Dwnrt Absrd - FAQ

Organizations make important decisions in complex systems such as supply chains, public health, sustainability, operations, and resource planning.

Today they face a difficult tradeoff.

Traditional simulation models and digital twins can produce reliable insights, but they require specialized expertise to build, maintain, and interpret. Most organizations lack the time, budget, or expertise needed to use them effectively.

Generative AI is easier to use and more accessible, but it can confidently produce incorrect answers and has limited grounding in how real-world systems behave.

SupaSim combines the strengths of both approaches. We make proven simulation and digital twin methodologies accessible through natural language interfaces, allowing users to ask questions and receive answers grounded in models rather than guesses.

SupaSim combines AI agents with digital twins of an organisation (DTO).

Users interact with an AI agent through natural language. The agent helps define scenarios, runs simulations against a digital twin, and interprets the results.

Rather than relying solely on statistical patterns learned from text, the agent grounds its responses in the behavior of a mathematical model representing the system being studied.

The agent also understands the limits of the model. When a question falls outside the model's scope, the agent can identify gaps and assist with extending the digital twin, with humans remaining in the loop where validation is required.

SupaSim is an AI-powered platform that enables users to build, interact with, and learn from digital twins of their organisation (DTOs) through natural language.

Users can describe a system, ask questions, explore scenarios, and receive answers grounded in mathematical models rather than purely statistical predictions. The platform combines AI agents with proven simulation and digital twin methodologies to make sophisticated decision-support tools accessible to a much wider audience.

SupaSim is also being built using a Forward Deployed Engineering (FDE) model. Rather than building the product in isolation, the company works closely with design partners to solve real-world problems and rapidly customize solutions for specific domains. Capabilities developed for one deployment are generalized, productized, and incorporated into the platform.

This approach allows SupaSim to combine the customer intimacy and learning velocity of services with the scalability of software. Each deployment improves the platform's modelling, simulation, and AI capabilities, creating compounding value for both customers and the company.

Over time, SupaSim aims to become the easiest way for organizations to create, maintain, and use digital twins for decision-making.

LLMs have reached a point where, with purpose-built tooling and expert guidance, they can help create, maintain, and operate valid digital twins from messy real-world context.

At the same time, base model capabilities appear to be converging. Increasing differentiation comes from orchestration, memory, tooling, domain expertise, and access to grounded reasoning systems.

Digital twins and simulation have been proven decision-support tools for decades. AI now provides an opportunity to make them dramatically more accessible, reducing the expertise required to build and use them.

Immediate Vision (0–3 Years)

Build AI-powered digital twin tools that make simulation-based decision support accessible to organizations that currently lack specialist modelling expertise.

Medium-Term Vision (3–7 Years)

Create agents capable of constructing, maintaining, and improving organizational digital twins while preserving institutional knowledge and supporting increasingly complex decision-making workflows.

Long-Term Vision (7–10+ Years)

Develop AI systems capable of building and evolving valid abstract world models through interaction with people, organizations, and eventually the broader world.

Growth-oriented organizations that regularly make strategic, operational, and resource-allocation decisions but lack digital twins of their organisation (DTOs) and dedicated modelling and simulation teams.

Initial opportunities include consulting firms, sustainability practitioners, operational planning teams, and digitally mature SMEs seeking decision-support capabilities that would otherwise be too expensive or complex to implement.

The beachhead market is estimated at approximately $2B.

The company's unique capability is generating, maintaining, and operating valid digital twins of complex systems.

These models powering digital twins can take many forms, including causal graphs, numerical models and statistical forecasting. They provide the grounding that enables AI agents to reason about systems, explore scenarios, explain outcomes, and support decision-making.

Validity of a digital twin is not binary. A model is only useful to the extent that it is fit for its intended purpose. Generating a model is relatively easy. Generating one that is valid and knowing when it is valid are substantially harder problems.

UX Flywheel

As users interact with SupaSim, the agent builds and refines user-specific digital twins and models. These models make future interactions increasingly useful and context-aware.

Data Flywheel

Generalized models distilled from many domain-specific digital twins become a source of unique, grounded synthetic data about how complex systems behave.

Model Flywheel

That synthetic data helps improve future model-generation systems, enabling SupaSim to create better digital twins faster and with less human intervention.

SupaSim is being built using a Forward Deployed Engineering (FDE) model.

We work closely with a small number of design partners to solve high-value decision-making problems using AI-powered digital twins. These engagements help validate customer needs, generate revenue, and accelerate product development.

Rather than treating each engagement as bespoke consulting, we identify common patterns, workflows, models, and tooling that can be generalized into the platform.

Over time, the proportion of reusable platform capabilities increases while the cost of customization decreases, allowing the business to scale beyond services alone.

Going too fast and going too slow.

Partnering exclusively with customers who prioritize short-term ROI can pull the company toward local optima rather than the broader long-term vision.

At the same time, failing to move quickly enough toward increasingly autonomous model generation and maintenance may allow competitors to close the window of opportunity.

They could and might.

However, building useful digital twins requires more than larger foundation models. It requires deep expertise in modelling methodology, validation, simulation, human-computer interaction, and decision-support workflows.

Historically, many transformative products have emerged not from frontier model providers themselves but from teams that combine frontier models with deep domain expertise and specialized workflows.

The founder combines:

• systems modelling expertise,
• consulting workflow experience,
• interface and product design,
• software development,
• and AI experimentation.

Over more than a decade, the founder operated a consulting and technology company focused on making complex models more usable for non-experts by building bespoke interactive software.

The founder has worked across domains including:

• healthcare,
• ecology,
• tourism,
• public health,
• organizational systems,
• mining,
• and sustainability.

The current venture represents a continuation and evolution of the founder's long-standing mission: making systems reasoning more usable, scalable, and operationally effective.

Dwnrt Absrd's parent company already possesses:

• 12+ years of systems modelling experience,
• an accumulated modelling and interface technology stack,
• simulation engines,
• reusable modelling templates,
• visualization systems,
• deployment infrastructure,
• Think in Cycles,
• and early SupaSim prototypes.

Relevant IP will be transferred and/or licensed to the company as appropriate.

Self-honesty, curiosity, usefulness, delight.

Build AI tools that help people make better decisions.

The world is increasingly complex, yet many important decisions are still made with incomplete information, limited analysis, and poor visibility into long-term consequences.

Better decision-making can improve the performance of organizations, the effectiveness of public policy, the allocation of resources, and the stewardship of social and ecological systems.

We believe AI should help people understand complex systems and explore consequences before acting, enabling better decisions at every scale.

AI systems capable of building and evolving valid abstract world models.

We believe intelligence requires more than pattern matching. It requires the ability to form, maintain, and refine representations of how the world works.

Our long-term vision is to develop AI systems that can build and evolve valid abstract world models through interaction with people, organizations, and eventually the broader world.

A digital twin is a computational model of a real-world system.

Depending on the use case, the twin may represent an organization, a supply chain, a population, a business process, an ecosystem, or another complex system.

Users can ask "what if" questions and explore scenarios through the twin before taking action in the real world.

AI excels at understanding language, summarizing information, and interacting with users.

Simulation excels at representing how systems behave over time.

By combining both approaches, SupaSim can provide answers that are both easy to access and grounded in explicit assumptions about how the system works.

SupaSim is designed to understand the scope and assumptions of the digital twin it is operating.

When a question falls outside the model's validity range, the agent can identify that limitation, explain it to the user, and recommend how the model might be extended or refined.

This is fundamentally different from systems that attempt to answer every question regardless of whether they have a reliable basis for doing so.