I got fired because Cursor admins can see token usage in private repos

Hello, World!

Since 2017, I've tried to build something of my own 7 times. Four of those attempts were in 2025 (thanks to LLMs and vibe-coding). Two attempts I can call successful. The rest failed with 0 users, or are sitting on my local machine with no public deployment.

My first successful project (attempt #2) was a Google Sheets Add-on for Investment Fund Management and Portfolio Tracking. I just got lucky with timing in 2017, when the crypto hype was starting and there were no tools or apps around it yet. When you hit the right timing, everything's easy—I built a tool for myself in a couple of days to solve my own pain point, and friends immediately wanted to buy it. I had a small Telegram channel on this topic (about 1,500 subscribers), and got 20 sales after the first post. I was literally selling a Google spreadsheet, processing payments manually. It was cool, but I initially didn't believe in the project's potential and didn't invest much in its development. Plus, I didn't have enough engineering skills at the time to turn it into a web app or proper SaaS. The project seemed somehow... not serious. So it gradually died, though it ran for 8 years (completely zero cost to maintain). Total revenue over 8 years was about $12k, which equals $142/month on average.

The second successful project is what I'm working on full-time now, and this time I think I've found "the one". I like how it's built and what value it brings to people. I like that it's a pure B2C SaaS. Overall, this is the project that fits me best so far. But I'm not yet sure if it has enough money-making potential, though I know it does because I've already gotten my first 3 sales from strangers with cold traffic from Telegram channels. I'll reveal the project idea later in this post, but first I want to talk about my development, launch, and most importantly, marketing experience.

Finding the Idea

I kept having thoughts about building something of my own, but this time something complete. Because now with LLMs it's super easy. But of course, you need an idea. A good, powerful, inspiring idea. And I was looking for it, brainstorming every day, talking with ChatGPT, but to no avail.

And when it seemed like I'd tried everything and wanted to just give up on "the idea of solo startup", I remembered something I was interested in exploring purely for myself and thought maybe I should try building around it. This time I didn't have initial excitement, but the idea captured me gradually. On November 19, 2025, I started development with wonderful expectations (thanks ChatGPT—you're the best at pouring maximally optimistic forecasts).

How I Got Fired

I had a great calm job at a wonderful company earning $6k/month. Short story of how I got fired: I was actively building my project through corporate Cursor (yes yes, very stupid), Cursor admin can see how many tokens you spent and on which project (your company admin sees all repositories names where you spent tokens in Cursor, even private ones), so they showed me that I was spending 95% of tokens on my project and 5% on main work. So they fired me fairly and justly. I had no fear; on the contrary, I felt surprisingly good. Probably because my "double life" was finally over.

So I got fired at the moment when the project was ready and deployed, domain configured, server security audit done. I was preparing to launch. I set a launch date 3 days after termination (this is important—set a clear date when you're starting your own business), bought wine to celebrate with my wife, bought delicious food. But fuck, I didn't buy any ads... My launch plan was: post 3 shorts on four social networks and tell all my friends about the project. I had no media presence. What a naive fool I was. I thought I'd sit in the evening drinking wine with my wife and enjoying notifications about new users. Here's tip #1:

Tip: "On launch day, if you don't have a media presence, buy ads from bloggers for sure! Otherwise you'll just sit and celebrate in silence."

By the way, your friends might not give a damn about your project or that it's important to you. Be ready for this 🙂

The next morning I woke up and frantically started thinking about where to quickly buy ads. The best thing I came up with was Telegram channels. My audience is divided into two language segments: Russian-speaking and English-speaking. The first is closer and more familiar to me, so I went there. I messaged seven channel owners in my niche, three simply ignored me, I made contact with one and boom—ad was on the channel by evening! I paid $25 for a post in a 5k subscriber channel. And got my first users! About 20 registrations total and 0 sales. At first this disappointed me, but then after buying ads on 7 other channels, I realized this was the best result in terms of registrations. What was important about this ad: the post was native, from the channel author, she tried the product and described her experience—probably that's what worked.

Insight #1: "Better to spend a bit more money for a blogger to try your product and make a native ad. It's 5-10x more effective"

Insight #2: "There are a lot of scam and bot-farm channels on Telegram. They skillfully disguise themselves, including inflating reactions. I got 2 clicks to the site from 2k views. This is abnormally low, considering the same post on another channel got 30 clicks from 2k views. There are entire farms of bot channels run by AI—you can see it in description patterns. Avoid them!"

Insight #3: "Don't rush to launch ads without analytics! Set up simple cookie-free analytics through Plausible—takes 20 minutes to configure. It doesn't require any cookie banners and complies with all laws. This way you can at least minimally track by UTM tags where traffic comes from and how it converts!"

I knew my product was good because I got great feedback from friends I gave it to for free. I knew the price was quite small and justified, there were registrations, but no sales.

I just wanted to throw money at a Telegram ad exchange to finally get that first sale. In total, I spent $150 blindly without much channel selection before it finally happened.

First Sale

Does everyone's first sales turn out strange, crooked, and unexpected? Mine was exactly like that: I got a purchase from someone where everything went wrong. The thing is, my product first shows a free preview, then sells the full version, but the client hit a bug and the preview didn't show at all—just a "Buy Full Version" button, and the client bought.

The second sale went through the normal flow, but the user couldn't access their account to get the product (because I didn't have a redirect there after magic link login), just couldn't figure it out. Thank God he wrote to support and I sent him the product personally.

The third sale finally happened as I intended, and the client got the product.

After the first sale—it was like an emotional weight lifted off me.

Monthly burn:

• Server: $25 (bought for a year)

• Third-party API: $59

• Google Workspace for three emails: $21

Unit economics:

• OpenAI and Gemini API per one product: $0.5

• Product price: $10

Total:

• Spent with ad costs: $320

• Earned: $30 + $40 (from direct sales) = $70

Profit: -250$, I am a businessman now 😅

What's the product anyway?

Now about the product idea. I deliberately left this for the end of the post because I'm afraid the idea might put some people off—it's, let's say, a bit "esoteric".

Basically, it's a service that automatically uses AI to generate a self-discovery book based on a person's date, place, and time of birth.

Based on this data, you can calculate quite a lot: natal chart, Vedic astrology, numerology, Human Design, Jung archetypes, and Chinese zodiac. The main idea is to synthesize interpretations of these knowledge systems and use LLM to create an interesting and engaging book about yourself.

As a result, the user gets a beautiful 80-100 page PDF/EPUB with a unique cover.

I want to say this is not a horoscope or prediction. The product doesn't play on people's fears or pains and doesn't promise to solve any problems. It's just entertaining, interesting reading about yourself that leaves positive emotions and "reinforces" good thoughts and insights about yourself. This is perhaps both its strength and weakness. It's called "SoulBook" - if you are interested link in my profile.

What's next?

It's been exactly 5 days since launch. I have 60 users, 3 sales to strangers, and 4 sales to friends directly (you could call them test sales). I know the product works and brings positive emotions to people. I'm happy about every new user. But I have no idea where to get traffic so my numbers add up. I'm not a marketer (at least not yet).

Current plan: post reels videos every day for 30 days on TikTok, Instagram, YouTube, and Pinterest + try to buy one native ad from an Instagram blogger. What do you think?
Maybe you can suggest another strategy?

P.S. Would love to connect!

If you have experience with B2C marketing in esoterics/self-discovery, or just want to discuss solo-founding—DM me. Always open to exchanging experience and collaborations.

Also, if you're interested in trying SoulBook—I'll give you a book for free in exchange for feedback.

Thanks for reading to the end! 🚀