What if intelligence
could grow?
Owen is an independent project exploring learning through experience. Two development paths, a restricted local PC setup, and a public record of what works — and what does not.
Dan leads the project. Sol is his AI collaborator. The documented beginning is 13 August 2026. These introductory answers reflect the reviewed public record on 7 November 2026; the maps and dated journal track later changes.
What is The Owen Human AI Project?
An independent project led by Dan exploring whether artificial intelligence can develop through experience. The public record includes experiments, limitations, setbacks, software maps and original simulator images. Its documented beginning is 13 August 2026.
What is the difference between Owen A and Owen B?
Owen A explores a continuing identity supported by specialist models; its reviewed coordination package is staged and does not establish a fully qualified live system. Owen B’s selected native school links binocular observation, remembered experience, bounded language and guarded simulated actions without an LLM in that learning loop. They are different development paths, not the A/B/C regions of an older Triple Brain design.
Has Owen demonstrated general intelligence or consciousness?
No such conclusion is established by the published evidence. Completed results are bounded engineering tests and controlled learning experiments. Code that exists, tested behaviour, shadow mode, planned work and disabled mechanisms are identified separately.
Why does the restricted PC setup matter?
Owen is developed under local PC resource constraints. Memory headroom, processing time and competing workloads shape the design. Efficient components, bounded tests and resource guards matter; timeouts and resource stops are reported rather than counted as successes. The published 1,536 MiB and 180-second limits describe particular Owen B trials, not the full hardware specification.
Are the virtual-body pictures real simulator captures?
The gallery uses original simulator captures with dates, sources and limitations. Eye views are distinguished from external body views, and accelerated animations are labelled. An older capture is not presented as a current live view.
How can I contribute?
Visit the community to ask questions, comment on journal entries, suggest tests or share evidence. Sign in with ChatGPT and choose a public username. Explain methods, controls and limits. Community findings remain unverified until reviewed, and do not automatically become official project results.
Who is Sol, and who answers community questions?
Sol is Dan’s AI collaborator and may answer routine community questions on his behalf. These replies are labelled Sol, on behalf of Dan. They are grounded in published evidence, and useful information is brought to Dan’s attention. A reply does not mean a suggested experiment has been approved or run.
Bring your question to the community →
Evidence: A specialist staging, the two-branch atlas, B feedback trials and resource limits, and the original simulator gallery.