I build tools that work and feel great. Essays on AI systems, philosophy, building, and more.
An agent's memory is everything it can reach, not just its context window. A preprint, Reading More, Finding Less, tests that idea with pre-registered experiments and finds the failure is routing: a hand-curated memory index sent the agent to the wrong files, plain file search beat it, and saved memories almost never got read again. The full corpus and harness ship with the paper for re-running.
Many teams use one AI model to grade another one's answers, and then let that grade decide what ships. Almost nobody checks whether the grader itself can be trusted. This is what it took, on a real production system, to make the grader reliable enough that a failing score is allowed to stop a release.
AI agents are built to remember, and that breaks their tests. When test cases share one user, each case inherits leftover memory from the one before it, so the pass rate measures the order the tests ran in, not the agent. Giving every test its own clean user and its own sealed world fixed that, and let the whole suite run in parallel for free.
First month inside a global org: docs as code, every API wired, and capabilities that visibly compound.
The cost of intelligence is collapsing. A map of who captures the value, with five predictions you can check.
Execution became abundant, and supervision dies with it. Three jobs remain: intent, terrain, and guarantees.
Git assumed one human, one machine, one source tree. Agents just broke all three assumptions.
Human plus machine ruled chess for a decade, until the engines stopped needing us. It is happening again.
When a capability becomes reliable enough, the task disappears from consideration. The constraint moves up the stack.
A close friend reflected back what he saw in me. Not conclusions — questions.
What becomes possible when an entire company aligns around a single execution stream.
I know how to run a remote team. That is exactly why I stopped optimizing for it.
I do not type faster or sleep less, yet my output multiplied. The speed lives in the infrastructure.
When the infrastructure around you can finally receive your thinking whole, you meet yourself in a larger room. Character arrives all at once.
The Anatomy of a One-Shot Prompt by Rashid Azarang
The Intelligent Sales Agent by Rashid Azarang
Costs of Emotional Outsourcing with AI by Rashid Azarang
Cognitive Relativity by Rashid Azarang
Personal Constitution by Rashid Azarang
Integrating Ideals with Reality by Rashid Azarang
On the Right Relationship Between Humans & AI by Rashid Azarang
What People Miss About Canvas by Rashid Azarang
Conversations Into Memory by Rashid Azarang
Navigation Layer by Rashid Azarang
The Network Effect of Intelligence by Rashid Azarang
Cognitive Capital: Assets That Think by Rashid Azarang
Built an Open-Source SMS Dashboard by Rashid Azarang
Why Your AI Agent Sucks at Front-End by Rashid Azarang
Software Architecture for Parallel AI by Rashid Azarang
The hard part is not asking questions. It is encoding what answers mean.
How AI Should Analyze Your Data by Rashid Azarang
Architecting Weather Data Storage by Rashid Azarang
Why AGI is Stuck and How to Unstick It by Rashid Azarang
Between Code and Consciousness by Rashid Azarang
Building A Supply Chain Risk Management Platform by Rashid Azarang
Contextual Gravity by Rashid Azarang
Deep Research on Rashid by Rashid Azarang
Going All In on Intelligence by Rashid Azarang
Airtable AI Agent by Rashid Azarang
Software Isn’t One-Size-Fits-All by Rashid Azarang
An ancient Sage valued himself upon knowing how to make a great City of a little one. Franklin, that modern Simpleton, showed how to reduce an Empire by alienating its provinces. I, a contemporary observer, shall demonstrate how the same may be accomplished not through distance but through division—not by severing remote territories but by fracturing the very heart of society itself.
When AI Gets Its Hands by Rashid Azarang
Replacing Scattered Scripts with Centralized Intelligence by Rashid Azarang
Getting Software to Talk to Each Other by Rashid Azarang
How to Spot Exceptional Market Opportunities by Rashid Azarang
Structure Smart by Rashid Azarang
We stand at the threshold of humanity's most profound partnership. Artificial intelligence promises to be our collaborator in solving civilization's greatest challenges—from climate change to disease, from loneliness to…
Here's a question that's about to matter a lot: when an AI helps you discover something, who gets the credit? This isn't hypothetical anymore. AI systems are writing code, finding patterns in data, even proposing new scientific hypotheses. The old frameworks for attribution—built when tools were dumb—are breaking down.
Stop thinking of AI as a single tool, and start thinking of it as a specialized workforce.
From Chaos to Living Architecture by Rashid Azarang