5 pieces
Compact intelligence01
Lab analysis / 1 October 2026
Exploratory resultsA smaller operator. The same behavior?
Our next question is whether compact hidden-state approximations survive inside the complete model.
Efficient architectures02
Lab analysis / 1 October 2026
Implementation under studyCan rational updates buy more expression per parameter?
Continued fractions give us a different primitive to explore. Stability and fair comparisons come first.
Adaptive graphs03
Lab analysis / 1 October 2026
Research hypothesisGrow the graph. Hold the active budget.
A graph that stores more capability without turning every training step into an irregular, expensive traversal.
Generative systems04
Research news / 29 September 2026
External research · our interpretationSmall controllers, large generative models.
A fresh control perspective on image generation—and a useful question for modular AI tools.
Generative systems05
Research news / 24 September 2026
External research · our interpretationCoherence needs a memory of the world.
Long-form generation raises a familiar creative-tool problem: keeping one world consistent across many decisions.
Research news edition: 24–30 September 2026. A dated editorial selection; our interpretations are labeled on each piece.