I made my website photo follow you (and wink)
The creator made an interactive photo of themselves that follows viewers around their website and winks when clicked. They accomplished this by converting a photo into a black-and-white dither pattern, using Claude to generate six directional versions of the image, and writing code to make it interactive.
Summary
The speaker describes a creative web project where they transformed a single photograph of themselves into an interactive element for their personal website. The process began with taking one photo and converting it into a black-and-white retro dither pattern, achieving a video game aesthetic that aligned with their site's design goals. They then uploaded this reference image to Claude and used Codex, Claude's coding agent with computer use capabilities, to generate six different versions of themselves in the same pose but looking in different directions: up, right, left, down, forward, and a winking expression. Codex also handled removing the backgrounds from these generated images and wrote the necessary code to implement the interactive functionality. The final result is a small animated version of the creator that tracks the user's cursor position on the website and performs a winking animation when clicked. The entire project took approximately five minutes to complete and required only minimal token usage. The creator has made the prompts and a step-by-step tutorial available on their website so others can create their own interactive 'mini me' versions.
Key Insights
- The creator used Claude's computer use feature to generate multiple pose variations (six directional versions) from a single reference image, demonstrating how AI can automate repetitive image generation tasks
- Converting a photograph to a black-and-white dither pattern was a deliberate aesthetic choice to match the creator's desired video game look for their personal website
- Claude's Codex not only generated the images but also automatically removed backgrounds and wrote the interactive tracking and winking code, handling both visual and functional components
- The entire interactive element—from photo conversion to coded functionality—was completed in approximately five minutes, indicating significant efficiency gains through AI assistance
- The creator democratized this technique by publishing the prompts and a step-by-step tutorial online, enabling others to create similar interactive self-portraits with minimal resources
Topics
Transcript
[0:00] They say that the best art looks like it's following you around. That's why I made this photo of me that follows you around on my website and even winks when you click it. Here's how I did it. I took one photo of me and turned it into a black and white retro dither pattern that gave it this video game look that I was going for on my personal site. Then I uploaded it to Flora and had Codex, Chachi PT's coding agent, use computer use to generate six versions of me in the same pose, looking up, right, left, down, forward, and that wink. Codex then removed the backgrounds and wrote the [0:31] code so that…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
Cybersecurity Is So Bad I Texted My Parents
The speaker discusses 'Black October,' a period marked by numerous critical cybersecurity vulnerabilities and attacks affecting major internet infrastructure, including PHP, GitLab, SSL issuers, and iOS devices. While acknowledging that AI will eventually improve security, the speaker warns that cybersecurity will worsen significantly over the next 18 months before improving.
What are channels in Claude Duet?
Claude Duet features a channels system inspired by Slack that allows users to organize conversations into separate subthreads within a main chat. This enables more structured collaboration by creating dedicated channels for specific projects, such as a sales channel for sales-related work.
Claude Duet needs multiplayer (please!)
The speaker argues that Claude and similar AI products are currently optimized for single-player use and lack proper multiplayer capabilities. Despite integrations like Claude tags and Slack, the speaker contends that sharing agents with their small team of five people remains problematic, preventing these tools from becoming part of their regular workflow.
Claude Duet is great, but SLOW
The speaker expresses frustration with the slowness of AI model Claude, which impacts their willingness to engage with it. They compare its responsiveness unfavorably against more responsive alternatives, despite acknowledging Claude's capabilities.
Claude Duet should be more annoying
The speaker emphasizes the need for agents like Claude Duet to be more engaging and proactive to avoid being forgotten in the competitive landscape of digital assistants. They critique the current methods of communication, such as morning briefings, as ineffective in attracting user interest.