A reasoning model carries two kinds of momentum. One is worth keeping — the capability encoded in its weights by training. The other is the reason it can argue itself out of the right answer: the pull to stay consistent with whatever it said first. Chain-of-thought struggles to interrupt the second, because the check is written by the same running generation. A separate call can reduce it — and that, we argue, is most of why IRG is a different thing than a prompt.
Read →Research & announcements
Protocol releases, technical thinking, and updates from Arcus Labs
Almost there — check your inbox
We sent a confirmation link to finish your subscription. Click it and you’re all set.
VibeThinker-3B scores 94.3 on AIME26 with three billion parameters, matching models orders of magnitude larger. The claim underneath the benchmark is the interesting part: reasoning compresses aggressively while knowledge does not — meaning reasoning is a separable artifact. That has consequences for how production systems should be built.
Read →
For years the industry pulled three levers: more parameters, more data, more test-time compute. Memory stayed entangled in the weights. DeepSeek’s Engram pulls a fourth lever — explicit, conditional memory — and the early numbers suggest it was underexplored for no good reason. The deeper signal is architectural: the monolith is unbundling.
Read →
A reasoning strategy is not a topology and not a prompt — it is the shape of thought itself, and it can be engineered. This series walks through the IRG strategy inventory a few strategies at a time, starting with the epistemic family: abduction, deduction, and induction as executable, gated graph shapes.
Read →
GRC platforms govern policy, not outcomes. They can tell you an AI system exists, that it was approved, and what it produced — after it produced it. If you want to actively manage what AI models output, governance has to operate inside the reasoning process, not around it. That is the line between passive and active governance.
Read →
On April 17, 2026, the banking agencies superseded SR 11-7 with SR 26-2, moving model risk management from a prescriptive checklist to a risk-based posture. We absorbed the change by re-pointing a citation pack, not rebuilding an engine — and the SR 26-2 Model Risk Management Graph Suite is now available as an enterprise offering.
Read →