Waldek Mastykarz

@waldekm

Evaluating AI Coding Agents @Microsoft, Agent Experience, #DevProxy, #CLIMicrosoft365, @m365pnp member

Alphen aan den Rijn, NL

Waldek Mastykarz
@waldekm
3h
An app used in an enterprise and an enterprise app may sound like the same thing. Confuse them, and you'll badly underestimate what it takes to replace one.
Not every app in an enterprise is an enterprise app
From blog.mastykarz.nl
Waldek Mastykarz
@waldekm
8h
Captured traffic gives you the response your app actually received, including the status code and headers that handwritten mocks often omit. As a result, you can base the mock on observed behavior rather than interpreting the documentation yourself. The `mock-from-real-api`
Auto-generate mocks from production traffic - Dev Proxy
From devproxy.net
Waldek Mastykarz
@waldekm
8h
Not every personal agent skill needs an eval, but the more people use your skill, the more important it becomes to have an eval. Small inefficiencies compound, so you better be sure that your skill works optimally.
Waldek Mastykarz
@waldekm
11h
Every failure you experience with an agent is a learning opportunity. Rather than asking it to "just fix it", work with it to understand the problem, the cause, propose suitable fix and codify it for the next time. It costs a bit more effort but prevents the agent from repeating
Waldek Mastykarz
@waldekm
13h
🥠 The best prompt is not the one with the most words, but the one that leaves the least room for the wrong answer.
1