I tried vibe coding for 90 days

Hey honeys and hustlers,

Sidequests are the spice of life. Or, at least that’s what I tell myself when I get distracted going down a rabbit hole that is seemingly unrelated to my main work. Vibe coding is one such rabbit hole I went down last spring. When I had the idea for CommunityOS, I somehow convinced myself that I needed to build a custom website with vibe coding tools because I didn’t like the templates I saw on Framer. A very dramatic response if I do say so myself. However, I stand by my original inclination that using AI to code is much better than using AI to write and produce videos. So here’s a little story about how my vibe coding journey, and what I believe is the true value proposition of vibe coding for creators and entrepreneurs

I saw a tweet from Lovable and was immediately intrigued. I could create a custom website by just describing what I want?! Count me in! I’m not a web developer or designer, so this felt like an easy yes for me. If only things were that simple. When I tried it for the first time in April 2025, the design language and token efficiency weren’t great. It was headache-inducing, in fact. I switched to Bolt, which had a noticeably improved design language, integrations, and a more efficient token usage option (plan vs. implement), and have pretty much exclusively used them ever since. I started by creating v1 of the CommunityOS website, then I built the Melanin MVP website, and I tinkered with a few other projects that haven’t seen the light of day.

Unfortunately, Bolt (and any other vibe coding tool, I suspect) won’t tell you if an idea isn’t best suited for vibe coding. It definitely won’t tell you the idea for an app you have is straight up bad. It won’t validate product-market fit or help you find potential customers. It just does what you tell it to, most of the time. But I do like that vibe coding tools give you a playground for creating solutions to problems, which is ultimately what creative entrepreneurship is, generally speaking – creative problem-solving. So I decided to put my newfound skills to the test.

I participated in the first iteration of the Hatch-a-thon, hosted and created by Hatch Fairhope in August. Hatch Fairhope is a coworking space that gathered problems from community members, organizations, and businesses in the Fairhope and Mobile, Alabama area. Using tech to solve real problems for working people, isn’t that the great promise they use to justify this huge investment in AI?! I wanted to see if I could create something meaningful, so I signed up. I selected a problem from the list given to participants and had two weeks to use AI to workshop a tool.

The problem I selected was from a business owner who wanted help identifying which SaaS tools might be helpful for their business. Many of you ask me for recommendations for digital tools based on what I’m using, have tried, and your specific use case. The last part is the hardest but most crucial part of the recommendation. So I decided to build a database and a quiz to help solve this problem. You can get specified recommended tools based on your quiz results, and compare the tools you may already be using to alternatives on the market. You can check it out here and tell me what you think!

I earned a whopping $5 for my efforts after winning the People’s Choice Award. I spent roughly $50 on Bolt to make it. Not a great trade-off. My rudimentary idea for generating revenue from this app/tool is lackluster and unpromising. Looking at it now with fresh eyes, I would remove the log-in portal to reduce friction, as there are tons of ways to track current digital tool subscriptions. I would likely put the tool database behind an email gate and just integrate the quiz with my current Please Hustle Responsibly website (likely creating an automation to email you the results that upsells you to a paid product of some kind).

I lost money on the creation, traveled to Mobile for the event, and I never moved forward with actually integrating this tool into my business (at least, not yet). So, was it worth it?

I took a break from vibe coding after the Hatch-a-thon, and recently returned to make two prototypes/scratch a creative itch: a receipt generator for writers (because Substack writers shouldn’t be the only ones who benefit from auto-generated graphics from their articles) and a camera finder tool (eerily similar to the SaaS tool finder, but with a better monetization plan, I think/hope). Did I take inspiration from current tools to craft these? Yes. Pete created Strava receipts, and Ryan Gilbert created Workspaces, which inspired the ideas above, respectively. Could I have made these without vibe coding? No. Pete knows how to code, and Ryan originally started Workspaces on Webflow, which I can’t even use to edit a title, let alone craft a complex website.

Vibe coding allows complete beginners like me to cheaply and efficiently create a usable v1 of an idea. Who knows how long it would take me to learn how to code or use Webflow? Who knows how much it would cost me to pay someone who knew how to code or create a website in Webflow? But I do know that it took me drastically less time and cost me staggeringly less money to create both of these using Bolt.

Colin Rocker made an interesting statement on Threads:

I think creators have a unique advantage in leveraging vibe coding to create extraordinary things: personalized solutions to problems, leveraging large datasets to curate informed searches and comparisons, and creating repeatable workflows for marketing. However, I think it is a fallacy that you will create a tool, app, or website in one weekend and get rich. It is as likely as any SaaS startup raising money to become a unicorn and have a big exit. That’s not to say you shouldn’t try vibe coding. But itis worth noting that, as creators, you shouldn’t rely on the tech startup monetization path. Paid subscriptions at scale and virality are all well and good, but if you’re reading this, you’re likely looking for something a little more sustainable.

I think pairing vibe coding with real-world applications, like helping small business owners make better decisions and enhancing community building, is a better way to build products that can sustainably generate revenue. The reasons that Strava Receipts will be a long-term project and product for Pete are that he’s leveraging the everyday athlete and weekend warrior community to share their small wins digitally and their big wins at events through printables. The reason Workspaces has reportedly generated 5 figures in the last 3 months is that it pairs a weekly newsletter (with sponsors and affiliate links) with a searchable online database of workspaces (with homepage/banner sponsors and a dedicated sponsorship page with custom offers for readers/viewers). 3 sources of revenue to complement the 50K social media followers and 18K newsletter subscribers.

AI isn’t a threat to creator jobs, as this random guy (Ben Affleck) so eloquently points out, but it is a threat to audience attention. How can we leverage AI to keep audiences and communities paying attention to things that matter, solve real problems, and enhance connection? I have a feeling that vibe coding will be a part of that answer for small but dedicated creative teams.