The following is an experiment that I conducted with the Claude 3.5 Sonnet LLM to see how well it could understand the article, synthesize a different type of content based on it, and implement a simple client-side web application. The article I used is a blog post by Glyph Lefkowitz, and I discussed this experiment in a Mastodon thread.
Note: This application has known accessibility issues, as discussed in the thread linked above. I'm leaving these issues intact to demonstrate some of the pitfalls of LLM-generated code.
The remainder of this page is unedited from the artifact generated by the LLM.
Are you more of a Nerd or a Geek?
Based on the definitions from the blog post, let's find out where you fall on the spectrum!
When watching a sci-fi movie, what's your typical reaction?
I find myself analyzing the physics and scientific accuracy of the special effects
I memorize the memorable quotes and scene details to reference later
What drives your social awkwardness more?
Noticing things others don't and having to filter my observations
Being super passionate about specific interests that others don't share
What's your approach to learning something new?
I dive deep into understanding the underlying systems and principles
I collect and memorize as many facts and details as possible
In a discussion about your interests, you're more likely to:
Explain how things work and their broader implications
Share trivia and specific details from your area of interest
What's your primary motivation when pursuing your interests?