I investigate learning, memory, evaluation and behavior in AI systems. Hypotheses, findings and limitations are reported separately.
OPEN EXPLANATION Orbital-dynamics research using restricted three-body simulations. Co-authored with Othon Cabo Winter, Ernesto Vieira Neto and Gislaine de Felipe. Earth, Moon and Planets, 100, 233–239.
The institutional record confirms authorship and publication. Access to the full article may be restricted.
OPEN EXPLANATION Study of learning, memory and evaluation in AI systems, including experimental-instrument validity and the classification of model, tool and measurement failures.
Internal manuscripts are not equivalent to peer-reviewed publications. A process finishing without an error does not establish a hypothesis.
OPEN EXPLANATION A testing program using protocols, comparisons and external verifiers. DeepSeek Harness (DSH) is an execution tool; the HARNESS program maintains its own scientific identity.
Invalid and inconclusive results are preserved. Technical canaries do not establish general improvements in intelligence.
OPEN EXPLANATION Combines hypothesis analysis with controlled reconstruction and ablation of functions such as memory, recurrence and self-modeling. Deep Psyche Harness (DPH) is this research line’s experimental architecture, distinct from DSH.
Convincing behavior and self-reports do not prove subjective experience. No claim of demonstrated artificial consciousness is made here.
An ablation removes a defined function, such as memory, to compare behavior under controlled conditions. A functional difference is not a measurement of consciousness.
NEXT STEP
Tell me what happens today, where the friction is and what needs to change. We can assess a path toward automation, a product or an AI system.