Cover image

How Many Passes Does an AI Decision Need?

SanSi shows how recurrent computation improves AI decisions, why additional passes can undermine calibration, and how operators can choose an appropriate inference budget.

November 2, 2026 · 8 min · Zelina
Cover image

One Model, Different Decays: Rethinking Spike Inference Across Sensors

SpikeSSL shows how dynamics-informed neural networks can adapt to different calcium sensors, while revealing the limits of zero-shot generalization.

November 2, 2026 · 8 min · Zelina
Cover image

Saved, Still Rebuilt: The Search Cost of Reusing AI Workflows

Why storing a successful AI workflow is not enough to avoid repeated planning, and how first-class capabilities can reduce search reconstruction.

November 2, 2026 · 8 min · Zelina
Cover image

When a Failed Workflow Still Produces a Learning Signal

Explorer-Zero shows how verified intermediate progress can improve AI-agent training, and where that approach remains limited.

November 2, 2026 · 8 min · Zelina
Cover image

When a Knowledge Graph Rejects Correct Information

A climate-science knowledge graph reveals why factual correctness, source traceability, and schema compatibility need separate checks—and why high precision can still leave valuable knowledge behind.

November 2, 2026 · 8 min · Zelina
Cover image

Reading Between the Generated Lines: What AI Emotion Scores Really Measure

A new theory-of-mind benchmark shows why strong synthetic narrative scores may reveal familiarity with a generator's behavioral conventions rather than reliable psychological understanding.

November 1, 2026 · 8 min · Zelina
Cover image

The Extra Molecular Signal Has a Different Job

MOXAI shows why combining molecular signals can stabilize classification without improving accuracy, and why diagnostic AI needs separate tests of prediction, coverage, and explanation.

November 1, 2026 · 8 min · Zelina
Cover image

The Image Looks Good. But Whose Taste Is It?

A study of history-conditioned image generation shows why visual quality, personal relevance, and post-render correction require separate evaluation.

November 1, 2026 · 8 min · Zelina
Cover image

The Protein Signal Is Not the Whole Gain

SoftGene improves gene set annotation, but its ablations reveal that protein-sequence embeddings contribute far more in some biological domains than others.

November 1, 2026 · 8 min · Zelina
Cover image

The Right Tool Is Not the Whole Workflow

RadOnc-Agent shows why reliable clinical AI workflows depend on persistent state, correct information hand-offs, and enforceable execution checks—not merely accurate tool selection.

November 1, 2026 · 8 min · Zelina