Can Meta-Learning Learn “Types”? — The Structural Relationship Between Meta-Learning and Ontology Design —
Meta-learning is often described as
“learning how to learn.”
It enables adaptation with very little data.It allows mod...
AI Approximates the World, but It Cannot Approximate Meaning — Continuous Approximation and the Discontinuity of Meaning: Why Meaning Remains in Logic
AI Approximates the World Well
AI approximates the world remarkably well.With astonishing precision and smoothness, it ...
How do we cultivate such designers? — The conditions required for people who can verbalize decision structures, and how to turn experience, failure, and conflict logs into learning.
Up to this point, we have discussed new perspectives required in the age of AI.
The problem is not the ...
Who Is the “Designer” in the Age of AI? — Neither Engineer nor PM, but the Author of Judgment Structure
In the age of AI, the same debate always emerges.
Engineers are critical.PMs are essential.We lack data...
A design that does not define boundaries will inevitably fail — implicit boundaries create accidents, and the cost of articulation becomes the cost of safety.
Accidents appear to happen suddenly.But in reality, they have been prepared for long before they occur.
And their commo...
Increasing model size does not make judgment wiser — discomfort with the scaling myth and giant models without design
AI has grown larger.
Parameters have increased, training data has expanded, and performance benchmarks continue to be s...
AI becomes more intelligent by not committing to decisions — Gradients, probabilities, and the posture of deferral, and architectures that resist premature conclusions
AI is fast.Computation finishes in an instant,and answers are returned immediately.
Because of this, we unconsciously b...
The Real Reason Human-in-the-Loop Fails — Why “Human-at-the-End” Loses Responsibility, and the Shift Toward Human-as-Author
Human-in-the-loop (HITL) has long been presented as a safety mechanism for the AI era.
The human checks at the end.The ...
Is a Multi-Agent System a Failure If It Fails to Reach Consensus? — Designing “Non-Agreement” as an Outcome, and the Value of Conflict Logs —
When people hear “multi-agent,”many expect the following:
Agents discuss with one another.They reconcile their differen...
Treat Cold-Start as a Contract — Transform Data Scarcity into a Design Problem —
Cold-Start Is Considered a Technical Problem
The cold-start problem
is usually treated as a technical issue.
There ...