<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Christopher Delaney on Medium]]></title>
        <description><![CDATA[Stories by Christopher Delaney on Medium]]></description>
        <link>https://medium.com/@CJDelaney?source=rss-c8b3e90eadff------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Qjs6RKpqhkqb1WZX3cS1jw.png</url>
            <title>Stories by Christopher Delaney on Medium</title>
            <link>https://medium.com/@CJDelaney?source=rss-c8b3e90eadff------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 05:54:47 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@CJDelaney/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[We had the tasks. We were missing the bigger picture.]]></title>
            <link>https://medium.com/@CJDelaney/why-we-built-withnettle-d1c8c90fd1fe?source=rss-c8b3e90eadff------2</link>
            <guid isPermaLink="false">https://medium.com/p/d1c8c90fd1fe</guid>
            <category><![CDATA[project-management]]></category>
            <dc:creator><![CDATA[Christopher Delaney]]></dc:creator>
            <pubDate>Thu, 22 Oct 2026 10:15:59 GMT</pubDate>
            <atom:updated>2026-10-22T12:29:03.175Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2s7lpjddlOtWLtZn_8gNxw.png" /><figcaption>An example hierarchy from the WithNettle homepage.</figcaption></figure><p><a href="https://withnettle.com/"><strong>Explore WithNettle — connected work for people and AI</strong></a></p><p>While building CoinKith, our household finance application, we found that keeping track of the work was becoming a job in itself.</p><p>We had tasks written down. We had plans, discussions and progress updates. We tried other ticketing systems, including GitHub’s built-in issues, but we found it hard to read the project as a whole.</p><p>We could see individual jobs. Understanding how they fitted together, what had changed and what needed our attention next took more effort.</p><p>As the project developed, decisions changed and work moved on. A task made sense when we created it, but returning to it later could mean going back through conversations to understand why it existed and whether the original plan still applied.</p><p>Breaking work into smaller tasks helped us get things done. It also made it easier to lose sight of the larger thing we were trying to achieve.</p><p>We wanted to see the project, focus on one area, then get into a specific task without losing its context.</p><p>That became the starting point for WithNettle.</p><p>Tools such as Jira and monday.com already offer ways to organise complex work. Jira has epics, tasks and subtasks, with additional hierarchy options. monday.com has boards, items and multi-level subitems. Having a hierarchy alone isn’t the difference. [1] [2]</p><p>The difference we wanted was in how you read and move through the work.</p><p>WithNettle puts the connected structure at the centre. Start with the larger goal, break it into branches, and follow a branch down to the task you need. When you need the wider picture again, move back up.</p><p>You don’t always want to look at everything. Sometimes you need to understand the whole project. Sometimes you just need to concentrate on one small part of it.</p><p>We wanted both to feel natural.</p><p>That matters when you’re managing several projects. A single view of every task can be overwhelming, while a narrowly filtered list can hide the reason those tasks exist. WithNettle’s approach is to let you focus within the structure, so the wider context remains connected to the work.</p><p>The tree shows how the work fits together. The board helps you follow its progress. The timeline helps you see it over time. They are different views of the same work, so you can choose the view that answers the question you have.</p><p>For us, that is a more useful starting point than adding another dashboard to an already complicated setup.</p><p>We also wanted decisions and explanations to stay close to the tasks they affect. When someone picks up a piece of work, they should be able to understand its purpose and follow the surrounding context without needing somebody else to reconstruct the story.</p><p>Working with AI made that need clearer.</p><p>An assistant could help with an individual job, but it needed the same background we did. What were we trying to achieve? What had already been decided? What had changed since the last session?</p><p>WithNettle gives people and connected AI tools a shared place to read and update that project context. The work can continue across conversations, rather than depending on one chat holding everything together.</p><p>Those questions matter just as much when the next person picking up the work is a teammate — or you a week later.</p><p>We’re building WithNettle around that everyday problem: making work easier to understand, easier to return to and easier to move forward.</p><p>CoinKith taught us that writing everything down was only part of the job. We also needed to make it readable.</p><p>We already had somewhere to store the tasks. We wanted somewhere we could see how they belonged together.</p><p><a href="https://withnettle.com/"><strong>Try WithNettle for your next project → withnettle.com</strong></a></p><h3>Sources</h3><p>[1] <a href="https://www.atlassian.com/software/jira/guides/issues/overview">Jira’s structure</a></p><p>[2] <a href="https://support.monday.com/hc/en-us/articles/29810815287570-Multiple-levels-of-subitems-on-monday-com">monday.com’s hierarchy</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d1c8c90fd1fe" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Human UX Matters More Than Ever in an AI-Driven Development Pipeline]]></title>
            <link>https://medium.com/@CJDelaney/why-human-ux-matters-more-than-ever-in-an-ai-driven-development-pipeline-cbbde8637437?source=rss-c8b3e90eadff------2</link>
            <guid isPermaLink="false">https://medium.com/p/cbbde8637437</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Christopher Delaney]]></dc:creator>
            <pubDate>Wed, 21 Jan 2026 17:33:26 GMT</pubDate>
            <atom:updated>2026-01-23T16:34:26.791Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uILLe-xNmg1Ur8T9SQ0SMg.png" /></figure><p><strong>Speed without judgement is now the biggest risk in software delivery</strong></p><p>I have worked in technology for over twenty-five years, the past seven years as a CTO. I have seen multiple waves of tooling promise efficiency and cost reduction. Most delivered incremental improvement. A few genuinely changed how teams worked. Large language models fall into the latter category.</p><p>AI has changed how software is built.</p><ul><li>Developers no longer search first. They ask AI.</li><li>Scaffolding, boilerplate, and exploration of unfamiliar frameworks now take minutes, not days or weeks.</li><li>Learning curves have shortened dramatically.</li></ul><p>Used properly, this reduces friction and accelerates delivery. Any experienced engineering leader can see the benefit.</p><p>That success is leading some organisations to draw the wrong conclusions.</p><p>Despite widespread automation, we still accept that certain roles remain essential. We still need project managers to define scope and priorities. We still need senior engineers to make architectural decisions. We still need junior engineers to implement, test, and challenge assumptions. We still need testing, even if AI helps generate it faster.</p><p>Yet UX is increasingly treated as optional at best.</p><p>This is a serious mistake.</p><p>UX is not visual polish. It is not a layer applied at the end of delivery. UX is the point at which business intent, technical constraints, and human behaviour are aligned. It is where decisions are made about what a system is supposed to do from the user’s point of view.</p><p>Almost every system we build exists to support a human activity. The goal is usually to make that activity faster, more reliable, or less error-prone. UX is how we determine whether that goal is being met.</p><p>I have seen executives suggest that AI can simply be used to generate the interface. This reflects a misunderstanding of both AI and UX.</p><p>A large language model does not understand business context, user incentives, or failure modes. It does not observe real behaviour or recognise where small design decisions create friction, errors, or long-term cost. It has no accountability for outcomes.</p><p>An LLM is a predictive system trained on historical data. It produces plausible outputs. It does not perform responsibility-bearing analysis.</p><p>UX exists to deal with consequences.</p><p>When UX is done well, the results are consistent and measurable:</p><ul><li>Lower error rates.</li><li>Faster onboarding and reduced training.</li><li>Lower support volume.</li><li>Higher task completion and adoption.</li></ul><p>These outcomes directly affect operating cost, retention, and revenue.</p><p>When UX is weak or absent, the opposite occurs. Support load increases. Workarounds appear. Adoption slows. Internal productivity degrades. In customer-facing products, churn increases and conversion suffers. These costs accumulate quietly and are often misattributed to execution rather than design.</p><p>AI excels at accelerating implementation. That is its strength. But acceleration without understanding increases the rate at which mistakes are made.</p><p>When UX is reduced or removed, AI does not compensate. It amplifies the gap. Teams generate more screens, flows, and features that are internally coherent but fail in real use. In this context, “garbage in, garbage out” is simply accurate.</p><p>This is the trade-off many organisations are misjudging. By attempting to automate away UX, they remove the only discipline in the pipeline focused on human outcomes.</p><p>The result is not efficiency. It is mistakes delivered faster.</p><p>From a commercial perspective, UX is one of the worst areas to cut. The cost of experienced practitioners is small compared to the downstream cost of rework, support, retraining, and lost adoption. Preventing misunderstandings early is consistently cheaper than correcting them later.</p><p>From a leadership perspective, the risk is structural. When organisations believe tools can replace understanding, they stop managing a development process and start running a content generation pipeline, hoping outputs align with reality.</p><p>AI changes how fast software is built.</p><p>It does not change who software is built for.</p><p>We build systems for humans. If those systems do not work well for them, speed of delivery is irrelevant.</p><p>Chris Delaney is a CTO with 25 years’ experience building and scaling software systems.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cbbde8637437" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>