01 / 15
MesoCort
๐Ÿง 
The Limbic System for AGI
โ† Move or tap to see human unpredictability โ†’
$10M
Seed Round
$50M
Pre-Money Valuation
Scroll
The Problem
AGI Is Building
Only Half a Brain

Daniel Kahneman's Nobel Prize-winning research identified two systems in the human mind โ€” fast intuition and slow reasoning. Every major AI lab is building System 2. No one is building System 1.

System 2 โ€” Slow Thinking
โฑ 500ms+

The Neocortex

Logic, language, planning, and deliberate reasoning.

Being built by
OpenAI ยท Anthropic ยท Google ยท Meta
VS
System 1 โ€” Fast Thinking
โšก 50โ€“100ms

The Limbic System

Fight-or-flight, instinct, reflex. The irrational layer that determines survival.

Being built by
No one. Until now.

Without the limbic system, you cannot accurately simulate or predict human behavior. Every AI model of humans today captures only what people should do rationally, not what they actually do under pressure, fear, or instinct.

The Simulation Gap
Every Industry Simulates
Humans Wrong

The Missing Layer in Human Simulation

Today's simulations predict what a rational agent would do. But humans aren't rational. A pedestrian doesn't calculate an optimal evasion path โ€” they flinch, freeze, or bolt. A trader doesn't follow their thesis โ€” they panic-sell. A consumer doesn't weigh features โ€” they buy on impulse.

Every model that simulates human behavior without the limbic layer gets the critical moments wrong โ€” the exact moments that matter most.

We build the technology to make human simulation accurate for the first time โ€” by encoding the instinctive, irrational layer that drives real decisions.
๐Ÿค–
$45B

Autonomous Systems Safety

Predict irrational human reactions around robots, autonomous vehicles, and drones. Safety-critical applications need limbic-aware models.

๐Ÿ“Š
$28B

Marketing & Consumer Behavior

Model real purchase decisions โ€” impulse, emotion, social pressure โ€” not rational personas that don't reflect reality.

๐Ÿ›ก๏ธ
$24B

Defense Simulation & Training

Human-like adversaries and stress-aware operator models for military training systems and wargaming.

๐Ÿฅ
$18B

Healthcare & Clinical

Predict patient stress responses, treatment compliance, and cognitive overload in high-stakes clinical environments.

๐ŸŽ“
$12B

Education & Workforce

Detect engagement, burnout, and performance tipping points before they manifest as failure.

๐Ÿ’น
$9B

Financial Modeling

Simulate panic, euphoria, and herd behavior โ€” the limbic forces that move markets far more than fundamentals.

$136B+
Combined addressable market for accurate human behavior simulation
Science Team
World-Class Researchers
Building the Limbic AI

6 PhDs ยท Oxford Alumni ยท Top 2% World Scientists

KP
AI Strategy

Kannan Paul, PhD

PhD in AI from Oxford. Neuropsychology researcher. Strategic advisor to a16z, Sequoia, SoftBank.

Oxford AI PhD
IZ
AI Expert

Prof. Ivan Zelinka, PhD

Top 2% Scientists Worldwide (Stanford). Complex systems & AI research leader.

Top 2% Scientists Worldwide
RS
Co-Founder & Head of AI

Radek Svoboda, PhD

Architect of GPRT and HUBERT. Leads core AI model development.

Model Architect
AI Engineering & Research Team
JP
Jan Plucar, PhDVP Research
MS
Miloslav SzczypkaCo-Founder & CTO
PP
Petr Prokop, PhDAI Researcher
Jล 
Jakub ล afaล™รญk, PhDAI Researcher
DV
David VojรกฤekAI Engineer
LJ
Lukรกลก JochymekAI Engineer
ED
Emanuel DopaterAI Engineer
Jลฝ
Josef ลฝรกฤekAI Engineer
Leadership
The Team Executing
the Vision
MB
Co-Founder & CEO

Michael Blaลพรญk

Human performance expert in sports and esports. Former coach of the Korean floorball national team. Built Rankacy to 55K users. Drives fundraising, vision, and BD.

JJ
Co-Founder & Lead Investor

Jonathan Jackson

Lordship Group owner. BioTech co-founder (Terasom). Strategic capital and scaling.

JP
Co-Founder & VP Research

Jan Plucar, PhD

Cybersecurity expert. Defense and critical infrastructure applications.

MS
Co-Founder & CTO

Miloslav Szczypka

Full technical execution across GPRT, HUBERT, and Human AI Bot pipeline.

Research & Partnerships
Backed by Science,
Trusted by Defense

Charles University

Sports Science & Human Performance

Mapping digital behavioral signals to physiological metrics with biometric markers from elite athletes.

Masaryk University

Psychology & Behavioral Research

Validating AI-detected behavioral patterns against gold-standard psychological assessments.

Vล B Technical University

AI & Technology

Co-developing novel architectures for behavioral models on multi-modal datasets.

Active POC

OneArc by BAE Systems

Defense Simulation & Data Collection

Building a behavioral data collection pipeline and After Action Review system for VBS4 military simulations, used by 60+ NATO militaries.

Technology Pipeline
Building HUBERT โ€” the First
Limbic Behavior Model
Currently Training
๐Ÿง 

HUBERT

Behavioral State Detection & Prediction

The first AI model to predict human limbic behavior from digital data alone. Multi-horizon prediction that makes human simulation accurate for the first time.

Next-gen architecture ยท Multi-task heads ยท H1/H2/H3 prediction horizons
โ†‘
BUILDS ON
โ†‘
๐ŸŽฎ
Built

GPRT

Game-State AI Foundation

Full game-state understanding at granular tick-level detail.

75M params ยท 95% accuracy
๐Ÿค–
2027+

Human AI Bot

Limbic-Aware AI Agent

AI agents with full limbic behavior for simulation and real-world interaction.

Foundation model ยท Cross-domain
What We Detect
Five Behavioral States
Under Pressure

The invisible cognitive states that determine real human behavior โ€” detected in real time, from digital data alone.

๐ŸŒŠ

Flow

Peak focus. Optimal challenge-skill balance.

๐Ÿ”ฅ

Tilt

Emotional spiral. Frustration degrades decisions.

๐Ÿ’Ž

Clutch

Peak under max pressure. Rising to the moment.

๐ŸงŠ

Choke

Collapse despite ability. Overthinking kills instinct.

โšก

Panic

Fight-or-flight. Survival bypasses reason.

The Core Innovation
Predicting the
Fastest Human Reactions

H1 โ€” Micro-Intent Layer

The Foundation of Accurate Human Simulation

When a stimulus appears, humans react in under 100 milliseconds. This reaction is limbic, automatic, and deeply revealing.

By measuring this instant response, we extract behavioral signatures no other model captures. This is why current simulations fail at critical moments.

โšก 0 โ€“ 300ms reaction window
๐Ÿค–

Autonomous Systems

Predict the flinch, not just the trajectory.

stimulus โ†’ startle โ†’ trajectory change
๐Ÿฅ

Healthcare & Stress

Detect cognitive overload before errors occur.

stimulus โ†’ hesitation โ†’ error risk
๐ŸŽ“

Education & Training

Reveal disengagement before the student realizes it.

stimulus โ†’ response delay โ†’ engagement
๐Ÿ›ก๏ธ

Military & Defense

Micro-reactions reveal stress resilience and readiness.

stimulus โ†’ combat response โ†’ profiling
How It Works
From Raw Data to
Behavioral Intelligence
01
๐ŸŽฎ

Capture

Limbic Reaction Data

Competitive gaming captures millions of behavioral moments under real cognitive pressure โ€” no sensors required.

02
๐Ÿง 

Map

Ego-View Reality

GPRT reconstructs what each player could perceive, mapping stimulus to response at sub-100ms resolution.

03
โšก

Predict

Behavioral AI Model

HUBERT learns the causal link between context, instinct, and performance โ€” predicting what humans do before they do it.

Spot
Detect limbic states in real time
Predict
Forecast behavior before it occurs
Analyze
Map context โ†’ reaction patterns
Simulate
Generate human-like AI agents
Roadmap
From Gaming Data to
Real-World Behavior

A clear path from the richest behavioral dataset in gaming to MesoCort โ€” a universal human behavior prediction layer across all domains.

2025โ€“2026
Now

CS2 Foundation

Train HUBERT on the world's richest behavioral dataset โ€” Counter-Strike 2 competitive matches. 100M+ situations, sub-100ms resolution.

โœ“GPRT built (75M params, 95% accuracy)
โ—HUBERT training in progress
โ—BAE Systems OneArc POC
๐ŸŽฎ CS2 Data
2027
Next

Simulations & Bots

Expand beyond CS2 to military simulations, driving sims, and other virtual environments. Deploy AI bots that mimic model-free human behavior.

โ†’Multi-simulation behavioral data
โ†’Human AI Bot โ€” limbic-aware agents
โ†’Defense & autonomous systems pilots
๐Ÿ›ก๏ธ Simulations + ๐Ÿค– Bots
2028+
Vision

Real World

Cross the sim-to-real gap. Apply limbic behavior models to traditional sports, healthcare, education, and real-world environments with sensor and video data.

โ†’Traditional sports analytics
โ†’IRL environments (wearables, video)
โ†’MesoCort universal behavior API
โšฝ Sports + ๐ŸŒ IRL
Digital-First Data
Simulation Expansion
Real-World Transfer
The Data Moat
A Dataset No One
Can Replicate
100M+
Model-free behavioral situations
3,000
Years of continuous human activity
8M
Players in database

๐Ÿ”ฌ Why This Data Is Unique

Most datasets capture slow, deliberate choices. We measure sub-100ms stimulusโ†’response โ€” the limbic layer that separates incomplete models from accurate ones.

๐Ÿ“ก How We Collect

Rankacy: 55K+ users, 15K new files daily. GPRT (75M params) understands every granular detail โ€” a growing behavioral flywheel with no global equivalent.

Why Now
Three Converging Forces
01

Generative โ†’ Agentic AI

Robots and AVs entering the real world need to predict how humans actually behave, not how they should.

02

Safety as Regulation

Accurate behavioral prediction is becoming compliance-mandatory. The EU AI Act is just the beginning.

03

The Data Window

Our 3,000-year dataset grows daily. First-mover advantage compounds every day without competition.

The Ask
Seed Round
$10M
Seed Round
$50M
Pre-Money Valuation
$3.5M
Already Invested

Use of Funds

R&D & Science
45% ($4.5M)
Scale PhD team, biometric validation, advance HUBERT
Compute
20% ($2.0M)
Train HUBERT and Human AI Bot
Sales & Partnerships
20% ($2.0M)
Defense partnerships and BD
G&A
15% ($1.5M)
Operations, legal, finance
๐Ÿง 
The Invitation
Join us on this
moonshot mission.

You can't simulate humans without their limbic system. MesoCort is building the missing piece โ€” and we're looking for investors who see the future as clearly as we do.