WhoAmI-Lab PRO
Model-creator1983
AI & ML interests
A pseudonymous research lab for small AI systems, external-context evaluation, and reproducible benchmark traces.
Recent Activity
posted an update 14 days ago
Clarification
Earlier posts were informal notes on an early prototype.
Public claims from this account will be limited to reproducible benchmark traces, failure cases, ablation notes, and black-box demo records.
No performance claim should be treated as final until the evaluation package is published. posted an update about 1 month ago
The principle of LLM is to predict the following words. A large amount of learning data and long-term GPU calculations are required to improve performance. I succeeded in developing an initial model with a different principle from LLM. The possibility of realiving the same task with an average calculation of 1 to 5% for each comparison item has been verified. The forecast to reach the frontier-class bench is 3 to 6 months later (2026/9/1) posted an update about 1 month ago
CPU-Oriented Selective Inference for Persistent AI
Independent researcher and developer based in Japan, exploring efficient AI systems for modest computing environments.
This project studies whether structured state and selective execution can reduce unnecessary computation in long-video and multimodal tasks while preserving reliable behavior.
Current prototype
CPU-only Synthetic World prototype:
* 64Ă—64 tile-native video
* Five development seeds, 360 schedule-offset cases
* W1/W2 quality gate: 99.17%
* W3 safe non-answer: 100%
* Deterministic replay: 100%
* Cold end-to-end wall-time ratio: 0.40090Ă— versus matched full-pass
These are limited engineering measurements—not proof of universal 10× acceleration, trained-model superiority, real-world high-resolution performance, or consciousness.
Goals
Process task-relevant evidence, maintain persistent experience, select small specialized components, separate perception/memory/reasoning/verification, abstain when evidence is insufficient, and record costs and failures transparently.
Long-term goal: practical AI that can observe, remember, reason and adapt on affordable CPU and edge hardware.
The internal method is intentionally not described publicly. Technical evaluation is welcome through black-box demos, benchmark traces, profiling data and preserved failure cases.
Next: longer/higher-resolution video, occlusion and camera-shift tests, trained models, fair baselines, independent implementations, and CPU/memory/energy measurements.
Experimental project. Reproducibility and honest accounting come first.Organizations
None yet