https://rochanewton.github.io/tags/discernment/Recent content in Discernment on Newton RochaHugo -- gohugo.ioen© 2026 Newton RochaSun, 13 Sep 2026 00:00:00 +0000https://rochanewton.github.io/posts/claude-cert-03-output-evaluation-and-validation/Sun, 13 Sep 2026 00:00:00 +0000https://rochanewton.github.io/posts/claude-cert-03-output-evaluation-and-validation/<h2 class="relative group">What this is about <div id="what-this-is-about" class="anchor"></div> <span class="absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100 select-none"> <a class="text-primary-300 dark:text-neutral-700 !no-underline" href="#what-this-is-about" aria-label="Anchor">#</a> </span> </h2> <p>Domain 1 of the <a href="https://anthropic-partners.skilljar.com/claude-certified-associate-foundations-certification" target="_blank" rel="noreferrer">Claude Certified Associate – Foundations</a> exam is <strong>Output Evaluation and Validation</strong>. It&rsquo;s worth 21% — more than any other domain on the exam. The whole domain comes down to one idea: <strong>good-looking output is not the same thing as validated output.</strong> Fluent, confident text doesn&rsquo;t tell you anything about whether it&rsquo;s actually correct. This post breaks down the framework for closing that gap, on purpose, instead of by accident.</p>