Erik Hanchett

Hanchett’s owned biography records a computer-science bachelor’s and an entrepreneurship MBA at the University of Nevada, Reno. His public account identifies earlier frontend engineering at AWS before developer advocacy. His educational work spans the Vue and Ember ecosystems and now includes agents and cloud services, using concrete applications to explain how the surrounding tools fit together.[2][1]

Keep model judgment separate from the application’s decision

In his Jev tutorial, Hanchett sends an application state and several narrow questions to an API that returns probabilities instead of conversational text. His pull-request example asks about requirements and credential handling, then maps the scores to pass, review or block in ordinary code. Conditions such as whether a change actually includes interactive UI also remain in code. A structured response makes integration predictable, while its judgments still need appropriate checks.[3]

Build an interface around a complete application

Vue.js in Action develops a web-store example with product listings, a shopping cart, checkout and an administrative interface. Hanchett and Benjamin Listwon use that application to teach reactive interfaces, data modeling, API use and shared state. The September 2018 edition documents an earlier stage of the Vue ecosystem, giving his teaching work a longer history than his current agent-focused material.[4]

His token-cost talk begins with a production agent whose model choices, growing context and long tool loops made it more expensive than expected. The original publisher description identifies prompt caching and model routing as interventions. The engineering question is which repeated input or task needs expensive computation, rather than assuming that a functioning agent is already using resources efficiently.[5]