https://toddwbucy.github.io/toddwbucy_research_blog/blog/Recent content in Blog on Todd W. Bucy — Research BlogHugoen-usThu, 19 Feb 2026 09:00:00 -0600https://toddwbucy.github.io/toddwbucy_research_blog/blog/2026/02/nobody-knows-what-an-agent-is/Thu, 19 Feb 2026 09:00:00 -0600https://toddwbucy.github.io/toddwbucy_research_blog/blog/2026/02/nobody-knows-what-an-agent-is/<p><img src="https://toddwbucy.github.io/toddwbucy_research_blog/images/blog/nobody-knows-what-an-agent-is/image2.jpg" alt="" /></p> <h2 id="who-this-is-for">Who This Is For<a class="anchor" href="#who-this-is-for">#</a></h2> <p><strong>This is not a research paper.</strong> It is a diagnostic framework for practitioners, and it is deliberately written as one. Researchers are busy building the next model. This is for the people trying to make agents work in production. The engineers, architects, and technical leaders who get up every morning and face the real problem: how do you build agent systems that are reliable, predictable, and economically viable?</p>https://toddwbucy.github.io/toddwbucy_research_blog/blog/2026/01/google-didnt-go-silent-you-stopped-looking/Thu, 29 Jan 2026 09:00:00 -0600https://toddwbucy.github.io/toddwbucy_research_blog/blog/2026/01/google-didnt-go-silent-you-stopped-looking/<p><img src="https://toddwbucy.github.io/toddwbucy_research_blog/images/blog/google-didnt-go-silent/image1.jpg" alt="" /></p> <p><em>A response to Delanoe Pirard&rsquo;s &ldquo;Transformers Are Dead&rdquo; and &ldquo;A 170M Model Beats GPT-4&rdquo;</em></p> <p>In December 2025, Delanoe Pirard published two articles that crystallized a question the ML community had been asking for a year: If Titans is so good, why has Google gone silent?</p> <p>The answer is this: Titans is one component of a seven-paper research program, not a standalone architecture. Google didn&rsquo;t go silent. The community stopped looking for the rest of the papers.</p>