Hello!
Over the past 2 months, I've been working on my SaaS business idea that helps people find better music recommendations provided by humans rather than biased algorithms.
The key idea is very similar to a forum. You post a thread (I personally call it a REQUEST song post) and the community can reply to the post by leaving a comment or a song recommendation based on what the user is asking for (or even both a comment and song recommendation), and the rest of the community has the opportunity to agree (or disagree) by voting on each song reply. You see, there's nothing amazing about the idea itself—it seems like I'm building a subreddit forum. For that reason, I decided to pay more attention to the execution rather than the idea by delivering value to the end user. The idea here is to make this process of finding music far more comfortable and easy without forgetting this sense of human connection.
In the early days of this project, I started pre-validating my idea by searching the web for lots of people complaining about how bad algorithms (particularly Spotify's) are at delivering accurate and authentic recommendations to some users. After I noticed how this issue was affecting the way people discover new music, I began conceptualizing the website using pen and paper and Figma to design which core features were needed to implement in order to achieve an MVP.
After I finished all my digital Figma designs, it was time to make a decision about my further progress. I asked myself: Which path should I take, the low-code solution (Webflow) or vibe-coding (Lovable, Cursor AI)?
Reviewing and comparing these two options, I found using vibe-coding more attractive so I could deploy my idea faster. With progress made on Figma, I tried to transfer all these designs into Lovable so I could transform them into something more functional with an established database. Unfortunately, the Figma plugin that enables me to transfer these designs only allows me to pass each design as a new project on Lovable, which ended up with me having around 21 active projects on Lovable, making the workflow a little bit complicated. Also, this AI platform didn't fit pretty well with my requirements to develop the core features of the project. For me, it doesn't have enough infrastructure to build something greater than a landing page. With that in mind, I decided it was better to move on to another more robust and powerful platform, so I started learning how the Cursor IDE works, its AI chat discussion, and the workflows that the community shares across the internet about vibe-coding.
Now that I had defined which would be my base platform (in this case, the Cursor IDE), I tried to move each Lovable project to GitHub so I could clone each repo to my local machine and therefore work on each file, but again, I ended up dropping this because Cursor's AI agent was unable to completely understand all these 21 files I gave it as a whole project. At this point, I decided that my UI/UX figma designs could be paused for a moment (without discarding the core features, of course) so I went straight to completely starting from scratch on Cursor.
At the beginning of this endeavor of vibe-coding, I knew it was necessary to learn how to actually make better prompts and how to develop a proper workflow that makes it more agile and less likely to get constant errors from the console. For that reason, I started investing in some subscriptions to vibe-coding websites where they share you some useful resources and tips that enable you to start a little bit faster. In my case, there's a GitHub repo called memory-bank that helped me a lot to achieve what I wanted to do. It's basically a library and some custom agents that help you build whatever feature you want through a series of steps. Each step has its own custom agent specialized in their task, and you can iterate this process whenever you want to add a new feature without worrying about the context because each custom agent keeps updating the cursor rules and the .md files (where it keeps a record of the entire project details, including technical ones such as frameworks and so on).
After a couple of weeks of prompting with a lot of trial and error, I successfully finished developing and deploying (using Vercel) a prototype so I could test it with a few friends and colleagues. They brought me little feedback, yet useful feedback to improve some UX details and consider other additional features.
Currently, I'm trying to figure out what steps I should take right now when it comes to getting potential users/clients logged into the platform so I can make real progress and not fall into the trap of unnecessarily polishing the product until it reaches perfection.
Here are some thoughts and insights that I have found along this process in general.
Unless the problem was within my reach to solve by myself, I deliberately let AI solve 90% of the problems. I mean, most of the time it's helpful, of course, to let AI fix all the upcoming issues. There were some times where I told AI to take another approach I knew could be better for the situation, but there were scenarios where I just mindlessly prompted "FIX IT" until it actually fixed it. Sometimes I felt this process was kind of boring because it doesn't challenge me to actively seek solutions.
It's worth mentioning that because of the lack of challenge vibe-coding has, it might be an opportunity to develop another skill outside the scope of technical knowledge such as programming. I'm pretty sure that currently I need to start learning more about how I should sell this product.
I know that for some of you guys this affirmation sounds obvious, but before reading a book related to marketing, I was thinking of finding my potential customers right in public places in my city. The issue here is that it might be harder to find people interested in my specific product due to the fact that most of the music listeners here where I live seem to be completely satisfied with algorithms, and that's okay. I must focus on this 1% of unsatisfied people who are looking for fresh music.
And that's my overall progress so far. Feel free to comment or suggest any idea you may find insightful. Thanks for reading!
Over the past 2 months, I've been working on my SaaS business idea that helps people find better music recommendations provided by humans rather than biased algorithms.
The key idea is very similar to a forum. You post a thread (I personally call it a REQUEST song post) and the community can reply to the post by leaving a comment or a song recommendation based on what the user is asking for (or even both a comment and song recommendation), and the rest of the community has the opportunity to agree (or disagree) by voting on each song reply. You see, there's nothing amazing about the idea itself—it seems like I'm building a subreddit forum. For that reason, I decided to pay more attention to the execution rather than the idea by delivering value to the end user. The idea here is to make this process of finding music far more comfortable and easy without forgetting this sense of human connection.
In the early days of this project, I started pre-validating my idea by searching the web for lots of people complaining about how bad algorithms (particularly Spotify's) are at delivering accurate and authentic recommendations to some users. After I noticed how this issue was affecting the way people discover new music, I began conceptualizing the website using pen and paper and Figma to design which core features were needed to implement in order to achieve an MVP.
After I finished all my digital Figma designs, it was time to make a decision about my further progress. I asked myself: Which path should I take, the low-code solution (Webflow) or vibe-coding (Lovable, Cursor AI)?
Reviewing and comparing these two options, I found using vibe-coding more attractive so I could deploy my idea faster. With progress made on Figma, I tried to transfer all these designs into Lovable so I could transform them into something more functional with an established database. Unfortunately, the Figma plugin that enables me to transfer these designs only allows me to pass each design as a new project on Lovable, which ended up with me having around 21 active projects on Lovable, making the workflow a little bit complicated. Also, this AI platform didn't fit pretty well with my requirements to develop the core features of the project. For me, it doesn't have enough infrastructure to build something greater than a landing page. With that in mind, I decided it was better to move on to another more robust and powerful platform, so I started learning how the Cursor IDE works, its AI chat discussion, and the workflows that the community shares across the internet about vibe-coding.
Now that I had defined which would be my base platform (in this case, the Cursor IDE), I tried to move each Lovable project to GitHub so I could clone each repo to my local machine and therefore work on each file, but again, I ended up dropping this because Cursor's AI agent was unable to completely understand all these 21 files I gave it as a whole project. At this point, I decided that my UI/UX figma designs could be paused for a moment (without discarding the core features, of course) so I went straight to completely starting from scratch on Cursor.
At the beginning of this endeavor of vibe-coding, I knew it was necessary to learn how to actually make better prompts and how to develop a proper workflow that makes it more agile and less likely to get constant errors from the console. For that reason, I started investing in some subscriptions to vibe-coding websites where they share you some useful resources and tips that enable you to start a little bit faster. In my case, there's a GitHub repo called memory-bank that helped me a lot to achieve what I wanted to do. It's basically a library and some custom agents that help you build whatever feature you want through a series of steps. Each step has its own custom agent specialized in their task, and you can iterate this process whenever you want to add a new feature without worrying about the context because each custom agent keeps updating the cursor rules and the .md files (where it keeps a record of the entire project details, including technical ones such as frameworks and so on).
After a couple of weeks of prompting with a lot of trial and error, I successfully finished developing and deploying (using Vercel) a prototype so I could test it with a few friends and colleagues. They brought me little feedback, yet useful feedback to improve some UX details and consider other additional features.
Currently, I'm trying to figure out what steps I should take right now when it comes to getting potential users/clients logged into the platform so I can make real progress and not fall into the trap of unnecessarily polishing the product until it reaches perfection.
Here are some thoughts and insights that I have found along this process in general.
- Indeed, vibe-coding could be a menace to critical thinking. Let me explain:
Unless the problem was within my reach to solve by myself, I deliberately let AI solve 90% of the problems. I mean, most of the time it's helpful, of course, to let AI fix all the upcoming issues. There were some times where I told AI to take another approach I knew could be better for the situation, but there were scenarios where I just mindlessly prompted "FIX IT" until it actually fixed it. Sometimes I felt this process was kind of boring because it doesn't challenge me to actively seek solutions.
1.1 Maybe I should focus on another challenge while I'm working on this process of vibe-coding:
It's worth mentioning that because of the lack of challenge vibe-coding has, it might be an opportunity to develop another skill outside the scope of technical knowledge such as programming. I'm pretty sure that currently I need to start learning more about how I should sell this product.
- It is more likely to find my potential customers in internet communities rather than going outside and asking people if they need my product:
I know that for some of you guys this affirmation sounds obvious, but before reading a book related to marketing, I was thinking of finding my potential customers right in public places in my city. The issue here is that it might be harder to find people interested in my specific product due to the fact that most of the music listeners here where I live seem to be completely satisfied with algorithms, and that's okay. I must focus on this 1% of unsatisfied people who are looking for fresh music.
And that's my overall progress so far. Feel free to comment or suggest any idea you may find insightful. Thanks for reading!
Dislike ads? Become a Fastlane member:
Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)
Access Restricted: Unlock 1,000,000+ Posts
Stop Peeking Through the Keyhole.
Open the Door.
You hit a wall because the best advice isn't free—it's earned. Since 2007, MJ DeMarco has been active here daily (99.9% active rate), building a war room for entrepreneurs who refuse to settle for mediocrity.
- Active Daily: Life-changing content posted dozens of times every day. No dead threads.
- MJ's Inner Circle: Direct access to the author of The Millionaire Fastlane.
- Vetted Network: Connect with founders scaling to 7 and 8 figures.
- Proven Roadmaps: Strategy over motivation. Execution over "hustle porn."
"SCALED TO 6-FIGURE MONTHS"
"Tips and advice here saved me from huge mistakes, others allowed me to scale my business to 6-figure months."— User MitchC
"INSANE QUALITY OF PEOPLE"
"The number of high quality people I've met from this forum has been insane. All of it predicated on building real value."— User Richard Peck
"You are the average of the five people you surround yourself with."
Are you surrounding yourself with success?
