Inspiration
Our inspiration was "Brain rot", which is the term which describes the mental exhaustion and attention fragmentation caused by excessive consumption of low quality short form content. We ourselves have wasted hundreds of hours on conventional social media and have very mixed views about it. On one hand, there is an infinite stream of great ideas and creativity. But at the same time, there is a general trend of content being designed for maximum engagement and attention rather than meaningful interaction.
The goal of MakeSpace is to fill the void of authentic creative content and make social media feel social again.
What it does
It consists of combining portfolio showcase, project tracking, social sharing, and collaboration matchmaking. The platform prioritizes creation over passive content consumption, serving both as a fun social app and a credible professional portfolio.
How we built it
We started by spending a couple days debating different visions for the app. Ryan was more focused on collaboration and I was more focused on content and sharing creative progression. Ultimately, we narrowed the vision down and began by giving MeDo the high-level vision of our project. To our surprise, it understood our vision pretty well right away. It created a long list of requirements which we barely had to change. Next, we specified the UI by drawing it out and describing it to MeDo. We defined dark modes, different tabs and ratios etc. Finally, MeDo generated a version 1 of the app. It was surprisingly great! However, it would take 15+ iterations for us to reach our vision. We did this by isolating each component and prompting the AI very specifically. We didn't end up writing a single line of code! All of it was done through AI.
Challenges we ran into
This was our first time using AI to build an application and we underwent many challenges. The use of AI to build an application instead of hands-on work was difficult since the results that the AI generated often didn’t match what we had in mind. Furthermore, we had to cut down on features that we had in mind due to limited tokens and the fear that adding more features will break down other parts of the app, we had to limit our requests. The issue of limited tokens worsened when requests took multiple inputs to implement or fix rather than one input. Sometimes, the AI would hallucinate some details we didn’t ask of which we had to specify to take down. Also, the AI would repeatedly make the same mistakes when we explicitly prompted it to do something, so we had to change our approach to prompting by making more detailed by making more detailed, isolated micro-prompts, feeding the AI only one specific task.
Accomplishments that we're proud of
The goal of this hackathon was to make a proof of concept, to refine ideas, and to gauge roughly the amount of work it would involve to undertake the challenge of making a social application. Part of that goal was also to consider what frameworks or languages were needed and the technicalities involved in making the app and to gain experience from it. It is amazing that the AI could build so much in 15 iterations!
What we learned
We learned just how high AI technology has peaked since AI's explosion in popularity. With the use of MeDo, we learned just how much potential AI has in speeding up development in websites or apps to mimic popular ui designs.
What's next for MakeSpace
We hope to write MakeSpace from scratch in the future and/or also include the features that we have cut out. One of the main ideas was to make our own algorithm from scratch to encourage innovation and originality on our app. Aside from watching posts from your vicinity, the algorithm would present another feed purely for content creation but the algorithm decides to queue videos that would only be original. The goal was to give users inspiration for their own projects, which they can draw from the video and interact with it to mark that they were inspired by a certain post. One of these ideas included a monthly challenge feature to spark project ideas for users and encourage creative problem solving through various design constraints that these challenges would include. This can help users add project ideas to their portfolios, in which they can present it in a professional setting, and encourage users to collaborate with others that are within their vicinity.
Built With
- medo
- supabase
- typescript
- vite
Log in or sign up for Devpost to join the conversation.