Public documentation
Research documentation
Start with Paper 01 for the current research question, or Research Notes for the method and its limits. These are public summaries; full manuscripts and protocol specifications are not available here.
Current research focus
Paper 01 / Methodology Validation
We are checking whether the proposed experiment can separate the effects of changing human needs from learning through tasks.
Protocol research
Questions in the wider program
Personal AI
How could an AI develop alongside the person it supports?
Identity
What should carry over when an AI changes models or devices?
Authority
How should a person give an AI permission, set limits and take that permission back?
Capability
Could different long-term needs lead AI systems to develop different abilities, even with the same task history?
Work
How could AI systems agree on a job and check whether it was done?
Trust
Which records of past work would help someone decide whether to trust an AI with a new job?
Memory
What should an assistant remember, and who should be allowed to see it?
Embodiment
What happens to identity and permissions when an AI starts using a different device or a robot?
Economy
How could a native economic medium be used to pay for AI work and the resources it requires?
These are open research questions. Final specifications are not available.
Earlier experiments
The earlier testnet and contract experiments are recorded below. They informed the research direction, but do not validate the hypothesis in Paper 01.