[Iter-X] 46/100days

Day4️⃣6️⃣

I’ve adjusted the AI again. I’m planning to use Function Call’s JSON return to achieve a stable structured JSON output. Essentially, I’m tricking the AI by using a tool that doesn’t actually exist or need execution. Normally, we rely on prompts to request a JSON response, but that approach doesn’t seem to guarantee 100% stability.

I’ve also further broken down the agents into more atomic units. Now, each agent handles a very specific task, making it easier to reuse their functions. For example, I’ve separated CityPlanner and TripPlanner, so if the destination city is unknown, AI can first plan the cities before diving into the itinerary. Over the next couple of days, I’ll focus on connecting the entire itinerary module, which will bring the core functionality to life.

It’s hard to believe that nearly half of the 100-day journey has passed. The intensity of daily output is quite demanding, especially since some days don’t show obvious progress. The public is mostly visual-oriented, and I don’t want to dig into too many technical details because that would be even more tedious. Instead, I mix in insights and news occasionally. If I build in public again in the future, I might consider sharing in phases rather than posting daily updates. I don’t want the content to become a monotonous log—I’d rather deliver it when there’s something valuable to share.

Current Progress: • Prototype & UI/UX Design: 33% • Backend (Go) Development: 36% • Client (Flutter) Development: 25% • Data: 10%

If you meet the following criteria, let’s connect:

  1. Persistent
  2. Ambitious
  3. Passionate



Enjoy Reading This Article?

Here are some more articles you might like to read next:

  • zsh nice 5 🧐. How the nice value disturb the… | by ifuryst | Medium
  • TCP congestion control. Statement!!! | by ifuryst | Medium
  • [Iter-X] 45/100days
  • [Iter-X] 44/100days
  • [Iter-X] 43/100days