http://programmer.ie/tags/release-gates/ Recent content in Release Gates on Programmer.ie: Modern AI programming Hugo en-us Wed, 02 Sep 2026 16:15:00 +0100 http://programmer.ie/books/browser-ai-from-first-principles/14-chapter/ Wed, 02 Sep 2026 16:15:00 +0100 http://programmer.ie/books/browser-ai-from-first-principles/14-chapter/ <p>A demo asks whether one carefully chosen example can work.</p> <p>A feature evaluation asks how often the product contract holds across the inputs, states and failures users will actually encounter.</p> <p>Both have value. Confusing them is the problem.</p> <hr> <h2 id="1-the-unit-of-evaluation-is-the-feature">1. The unit of evaluation is the feature</h2> <p>“Evaluate Gemma 4” is too broad for our application. “Evaluate this prompt” is too narrow.</p> <p>The useful unit is something a user attempts:</p> <div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;"><code class="language-text" data-lang="text"><span style="display:flex;"><span>summarize selected documentation </span></span><span style="display:flex;"><span>rewrite permission copy without changing authority </span></span><span style="display:flex;"><span>detect language and decide whether to translate </span></span><span style="display:flex;"><span>produce structured arguments for a read-only tool </span></span></code></pre></div><p>Each feature includes input preparation, capability selection, model output, validation, fallback and UI behavior.</p>