<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Lukas Rodrigo</title>
    <description>The latest articles on DEV Community by Lukas Rodrigo (@lukasrar).</description>
    <link>https://dev.to/lukasrar</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4250758%2Ff5253fb9-3f39-41b5-a804-ba1bca211487.jpg</url>
      <title>DEV Community: Lukas Rodrigo</title>
      <link>https://dev.to/lukasrar</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/lukasrar"/>
    <language>en</language>
    <item>
      <title>Beyond the Prompt: Why Your AI Chose a Spinner Instead of a Skeleton</title>
      <dc:creator>Lukas Rodrigo</dc:creator>
      <pubDate>Fri, 20 Nov 2026 17:45:57 +0000</pubDate>
      <link>https://dev.to/lukasrar/beyond-the-prompt-why-your-ai-chose-a-spinner-instead-of-a-skeleton-23oc</link>
      <guid>https://dev.to/lukasrar/beyond-the-prompt-why-your-ai-chose-a-spinner-instead-of-a-skeleton-23oc</guid>
      <description>&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1e40tkdy7tg91qx2b4zn.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1e40tkdy7tg91qx2b4zn.png" alt="AI-assisted software development illustrating the engineering decisions behind generated code" width="800" height="320"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The real work of an AI agent doesn’t necessarily begin when you press Enter.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;"Create a loading component."&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;That's it.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;One sentence.&lt;/p&gt;

&lt;p&gt;No extra context.&lt;/p&gt;

&lt;p&gt;No design guidelines.&lt;/p&gt;

&lt;p&gt;No UX requirements.&lt;/p&gt;

&lt;p&gt;No product description.&lt;/p&gt;

&lt;p&gt;Yet an AI coding agent somehow returns a complete component.&lt;/p&gt;

&lt;p&gt;Maybe it's a spinner.&lt;br&gt;&lt;br&gt;
Maybe it's a skeleton screen.&lt;br&gt;&lt;br&gt;
Maybe it's a progress bar.&lt;br&gt;&lt;br&gt;
Maybe it doesn't render anything at all.&lt;/p&gt;

&lt;p&gt;At first glance, it feels like magic.&lt;/p&gt;

&lt;p&gt;But the interesting part isn't the component that was generated.&lt;/p&gt;

&lt;p&gt;The interesting part is everything the AI decided before writing the first line of code.&lt;/p&gt;

&lt;p&gt;That's what this series is about.&lt;/p&gt;

&lt;p&gt;Welcome to Beyond the Prompt.&lt;/p&gt;


&lt;h2&gt;
  
  
  The Prompt Is Only the Beginning
&lt;/h2&gt;

&lt;p&gt;When people talk about AI-assisted development, they usually focus on prompts.&lt;/p&gt;

&lt;p&gt;"Prompt engineering is the future."&lt;br&gt;&lt;br&gt;
"Just write better prompts."&lt;/p&gt;

&lt;p&gt;But prompts are only inputs.&lt;/p&gt;

&lt;p&gt;The real engineering happens after the model receives them.&lt;/p&gt;

&lt;p&gt;An autonomous coding agent doesn't simply autocomplete code.&lt;/p&gt;

&lt;p&gt;It starts making decisions.&lt;/p&gt;

&lt;p&gt;Lots of them.&lt;/p&gt;

&lt;p&gt;Some are obvious.&lt;br&gt;&lt;br&gt;
Most are invisible.&lt;/p&gt;

&lt;p&gt;For something as simple as a loading component, the agent has to answer questions that the user never asked.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Is this page loading once or continuously?&lt;/li&gt;
&lt;li&gt;Is there existing UI that should remain visible?&lt;/li&gt;
&lt;li&gt;Is the loading duration expected to be milliseconds or several seconds?&lt;/li&gt;
&lt;li&gt;Does the application already have a design system?&lt;/li&gt;
&lt;li&gt;Is accessibility required?&lt;/li&gt;
&lt;li&gt;Is the component reusable?&lt;/li&gt;
&lt;li&gt;Is this mobile?&lt;/li&gt;
&lt;li&gt;Desktop?&lt;/li&gt;
&lt;li&gt;Dashboard?&lt;/li&gt;
&lt;li&gt;Form?&lt;/li&gt;
&lt;li&gt;Table?&lt;/li&gt;
&lt;li&gt;Search results?&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;None of these questions were in the prompt.&lt;/p&gt;

&lt;p&gt;Yet every answer changes the implementation.&lt;/p&gt;


&lt;h2&gt;
  
  
  The First Decision: What Kind of Loading Is This?
&lt;/h2&gt;

&lt;p&gt;One of the earliest choices an AI agent makes is identifying what is actually loading.&lt;/p&gt;

&lt;p&gt;Not every loading state represents the same user experience, which means not every situation deserves the same loading component.&lt;/p&gt;
&lt;h3&gt;
  
  
  Fetching an Entire Page
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F7j876r5rpey682rt8nr0.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F7j876r5rpey682rt8nr0.png" alt="Skeleton screen showing placeholder content while an entire page loads" width="376" height="337"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;A Skeleton is usually the best choice.&lt;/p&gt;

&lt;p&gt;The user is waiting for content that hasn't been rendered yet, so showing the page structure helps reduce uncertainty and makes the application feel faster.&lt;/p&gt;
&lt;h3&gt;
  
  
  Waiting for a Button Action
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F7ewdhrafxblg5sm82rqc.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F7ewdhrafxblg5sm82rqc.png" alt="Button with an inline spinner indicating that an action is being processed" width="393" height="318"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;A Spinner inside the button is often enough.&lt;/p&gt;

&lt;p&gt;The rest of the interface is already available. The user only needs confirmation that their action is being processed.&lt;/p&gt;
&lt;h3&gt;
  
  
  Uploading a File
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ftgh0y6fotmiylnz1z9vp.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ftgh0y6fotmiylnz1z9vp.png" alt="Progress bar showing the progress of a file upload" width="432" height="253"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;A Progress Bar is the better option.&lt;/p&gt;

&lt;p&gt;Unlike a spinner, it tells users how much work remains, which becomes increasingly important as the operation takes longer.&lt;/p&gt;
&lt;h3&gt;
  
  
  Streaming an AI Response
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fsc605c5bxqztrld8vpu3.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fsc605c5bxqztrld8vpu3.png" alt="AI chat interface displaying a response as it is generated" width="371" height="293"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;In many cases, the best loading state is no traditional loader at all.&lt;/p&gt;

&lt;p&gt;Displaying the content as it's generated creates a more natural experience than making users wait for the complete response.&lt;/p&gt;
&lt;h3&gt;
  
  
  Refreshing Data in the Background
&lt;/h3&gt;

&lt;p&gt;Sometimes the best loading component is none.&lt;/p&gt;

&lt;p&gt;If users can continue interacting with the application, a subtle indicator - or no indicator at all - often creates a smoother experience than interrupting the interface with a visible loader.&lt;/p&gt;

&lt;p&gt;Notice something.&lt;/p&gt;

&lt;p&gt;The loading component isn't chosen because it looks better.&lt;/p&gt;

&lt;p&gt;It's chosen because it communicates a different kind of progress.&lt;/p&gt;

&lt;p&gt;A spinner says:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Something is happening.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;A skeleton says:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;The content is coming, and this is roughly where it will appear.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;A progress bar says:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;We know how much work is left.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;These aren't just different UI components.&lt;/p&gt;

&lt;p&gt;They're different ways of communicating with the user.&lt;/p&gt;


&lt;h2&gt;
  
  
  Why Skeletons Often Feel Faster
&lt;/h2&gt;

&lt;p&gt;One of the biggest UX discoveries over the last decade is that perceived performance matters almost as much as actual performance.&lt;/p&gt;

&lt;p&gt;Imagine waiting three seconds.&lt;/p&gt;
&lt;h3&gt;
  
  
  Version A
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Frwenfibrbxirvcjrnfeq.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Frwenfibrbxirvcjrnfeq.png" alt="Version A illustrating a loading experience without skeleton placeholders" width="799" height="149"&gt;&lt;/a&gt;&lt;/p&gt;
&lt;h3&gt;
  
  
  Version B
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fn4bd5wp3kkpwdaluovph.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fn4bd5wp3kkpwdaluovph.png" alt="Version B illustrating a loading experience with skeleton placeholders" width="800" height="188"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Still three seconds.&lt;/p&gt;

&lt;p&gt;Yet users consistently report the second experience as faster.&lt;/p&gt;

&lt;p&gt;Nothing changed technically.&lt;/p&gt;

&lt;p&gt;Everything changed psychologically.&lt;/p&gt;

&lt;p&gt;Skeleton screens reduce uncertainty.&lt;/p&gt;

&lt;p&gt;They tell users:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;"The page already exists. We're just filling it in."&lt;/p&gt;
&lt;/blockquote&gt;
&lt;h2&gt;
  
  
  Why an Agent Might Ignore Skeletons
&lt;/h2&gt;

&lt;p&gt;Now here's something interesting.&lt;/p&gt;

&lt;p&gt;Many developers assume that skeletons are always better.&lt;/p&gt;

&lt;p&gt;They're not.&lt;/p&gt;

&lt;p&gt;Imagine clicking a button that triggers a network request lasting 400 milliseconds.&lt;/p&gt;

&lt;p&gt;Rendering a complex skeleton for less than half a second creates unnecessary visual noise.&lt;/p&gt;

&lt;p&gt;An AI agent evaluating this scenario might decide:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Spinner&lt;/li&gt;
&lt;li&gt;No animation&lt;/li&gt;
&lt;li&gt;Optimistic update&lt;/li&gt;
&lt;li&gt;Keep previous content visible&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;instead of replacing the UI with placeholders.&lt;/p&gt;

&lt;p&gt;Again, the decision isn't aesthetic.&lt;/p&gt;

&lt;p&gt;It's contextual.&lt;/p&gt;


&lt;h2&gt;
  
  
  Context Changes Everything
&lt;/h2&gt;

&lt;p&gt;Suppose the prompt is still exactly the same.&lt;/p&gt;

&lt;p&gt;Create a loading component.&lt;/p&gt;

&lt;p&gt;Now imagine four different applications.&lt;/p&gt;
&lt;h3&gt;
  
  
  Banking App
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fqe3386jypzaib1wbpvxw.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fqe3386jypzaib1wbpvxw.png" alt="Banking application interface illustrating a loading state in a financial environment" width="752" height="502"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Trust matters.&lt;/p&gt;

&lt;p&gt;Aggressive animations can make financial operations feel unreliable.&lt;/p&gt;

&lt;p&gt;A subtle progress indicator is usually preferable.&lt;/p&gt;


&lt;h3&gt;
  
  
  Social Media Feed
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1z50uie3brxw9c9n9ek4.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1z50uie3brxw9c9n9ek4.png" alt="Social media feed displaying skeleton cards while content loads" width="752" height="502"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Content is consumed continuously.&lt;/p&gt;

&lt;p&gt;Skeleton cards preserve layout stability.&lt;/p&gt;


&lt;h3&gt;
  
  
  File Upload
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F9stpjx00z4d5f076u497.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F9stpjx00z4d5f076u497.png" alt="File upload interface showing upload progress" width="752" height="502"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Users care about completion.&lt;/p&gt;

&lt;p&gt;A progress percentage is far more valuable than an infinite spinner.&lt;/p&gt;


&lt;h3&gt;
  
  
  AI Chat
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fs69ne0m7ea5iwxu026y5.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fs69ne0m7ea5iwxu026y5.png" alt="AI chat interface illustrating token-by-token response streaming" width="752" height="502"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The response is generated token by token.&lt;/p&gt;

&lt;p&gt;The best loading state may be…&lt;/p&gt;

&lt;p&gt;No loading state at all.&lt;/p&gt;

&lt;p&gt;Streaming content often feels more natural than waiting for completion.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Same prompt.&lt;br&gt;&lt;br&gt;
Four different answers.&lt;/strong&gt;&lt;/p&gt;


&lt;h2&gt;
  
  
  Invisible Constraints
&lt;/h2&gt;

&lt;p&gt;Modern coding agents rarely work in an empty project.&lt;/p&gt;

&lt;p&gt;They're usually analyzing an existing codebase.&lt;/p&gt;

&lt;p&gt;That changes everything.&lt;/p&gt;

&lt;p&gt;Before writing a loading component, an agent may inspect:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Existing design system&lt;/li&gt;
&lt;li&gt;UI library&lt;/li&gt;
&lt;li&gt;Component naming conventions&lt;/li&gt;
&lt;li&gt;Theme tokens&lt;/li&gt;
&lt;li&gt;Accessibility helpers&lt;/li&gt;
&lt;li&gt;Animation utilities&lt;/li&gt;
&lt;li&gt;Existing Loader component&lt;/li&gt;
&lt;li&gt;Storybook documentation&lt;/li&gt;
&lt;li&gt;ESLint rules&lt;/li&gt;
&lt;li&gt;TypeScript configuration&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If the project already contains:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight jsx"&gt;&lt;code&gt;&lt;span class="o"&gt;&amp;lt;&lt;/span&gt;&lt;span class="nc"&gt;LoadingSpinner&lt;/span&gt; &lt;span class="n"&gt;size&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s2"&gt;"md"&lt;/span&gt; &lt;span class="o"&gt;/&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;creating another spinner component would actually be a bad decision.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The best code is often the code that isn't written.&lt;/strong&gt;&lt;/p&gt;




&lt;h2&gt;
  
  
  The Accessibility Decision
&lt;/h2&gt;

&lt;p&gt;Loading indicators aren't only visual.&lt;/p&gt;

&lt;p&gt;Screen readers also need to understand what's happening.&lt;/p&gt;

&lt;p&gt;An experienced engineer immediately thinks about:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight html"&gt;&lt;code&gt;aria-busy="true"
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;or&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight html"&gt;&lt;code&gt;role="status"
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;or announcing completion once loading finishes.&lt;/p&gt;

&lt;p&gt;An AI agent trained on production-quality patterns may include these automatically.&lt;/p&gt;

&lt;p&gt;Not because the prompt requested accessibility.&lt;/p&gt;

&lt;p&gt;Because accessibility is part of good engineering.&lt;/p&gt;




&lt;h2&gt;
  
  
  Performance Is Also a Design Decision
&lt;/h2&gt;

&lt;p&gt;Performance Is Also a Design Decision&lt;/p&gt;

&lt;p&gt;Suppose your skeleton consists of twenty animated placeholders.&lt;/p&gt;

&lt;p&gt;Each uses gradients.&lt;/p&gt;

&lt;p&gt;Each animates indefinitely.&lt;/p&gt;

&lt;p&gt;Looks beautiful.&lt;/p&gt;

&lt;p&gt;Now place fifty of them inside a virtualized table.&lt;/p&gt;

&lt;p&gt;Suddenly your loading animation becomes one of the most expensive parts of the page.&lt;/p&gt;

&lt;p&gt;Good loading components should disappear quickly.&lt;/p&gt;

&lt;p&gt;That doesn't mean they should be expensive while they're alive.&lt;/p&gt;




&lt;h2&gt;
  
  
  Design Systems Change the Question
&lt;/h2&gt;

&lt;p&gt;In mature engineering organizations, developers rarely ask:&lt;/p&gt;

&lt;p&gt;"Should we use a spinner or a skeleton?"&lt;/p&gt;

&lt;p&gt;That decision has often already been made.&lt;/p&gt;

&lt;p&gt;A well-designed Design System defines when each feedback pattern should be used, creating consistency across the entire product.&lt;/p&gt;

&lt;p&gt;For example, the guidelines might look something like this:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Component&lt;/th&gt;
&lt;th&gt;When to use&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Spinner&lt;/td&gt;
&lt;td&gt;When the duration of an operation is unknown.&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Skeleton&lt;/td&gt;
&lt;td&gt;When replacing content that hasn't loaded yet.&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Progress Bar&lt;/td&gt;
&lt;td&gt;When the application can measure how much work remains.&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Linear Progress&lt;/td&gt;
&lt;td&gt;For background operations that shouldn't interrupt the user.&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Toast&lt;/td&gt;
&lt;td&gt;To communicate that an operation has completed successfully.&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;These aren't hard rules.&lt;/p&gt;

&lt;p&gt;They're conventions that help every team build a consistent user experience.&lt;/p&gt;

&lt;p&gt;An AI agent working inside an established codebase isn't inventing a loading pattern from scratch.&lt;/p&gt;

&lt;p&gt;Instead, it's looking for the product's existing language.&lt;/p&gt;




&lt;p&gt;If the Design System already defines that lists always use skeletons, generating a spinner would actually be the wrong decision - even if it technically works.&lt;/p&gt;

&lt;p&gt;This is one of the biggest differences between writing code and building products.&lt;/p&gt;

&lt;p&gt;The correct answer isn't always the one that solves the problem.&lt;/p&gt;

&lt;p&gt;It's the one that solves the problem the same way the rest of the product does.&lt;/p&gt;

&lt;p&gt;Understanding those conventions often matters more than understanding React itself.&lt;/p&gt;




&lt;h2&gt;
  
  
  What Actually Happened Inside the Agent?
&lt;/h2&gt;

&lt;p&gt;From the outside, it looks like this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;Prompt
  ↓
Code
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Reality is much closer to this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;Prompt
  ↓
Understand intent
  ↓
Identify user experience
  ↓
Inspect project conventions
  ↓
Find reusable components
  ↓
Evaluate accessibility
  ↓
Evaluate performance
  ↓
Choose loading strategy
  ↓
Generate implementation
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The code is almost the final step.&lt;/p&gt;




&lt;p&gt;The interesting work happened long before.&lt;/p&gt;




&lt;h2&gt;
  
  
  How to Apply This Today
&lt;/h2&gt;

&lt;p&gt;The next time you ask an AI agent to build something, don't stop at the generated code.&lt;/p&gt;

&lt;p&gt;Instead, ask questions like:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Why did you choose a spinner?&lt;/li&gt;
&lt;li&gt;Why not a skeleton?&lt;/li&gt;
&lt;li&gt;What alternatives did you consider?&lt;/li&gt;
&lt;li&gt;What assumptions did you make?&lt;/li&gt;
&lt;li&gt;What information was missing from my prompt?&lt;/li&gt;
&lt;li&gt;How would this change for a mobile app?&lt;/li&gt;
&lt;li&gt;Would your decision be different if this were a design system component?&lt;/li&gt;
&lt;/ul&gt;




&lt;p&gt;These follow-up questions expose the reasoning behind the implementation.&lt;/p&gt;

&lt;p&gt;And that's often more valuable than the code itself.&lt;/p&gt;

&lt;p&gt;The goal isn't to make the AI justify every line.&lt;/p&gt;

&lt;p&gt;It's to understand the trade-offs that led to the solution.&lt;/p&gt;

&lt;p&gt;That's how you move from using AI as an autocomplete tool to collaborating with it as an engineering partner.&lt;/p&gt;




&lt;h2&gt;
  
  
  Final Thoughts
&lt;/h2&gt;

&lt;p&gt;A loading component seems trivial.&lt;/p&gt;

&lt;p&gt;It's one of the first things most frontend developers learn to build.&lt;/p&gt;

&lt;p&gt;But it's also a perfect example of something larger.&lt;/p&gt;

&lt;p&gt;The prompt may contain only seven words:&lt;/p&gt;

&lt;p&gt;"Create a loading component."&lt;/p&gt;

&lt;p&gt;Yet those seven words trigger dozens of engineering decisions.&lt;/p&gt;

&lt;p&gt;Some relate to UX.&lt;br&gt;&lt;br&gt;
Others to accessibility.&lt;br&gt;&lt;br&gt;
Some to performance.&lt;br&gt;&lt;br&gt;
Others to consistency with an existing design system.&lt;/p&gt;

&lt;p&gt;The component you see is only the visible result of those decisions.&lt;/p&gt;

&lt;p&gt;The real work happened long before the first return ( appeared in the editor.&lt;/p&gt;

&lt;p&gt;That's what Beyond the Prompt is really about.&lt;/p&gt;

&lt;p&gt;Not the prompts we write.&lt;/p&gt;

&lt;p&gt;But the engineering decisions that happen after them.&lt;/p&gt;

&lt;p&gt;Because once you start seeing those invisible decisions, you'll realize something important:&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Great AI-assisted development isn't about getting code faster. It's about understanding why that code was the right choice in the first place.&lt;/strong&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>softwareengineering</category>
      <category>webdev</category>
      <category>ux</category>
    </item>
  </channel>
</rss>
