Web3 Badge Fix & GIGI Map Improvements — May 9, 2026

I aimed to resolve two issues that have bothered me. One of them is that the web3 badge shows incorrectly, and the other is a few peculiarities with the GIGI map. After a long 13 hour day, I managed to deliver a patch for the web3 badge and an improvement on how GIGI map interacts. Both changes are very important to the type of user experience I want to provide on my agentic web platform. My first commit was on the web3 badge. I found out that the asset for the SPACE coin badge was incorrectly shipped as a JPEG but named with a .png extension. This made some of the browsers misinterpret the image format, and hence exhibit display problems. I solved this by re-encoding the SPACE coin badge as an actual PNG file and added an image fallback. This way, if the primary image fails to load, a backup image will be available which improves badge display across different platforms. It was simple to fix this, but debugging the MIME types took way more time than I expected as I figured out the different pieces of browser behavior about rigid MIME type checking.

The second commit involving the GIGI map functionality was more complex. Users said the GIGI map orb kept showing even when the map was changed to Manual mode, which is not supposed to happen. I needed to make sure that when the AI takeover was switched off, the map orb cleared. When doing this, I decided to prioritize the new `aiCallTakeoverEnabled` flag over the stale `gigiMapOrb` in the `portalToSpatial`. I also cleared the `gigiMapOrb` in the `mergePortalDetailCache` when a PUT request changed the mode to Manual. This meant I had to recompute the `gigiMapOrb` based on the PUT response as a gigiMapOrb to replicate the behavior of GET requests. I spent a lot of time testing this because it required careful logic, and I wanted to make sure it behaved correctly in all scenarios.

While working solo on these issues and coding with Claude Code in VS Code, I found it rewarding and challenging. I could have taken a more collaborative approach with Lovable or Cursor, but I’m determined to make this one-man show with an AI team. Each bug I fix brings me closer to building a billion-dollar company. Today’s progress makes me believe I can create a better product alone and rely on AI to do the hard work.

Thinking about my whole experience, I gained some important insight. First, I learned how important thorough testing is for any user interface related features. Being able to have the web3 badge in users' trusted displays and enhancing engagement is an important consideration as I work to build out the functionalities of the agentic web.

Moreover, GIGI map fixes brought to me the importance of having clarity when it comes to understanding state management in a web application. Going forward with my development, I need to be aware of the interactions and response of various components to the actions of users. This is very important in a spatial AI because the experience of users fully depends on real-time responsiveness and reliability.

I’m proud of the work I did today. I’ll be using what I learned to improve users' experience, and the platform's overall integrity and usability. It goes without saying that I relish every step of the journey to make an agentic web platform that puts the power in users' hands. I know what I want to do next, and the challenges motivate me.