it’s never seen before, and even aligns more closely with patterns in brain activity. This is a big deal. It moves us closer to a more unified model of how humans make decisions — and opens up new possibilities for psychology, AI, and behavioral science. Well worth a read. https://lnkd.in/gWnRZvUd