Posts

I replicate a viral post on weight programming, add complexity analysis and share some opinions.

Summary: This post is aimed to be an accessible introduction to the idea of AI Institutions. The concept of institution is the root that enables the large-scale human cooperation. Understanding its core dynamics can be seen as a necessary step to build AI institutions. The approach is somewhat analogous to this: humans studied birds to understand the physics of flight to build the plane. We here study the principles of human intelligent dynamic cooperation to build the AI institutions.

A small story on the effects of progress, and understanding of the human intellectual endeavour.

We analyze briefly phenomena of social-scientification of computer science, of measuring historically human things, and the classical idea of limits of human reach, contrapposed with the longing of wanting more and more.

This short note explores the connection of rules to the language and the difference between precise and ambiguous rules. It analyzes briefly the concept of intentions and the legislative process, with a focus on artificial societies and its future.

Taking some concepts of punishment from the Human sphere and applying it to AI Agents.

What we shall expect from AI, and how as a society we will need to engage with this problem.

Re-engineering the blog stack and some random thoughts on academia's publish-or-perish culture as a tragedy of the commons.

Fully exploiting the abilities of modern hardware is hard. Commonly written software does not fully exploit all of the underlying hardware optimization abilities, and often needs specialized software to achieve better utilization. This post will explore some of the most common optimization techniques for single-core binaries executing in modern hardware. We will cover concepts as Instruction-Level Parallelism, cache analysis for specific computations and briefly touch over vectorization. We conclude with a short example showcasing how these techniques are actually used and an analysis of these results. We report a maximum of 10x speedup over the standard naïve matrix-matrix multiplication code.