Patch, Don’t Preach: The Coming Era of Modular AI Safety
A mechanism-first look at safety policy patching, a lightweight way to update LLM safety behaviour without redeploying full model weights.
Cognaptus Insights
Industry thoughts, trends, and commentary.
Proof, Policy, and Probability: How DeepProofLog Rewrites the Rules of Reasoning
DeepProofLog reframes symbolic proof search as policy learning, showing how neurosymbolic AI can scale reasoning without throwing away proof-level interpretability.
The Gospel of Faithful AI: How FaithAct Rewrites Reasoning
FaithAct turns multimodal reasoning from fluent narration into evidence-checked planning, making hallucination less a personality flaw and more an engineering defect.
The Problem with Problems: Why LLMs Still Don’t Know What’s Interesting
A study of math-problem interestingness shows why AI systems need calibrated taste, not just stronger solving ability.
DeepPersona and the Rise of Synthetic Humanity
DeepPersona shows that synthetic users become useful not by becoming longer, but by becoming structured, controllable, and empirically testable.
Forget Me Not: How IterResearch Rebuilt Long-Horizon Thinking for AI Agents
IterResearch shows why long-horizon AI agents need disciplined workspace reconstruction, not merely longer context windows.
Parallel Worlds of Moderation: Simulating Online Civility with LLMs
A mechanism-first reading of COSMOS, an LLM-powered counterfactual simulator for testing moderation strategies before exposing real communities to policy experiments.
Touch Intelligence: How DigiData Trains Agents to Think with Their Fingers
A mechanism-first reading of DigiData, Meta’s dataset and benchmark for training mobile agents to complete real app tasks rather than merely imitate taps.
When Agents Think in Waves: Diffusion Models for Ad Hoc Teamwork
A mechanism-first reading of PADiff, showing why diffusion policies may help agents preserve multiple cooperation plans when working with unfamiliar teammates.
When AI Argues Back: The Promise and Peril of Evidence-Based Multi-Agent Debate
ED2D shows that evidence-grounded AI debate can make misinformation correction more persuasive, but also more dangerous when the system is wrong.