Small Models, Big Mouths: Why Game AI Doesn’t Need Giant Brains
A mechanism-first reading of DefameLM: why narrowly scoped small language models may be more practical than giant cloud LLMs for real-time game AI and some business automation loops.
Cognaptus Insights
Industry thoughts, trends, and commentary.
Thinking in Panels: Why Comics Might Beat Video for Multimodal Reasoning
A business-focused reading of Thinking with Comics, a paper arguing that comic panels may offer a cheaper and more structured middle path between static images and video for multimodal reasoning.
ThinkSafe: Teaching Models to Refuse Without Forgetting How to Think
A mechanism-first reading of ThinkSafe, a self-generated safety-alignment method that restores refusal behavior in reasoning models without paying the usual teacher-distillation tax.
When Language Learns to Doubt Itself: Self-Contradiction as an Upgrade Path for Multimodal AI
Self-contradiction in multimodal models is not just a failure signal; it may be a cheap diagnostic for aligning generation with understanding.
When LLMs Meet Time: Why Time-Series Reasoning Is Still Hard
A close reading of TSAQA shows why turning time series into question-answering tasks helps evaluate LLMs—but does not magically give them temporal reasoning.
When One Patch Rules Them All: Teaching MLLMs to See What Isn’t There
A mechanism-first reading of how one reusable visual perturbation can steer closed-source multimodal models toward a chosen target across unseen images.
Agentic Systems Need Architecture, Not Vibes
A mechanism-first reading of why reliable AI agents need subsystem architecture, reusable design patterns, and clearer diagnosis than another enthusiastic list of agent tricks.
Algorithmic Context Is the New Heuristic
A new A* heuristic-design paper shows why algorithmic context can matter more than vague domain prompting when LLMs are used inside constrained optimization workflows.
Ask Once, Query Right: Why Enterprise AI Still Gets Databases Wrong
A mechanism-first reading of why enterprise database routing fails when it relies on embeddings or prompt-only LLM reranking, and why schema coverage plus connectivity checks matter.
GAVEL: When AI Safety Grows a Rulebook
A mechanism-first reading of GAVEL, a rule-based activation monitoring framework that turns model-internal signals into auditable AI governance logic.