https://dylanwatt.com/Recent content on Dylan WattHugoen-usThu, 12 Jun 2025 00:00:00 -0700https://dylanwatt.com/posts/2025-06-12-5-things/Thu, 12 Jun 2025 00:00:00 -0700https://dylanwatt.com/posts/2025-06-12-5-things/<style> div.title h1 { display: none; } </style> <div style="background:rgb(35, 35, 51); padding: 2rem 2rem; position: relative; overflow: hidden;"> <h1 style="color: white; font-size: clamp(2rem, 5vw, 3.5rem); font-weight: 800; line-height: 1.1; margin: 0 0 1rem 0; max-width: 900px;"> 5 Essentials for getting the most from Coding Agents </h1> <p style="color: #9ca3ff; font-size: 1.25rem; margin: 1rem 0 0rem 0; max-width: 600px;"> Spend the time to teach them your Project </p> </div> <p>There are a lot of AI coding agents out there, but they all stall out if you don&rsquo;t tool your codebase for them. There&rsquo;s a ceiling of how effective they can be if they&rsquo;re running in a boilerplate environment, with no customized context or guidance.</p>https://dylanwatt.com/posts/2025-06-05-mcp-heel/Tue, 03 Jun 2025 00:00:00 -0700https://dylanwatt.com/posts/2025-06-05-mcp-heel/<p>I&rsquo;ve been getting into personalizing Claude desktop for more much more detailed, powerful responses.</p> <p>Model Contect Protocol (MCP), is a way to expose tools (APIs) to an AI agent to get custom data, or do custom actions. The issue is that the request and response are all fully written and read by the LLM, so for large datasets, you&rsquo;re making the LLM your data transport layer as well as an AI.</p>https://dylanwatt.com/posts/2025-05-29-computer-use/Wed, 21 May 2025 04:00:00 -0700https://dylanwatt.com/posts/2025-05-29-computer-use/<meta property="og:image" content="{{ .Site.BaseURL}/computer_use/image.png" /> <style> @media screen and (max-width: 600px) { #desktop-text { display: none; } } @media screen and (min-width: 601px) { #mobile-text { display: none; } } } </style> <div style="display: flex; gap: 2rem; align-items: flex-start;"> <div style="flex: 1;"> <p>I recently had a well upvoted comment on hacker news regarding an AI agent and, as is the way of the internet, somebody decided to stalk my online presence to find something unrelated to be mean about. </p>https://dylanwatt.com/posts/2025-04-29-loops/Tue, 18 Mar 2025 00:00:00 -0700https://dylanwatt.com/posts/2025-04-29-loops/<p>The City upon a hill for AI coding is that a well specified ticket should get an <a href="https://x.com/brian_armstrong/status/1891921255290503545">automated pull request implementing it.</a> The act of coding itself will go from carpentry to CNC, where you still must understand the properties of the material you work with, but shaping that material is a easily repeatable task, so you just focus on the goal.</p> <p>Many of the day to day tickets a software engineer runs into are already at the point where a well instrumented AI Agent can handle the changes. Updating a title, creating a well specified screen, and even implementing a simple CRUD endpoint are almost fully solved. However, many tasks, like creating a database migration, managing security concerns, and other tricky tasks are beyond current agents.</p>https://dylanwatt.com/posts/2025-02-13-meta/Thu, 13 Feb 2025 00:00:00 -0700https://dylanwatt.com/posts/2025-02-13-meta/<p>When I first started writing code, I had a sense of joy in the fact I was learning the arcane incantations that brought modern day magic to life. Over time that feeling faded as I faced the realities of building code at scale. Bogged down by so many opinions of the &ldquo;right&rdquo; way to do things, instead of the joy of creation I now felt the dread of knowing I hadn&rsquo;t tested the code well enough, or that dependencies weren&rsquo;t injected, or that I had 327 npm deprecation warnings.</p>