Moving my blog
I`m in the process of moving my blog to a new hosting provider and a new blogging platform, Ghost. In the next few weeks I'll be copying all the post to the new platform.
Read more
From Plato to Gradient Descent: The Cage We Forgot We Built
We talk endlessly about what AI can and cannot do. We talk very little about the assumptions built into the question itself. This dialogue follows a philosopher and an AI practitioner through twenty-five centuries of choices, Plato, Aristotle, Galileo, Newton, Turing, Shannon, Hodgkin and Huxley, each one narrowing what counts
There Will Be More Programmers, Not Fewer
AI won't reduce the number of programmers. It will create millions more, because running code you cannot understand is not programming. It is faith. Part I — The Hypothesis The Democratization Arc, and What It Actually Produced. The prevailing narrative about AI and programming runs like this: LLMs will
The Understood-Understanding Gap: Why LLMs Hit a Wall
Large Language Models have become ubiquitous in knowledge work, yet a curious pattern has emerged. Talk to consultants at McKinsey, BCG, Bain, and you'll hear the same refrain: "I don't really know how I can leverage AI to boost my productivity beyond research and brainstorming.
The Shared Delusion: Both doomers and dreamers assume current technology leads to AGI. They’re both wrong
Excel didn’t kill accountants; it gave them better tools. Yet, with AI, we’ve abandoned this logic in favor of a collective fantasy. We’re treating a sophisticated calculator as a human replacement rather than a human multiplier. The reason for this shift is the "2-to-5-year" horizon.