Most of the Intentional AI series has tested AI in areas where judgment, creativity, and context matter a lot. This episode is a little different. Virgil, Cole, and returning guest Chad take a look at AI and web development, a domain where patterns, repetition, and known best practices are the whole point. On paper, that should be where AI shines.

Chad makes clear that for working developers, AI is already genuinely useful. Generating boilerplate, producing code blocks for well-understood functionality, and cutting down on time spent typing out repetitive structures. These are real wins. The catch is that getting value out of AI-generated code still requires knowing what you're looking at. If you can't read the output, you can't catch the errors, and you can't fix what's wrong.

That gap becomes more visible when you consider who these tools are being marketed to. The pitch is often aimed at business users and non-developers, promising a fast path from idea to working product. The episode digs into why that gap -- between what gets generated and what is actually usable -- is harder to close in code than it is in content or images. A piece of writing that's 80% there can be polished. Code that's 80% there can be a liability, especially if the person using it doesn't know what the other 20% is.

Virgil tested Claude, ChatGPT, and GenSpark against the same prompt: build a visually appealing, fully accessible accordion web component using the series source article. All three produced something workable. None were perfect. Claude handled screen reader accessibility well but had a JavaScript bug that prevented the drawers from opening and used a low-contrast color scheme. ChatGPT produced the most functional but visually flat result, with the worst screen reader compliance. GenSpark produced the most polished visual, with the most helpful follow-up prompts, and landed in the middle on accessibility. As Virgil put it, these were the least failures the series has generated, which is saying something.

Previously in the Intentional AI series:

Episode 1: Intentional AI and the Content Lifecycle

Episode 2: Maximizing AI for Research and Analysis

Episode 3: Smarter Content Creation with AI

Episode 4: The role of AI in content management

Episode 5: How much can you trust AI for accessibility

Episode 6: You’re asking AI to solve the wrong problems for SEO, GEO, and AEO

Episode 7: Why AI can make your content personalization worse

Episode 8: The real value of AI wireframes is NOT the wireframes

Episode 9: Just because AI can create images doesn't mean you should use them

Episode 10: The Super Bowl didn't sell AI, it exposed it

Episode 11: AI video rewards planning, not your ideas

New episodes every other Tuesday.

For more conversations about AI, design, and digital strategy, visit https://www.highmonkey.com/podcast and subscribe on your favorite podcast platform.

(0:00) - Intro

(0:44) - Today's topic: Intersection of AI & coding

(2:46) - The "just type and get a website" myth

(3:44) - Where AI actually helps with coding

(6:08) - When "good enough" works and when it doesn't

(8:13) - The real win: Using AI as a developer

(11:06) - Tool test: building an accordion with AI

(13:14) - Testing Claude

(16:23) - Testing ChatGPT (Codex)

(18:25) - Testing GenSpark

(20:25) - Using AI to fix AI code

(23:11) - Tons of opportunity with AI and coding

(24:16) - Outro

Subscribe for email updates on our website:

https://www.discussingstupid.com/

Watch us on YouTube:

https://www.youtube.com/@discussingstupid

Listen on Apple Podcasts, Spotify, or Soundcloud:

https://podcasts.apple.com/us/podcast/discussing-stupid-a-byte-sized-podcast-on-stupid-ux/id1428145024

https://open.spotify.com/show/0c47grVFmXk1cco63QioHp?si=87dbb37a4ca441c0

https://soundcloud.com/discussing-stupid

Check Us Out on Socials:

https://www.linkedin.com/company/discussing-stupid

https://www.instagram.com/discussingstupid/

https://www.facebook.com/discussingstupid

https://x.com/DiscussStupid