01Language
Model meaning
Large language models learned powerful statistical representations of words, concepts, and the relationships between them.
A Life Design Technologies thesis
AI is learning the world around us.It also needs to understand the world within us.
World models help machines predict and act in physical environments. An internal world model is a permissioned, correctable working picture of one person’s lived experience. Its purpose is to help that person understand their life and make more aligned choices—not to define the life they should want.
ONE PERSONInternal
worldPermissioned · evolving · corrigible
ContextStateValuesChoicesOutcomesPurposeworldPermissioned · evolving · corrigible
SubjectiveLongitudinalWhole-lifeHuman-directed
The observation that changed our work
People often know what matters to them. Knowing is not always enough to change.
For years, Jeff and Jaqueline Lerner worked with people trying to create more freedom, meaning, stability, and possibility. They heard thousands of people describe the lives they wanted—and, just as important, why those lives mattered.
They also watched many of those same people struggle to act consistently on things they genuinely cared about. The pattern was too common to explain as laziness, lack of intelligence, or lack of desire.
Needs, beliefs, identity, nervous-system state, habits, relationships, history, resources, and present circumstances all shape what a person can see and do. Advice given without that context will often miss the real constraint.
We came to believe that lasting change begins with understanding: not another system telling people what to do, but a clearer picture of the particular life in which change has to happen.
The intelligence frontier
We are teaching machines how the outside world works.
AI’s representation of reality keeps expanding—from language, to perception, to physical environments that can be simulated, predicted, and acted within.
That work can help robots take on labor that is dull, dirty, or dangerous. It can expand what people are able to build, discover, and protect.
It is essential progress. It is also only half of the human augmentation problem.
02Sight + sound
Model perception
Multimodal systems are learning the visual and sonic patterns that let machines interpret more of what people can see and hear.
03Physical reality
Model the world
World models aim to predict how three-dimensional environments change so agents and robots can plan, act, and learn safely.
The missing model
A map of physical reality cannot tell us what makes one life worth living.
AI can optimize a route, a factory, a supply chain, or a robot’s next movement. But helping a person navigate grief, ambition, family, work, identity, tradeoffs, meaning, or purpose requires a different kind of representation.
It requires a model of lived experience: how this person understands their situation, what they value, what their nervous system and attention can hold, what relationships and responsibilities shape their choices, and what a fulfilling life means to them.
Without an internal world model, human-centered AI risks optimizing everyone toward the same externally legible version of a good life.
Working definition
A working picture of one person’s world within.
An internal world model is a partial, consent-based, and correctable representation that helps AI relate a person’s present moment to the wider system of their life. It is never the person, and it should never be treated as complete.
Lived context
What is happening around me?The roles, relationships, responsibilities, resources, constraints, and circumstances that shape a person’s real options.
Experienced state
What is happening within me?What the person reports about attention, emotion, energy, stress, needs, beliefs, and capacity—held with uncertainty, not treated as objective fact.
Personal direction
What matters, and where am I going?Values, goals, commitments, meaning, purpose, and the person’s own definition of a life worth living.
Choice + consequence
What happened when I acted?Decisions, actions, outcomes, and feedback that help the model learn what actually supports this person in this context.
Whole-life relationships
How does one part change the rest?How work, money, health, family, identity, contribution, and purpose affect each other across time.
What it is not
Not mind readingNot a diagnosisNot a permanent personality scoreNot a digital owner of the selfNot surveillance with a friendlier nameNot a machine-defined purpose
Conceptual operating loop
Understanding before steering.
The model earns usefulness through an ongoing loop of permission, representation, reasoning, action, correction, and outcomes. The person remains both the source of authority and the judge of fit.
Receive
Begin with what the person deliberately shares and permits the system to use.
Represent
Build a connected, uncertainty-aware picture instead of reducing the person to a profile or score.
Reason
Relate the current moment to the person’s wider life, direction, capacity, and available choices.
Support
Offer explanations, options, plans, and tools that fit—without taking ownership of the decision.
Learn
Let correction, action, and outcomes improve the model while preserving history and user control.
Design requirements
The architecture only works if human agency is a system requirement.
Permission before inference
A useful model should begin with context a person knowingly shares, not hidden surveillance.
Agency before optimization
The objective is not compliance or engagement. The person defines what better means and keeps the final choice.
Pluralism before one best life
There is no universal template for fulfillment. The architecture must preserve individual values, cultures, relationships, and paths.
Correction before certainty
A person must be able to inspect, challenge, update, and delete what the system thinks it knows.
Continuity with boundaries
Long-term context can make AI more useful, but memory should be scoped, legible, and revocable.
Human support where it matters
Software should recognize when qualified professionals, trusted relationships, or urgent services are the right next layer of help.
From thesis to product
The internal world model is the intelligence foundation beneath our work.
Life Design Technologies is testing and developing this architecture through products that help people understand their lives, make decisions, follow through, and reach human support when it is useful.
Personal intelligence
LD
The relational intelligence that uses an internal-world representation to understand the person behind the prompt and help them think, choose, and follow through.
See how LD worksLife Design application
Life Design OS
The software environment where context, memory, planning, Life Layers, and specialized tools can adapt around one person’s changing life.
Explore Life Design OSExperience architecture
Adaptive Software
A new software model in which the experience learns the person—with permission—and adjusts its guidance, tools, organization, and timing around what they actually need.
Explore Adaptive SoftwareExpanded access
The 4th P
The nonprofit expanding access to Life Design technology, programs, and support for people whose financial or social circumstances might otherwise put them out of reach.
Explore the nonprofit missionResearch + product company
Life Design Technologies
The company developing the architecture, adaptive software systems, and product ecosystem that turn this thesis into practical support.
Explore the technologyWhy access matters
AI should not make fulfillment another advantage reserved for the wealthy.
People with more money often have more access to time, support, information, networks, and options. AI can widen that divide—or help narrow it.
The 4th P exists to expand access to Life Design technology, programs, and community for people who might otherwise never receive them.
Explore the nonprofit missionAn open research agenda
This is a direction, not a solved problem.
The work spans world models, predictive processing, interoception, subjective wellbeing, memory, human-computer interaction, privacy, safety, and personalized decision support. We separate established research, product hypotheses, and long-term ambitions.
Read the research mapPut the model to work
Build a more fulfilling life with intelligence that learns yours.
Life Design OS is the first practical application of the Internal World Model: personal intelligence designed to understand your context, connect the parts of your life, and help you act on what matters.