Revonex publishes in-depth analysis on large language models, autonomous systems, and the engineering behind intelligent software. Clear thinking on what actually matters.
Model releases, research worth reading, and tools reshaping how we build — curated every week.
How large language models actually work — attention mechanisms, emergent behavior, alignment, and where the field is heading.
Ongoing researchThe craft of building software — system design, trade-offs, working with AI tools day-to-day, and lessons from shipping things.
Regular writingBreakdowns of interesting papers, ideas from cognitive science, and experiments I'm running on the side.
Deep divesTools I'm building, experiments in the open, and thoughts on the culture of sharing knowledge in the AI era.
ProjectsI'm a builder who got obsessed with AI around the time transformers started doing things that felt genuinely strange. Since then I've been deep in the weeds — building autonomous agents, studying cognitive architectures, and trying to understand where the interesting failures happen.
Revonex is an independent publication focused on AI and engineering. I write here to think in public — sometimes half-formed notes, sometimes long-form deep dives. Not to teach, but to understand. I hope some of it is useful to you too.
When I'm not writing or building, I'm probably reading something about emergence, probability, or why systems fail in unexpected ways.