Architecture on Todd W. Bucy — Research Blog

Productive Friction: The Variable You're Already Trying to Control

Thu, 28 May 2026 00:00:00 +0000

Part 2 of a 3-article series on agent memory architecture. Part 1 made the cognitive-architecture argument that memory belongs in a categorically different component from the reasoning engine. This article names the formal variable that determines whether the architectural split is doing real work, and gives practitioners a diagnostic they can use on their own systems. Part 3 describes what the engineering commitments look like.

You cannot move a car on a perfectly smooth road with perfectly smooth tires. The engine is fine. The transmission is fine. The wheels rotate exactly as designed. Nothing is broken. The car doesn’t move because friction between tire and road is the medium through which rotational work becomes translational motion, and without it, the work has nothing to push against. The car doesn’t fail to accelerate. It has no purchase from which forward momentum is possible no matter how fast the wheels turn.

Latency Is the Enemy of Agency

Mon, 18 May 2026 00:00:00 +0000

Why an Engineer with the Hardware in Front of Him Wants AI to Do Better

Personal note in a series that has been making architectural arguments. The architectural papers describe what the framework is. This piece describes why someone would build it. Specifically: why I built it, what about current AI deployment offends me as someone who actually understands the hardware, and what the architectural commitments come from.

Your Model Has Humanity's Cortex. It Needs Its Own Hippocampus.

Sun, 17 May 2026 00:00:00 +0000

Part 1 of a 3-article series on agent memory architecture. This article makes the cognitive-architecture argument: the model has the cortex part, humanity’s verbal and reasoning capacity in compressed form, but no working hippocampus of its own, and adding more text to its context will not give it one. Part 2 names the variable that determines whether the architectural alternative works. Part 3 describes what the engineering commitments actually look like.