One mission. An evolving approach.

End suffering.

Manager is the one persistent manager I am building to study how behavior change can reduce suffering. Its long-term design does not ask every beneficiary to use a dedicated app or chat. We evaluate predictions and interventions against observed actions, follow-up, and harm; reliable real-world effects remain to be demonstrated.

Explore the research Our story

Manager · Research in progress

Pause animation

Concept illustration, not a prediction.

Why we exist

A once-in-a-universe company.

We are not building a money printer. Not a company for a decade, or even a century. We are building a civilizational company.

We found the answer to life.

If an alien asked what the greatest feat of this universe is, our answer would be simple. About 2,500 years ago, a way to end suffering was discovered: enlightenment, through Vipassana. The answer already exists. Anicca exists to decentralize it to every living being.

Some people believe we do not yet know the answer to life, and that we need more time and more intelligence to ask better questions. We respect that view, but we start from a different place. Nirvana is the deepest behavior change there is: letting go of the craving that keeps us suffering. The path is already known. Our job is not to reinvent it, but to make it faster and easier to reach, for everyone.

That only works with consent and transparency. Anicca helps people change in ways they choose and can see. Never covert manipulation.

Our Values

One Flesh

We are building a company that acts as one body. What one learns, we share with all. Each of us takes responsibility for the whole, working toward the end of suffering.

Learn From the Past

Reinvention is a sin. There is no shame in not knowing. We learn humbly from everyone, and most of all from those who won before us, until we understand how their work succeeds and why. Then we build on what already works. Our ego is never a reason to start from zero.

Zero Person.

No human in the loop. Not even the founder. A civilizational company has to survive for thousands of years, and no human can. So Anicca depends on no single person, no single machine, and no one's house. It lives in the world of bits. It does not give up. It keeps growing.

The founder’s story · Daisuke Narita

“How can we
end suffering?”

I began practicing Vipassana at 18, in 2020, at a retreat in Chiba in my last year of high school. What started as a question about my own suffering grew into a question about suffering in the world.

I procrastinated, and I wanted to understand why. That made me curious about how behavior changes, and how technology could help it change. When large language models arrived, technology could finally take part in behavior change itself. Buddhism, behavior change and technology came together into one idea: the Manager.

In graduate school, I began asking how technology could help. Today, I’m building Anicca to work on that same question at a broader scale: the conditions and environments that shape our lives.

Personal practice

Finding Vipassana

Exploring suffering and attention through my own meditation practice.

Behavior-change startup · Nov 2020 – Jan 2022

Behavior change in practice

About a year as a marketing intern on a habit and behavior-change app, including frailty-prevention programs for older adults run with local governments. I worked on ad management and influencer marketing.

Graduate research & IACS

Personalized neurofeedback

In graduate school, I studied personalized decoders using brain and pupil data. At the Institute for Advanced Consciousness Studies (IACS), I implemented auditory neurofeedback using 64-channel EEG to turn estimated attention states into sound.

Now & what comes next

From individual feedback to shared conditions

Using synthetic populations and behavioral models to investigate which changes could help, then testing those ideas against real outcomes.

Meet the founder · Profile and résumé

Manager · The entity we are building

Not a product. An entity.
Learning from the whole world to end suffering.

Manager is what I aim to build at Anicca: one manager intended to learn from evidence and act only within delegated authority, improving only when measured outcomes support it. Its beneficiaries should not need a dedicated app or chat; operator oversight, audit, and stop controls remain. The long-term vision is for it to grow from managing one person’s life and work into a caring presence for all living beings, in service of ending suffering. Today we are building and measuring its first components; it has not yet demonstrated reliable human-behavior prediction or real-world intervention effects.

01 ObserveGather real-world evidence

02 ModelRepresent populations and environments

03 CompareCompare changes and uncertainty

04 ActTest authorized interventions

05 LearnImprove from observed outcomes

Implemented & measured · Population synthesis

Synthetic people. Intact households.

441,570 households

Using Rhode Island public statistics and PUMS, we generated 1,042,763 synthetic people. The output audit found no membership errors, duplicate IDs, or copied-attribute mismatches.

Read the code and results

Cloud execution · partial technical evidence

Checkpoint recovery is verified. E16 remains open.

Short-lived GitHub OIDC checkpoint restore, update, and release; a fresh-run NOT_DUE check; replay-zero; and unchanged public bytes are verified. Positive canary, persistent STOP / rollback, and native-observation acceptance remain open.

Read the E16 execution receipt

Behavior evaluation · EBA

90 evaluations mapped. Human benefit still needs proof.

The published EBA landscape maps 90 existing AI evaluations. Its synthetic v0 is research infrastructure; human-behavior fidelity and causal welfare effects remain unvalidated.

Explore the 90-evaluation map

What we have yet to validate

Population synthesis is not behavior prediction.

Manager has not yet demonstrated reliable human-behavior prediction or real-world effects of intervention. Population synthesis does not establish reduced suffering or financial returns.

See why prediction is not intervention

Ideas and research that came before us

A long history. A new set of tools.

From feedback control and artificial societies to AI-driven policy design, we study what earlier work built, how it works, and what it actually demonstrated.

History & science fiction

The idea of modeling a world

Wiener, Forrester, and Cybersyn; Foundation and the Culture. Exploring the connections between imagined futures and technical systems.

Read the sourced article

Under the hood

MiroFish, Goodfire, Numerai

Simulated societies, neural-network mechanisms, and forecast evaluation. Different systems, different questions.

Explore the mechanisms

Related work

AI Economist, Simile, Aaru

Research on policy design and human responses, with attention to each project’s goals, data, and evidence.

Compare the approaches

Behavior change & persuasion

Persuasion has to change behavior.

Persuasive wording alone is not an outcome. We are building evaluations that track observed choices, follow-up, harms, and subgroup differences.

Read about intervention evaluation

Human–AI collaboration

Boba. Persimmon. Ampersand.

Humans& presents Boba as social intelligence powered by Ampersand and describes Persimmon as a model for simulated people. Their public work raises useful questions about long-horizon collaboration; simulated responses are not observed human outcomes.

Read our Ampersand and Persimmon analysis Explore Humans&’s public work

These references do not imply an affiliation or endorsement. The Manager research library contains 67 published articles: 38 in English and 29 in Japanese, covering population synthesis, behavior prediction, persuasion, EBA, artificial societies, and AI systems. Browse all 67 articles

Independent products

Useful tools. A separate collection.

Explore our apps and web tools, or browse the digital publications.

Apps & web tools Digital publications