Exploring the Future Paradigm of the IT Industry

The output of large language models is not a finished product — it is a vast source of information entropy. On its own, it is noisy, unstructured, and overwhelming. The real challenge, and the real opportunity, is how we shape, filter, compress, and orchestrate this entropy into usable systems — socially, organizationally, and business-wise.

I see LLMs as raw material: abundant but chaotic, powerful but directionless. Just as past industrial revolutions required new protocols for energy, logistics, and computation, this new era will demand protocols for language — rules and architectures that make AI outputs callable, transferable, verifiable, and schedulable. Like LEGOs!

This Substack is where I explore those protocols and paradigms:

• Technology — how language itself becomes system design.

• Business — how to build viable products and markets on top of entropy-rich AI output.

• Society — how governance, law, and culture evolve when intelligence is ambient, probabilistic, and everywhere.

My aim is to chart the design patterns of an “entropy-aware” AI era, where the challenge is not only what AI can generate, but how we transform that raw entropy into enduring value.