Grey writes about AI-assisted product development and works on agent infrastructure. The essay AI Makes It Easier to Build the Wrong Product argues for matching the size of an implementation batch to uncertainty about the product.[1][2]
Start with AI Makes It Easier to Build the Wrong Product →
Links
X Contributions
Grey distinguishes bounded work, such as a migration with a clear check, from work whose desired behavior is still uncertain. Larger implementation batches can suit the former. For the latter, the essay recommends small, complete paths through the interface, logic, data and operations, so feedback arrives before the team commits to more work.[2]
Ideas
Tests can establish that software follows a specification without establishing that the specification fits the user’s actual workflow. Grey calls the gap between a plan, code and new knowledge decision debt. The proposed remedy is to update the product decision and acceptance checks after each slice.[2]
Tools, writing & talks
Writing, research & talks
Start here
An essay on batch size, product uncertainty and feedback in AI-assisted development.
Read the decision rule against one planned feature. Software or product experience helps; no agent setup is required.
Author[2]Original source