Check it out on the Chrome Web Store bro!
I spent about three days building a LinkedIn Net Worth Calculator. The idea was inspired by another viral mini app i saw, the Instagram Body Count Calculator.
Under one of the posts someone jokingly suggested a LinkedIn version and I thought fuck it let me cook something up.
Originally the plan was to make it completely RNG. Generate a random but plausible number and show it as a profile’s net worth. But the more I worked on it, the more I realized there were actually a few real data points you can extract from LinkedIn that make the estimate slightly less ridiculous.
The main limitation is that LinkedIn only really gives you career data. I can see someone’s job history, company, role, and tenure. What I can’t see is whether someone:
- started a company on the side
- inherited money
- made a fortune trading crypto in 2017
- sells drugs
Those things create massive outliers, and there’s no reliable way to account for them using LinkedIn data alone.
So the calculation mostly assumes a career-based trajectory.
Tech
Technically the project is pretty simple. There are two components.
The frontend is just a chill Chrome extension.
All it does is parse a user's work history from the page and send it to the backend. The backend then runs some calculations to estimate a net worth.
For the backend I used Cloudflare Workers.
Cloudflare Workers have a very generous free tier and are perfect for tiny services like this. You can also attach a database easily, so I used that to store cached results.
High level flow:
- User visits a LinkedIn profile
- Extension extracts job history
- Sends it to the worker
- Worker checks cache
- If not cached, compute estimate
- Store result and return it
There is no good reason for the backend to exist. This can all be done client side. I just wanted to set up a backend tbh
Vibe Coding vs Actually Coding
I built most of this using Codex.
The whole thing took about 3 days, maybe 9 hours total of actual work (not counting this writeup or waiting for Chrome to approve the extension).
I’d estimate the project was roughly:
85% vibe coding
15% actual coding
Lately I’ve been trying to vibe code most of my projects but there's still a big gap between AI assisting and AI actually being the driver. My goal is to be able to drive somewhere, spin up a voice chat with Codex and by the time my drive ends I have an app spun up. My process is not good enough for this yet tho.
Looking back, I could probably have built this in a single day if I just sat down and implemented it myself. the core problem is actually very simple. the model really needed was one equation.
net_worth ≈ role
× company_tier
× salary_estimate
× years_worked
You can add modifiers like:
- company tier (FAANG vs random startup)
- role seniority
- time spent in each role
- geography
But fundamentally it’s just one equation.
What I’d Do Differently
If I were starting this project again I would define a single canonical equation and enforce it with a test.
net_worth = f(role, company_tier, years, location)
I didn’t realize this would become the core algorithm until a bit of iteration. Once I understood that, I really should have scrapped the implementation and restarted cleanly. Had i done that i could have likely killed 70% of the code while increasing accuracy.
Also a really big thing. I dont expect this to go viral simply because the friction to download a chrome extension is too great. I couldnt figure out how to scrape Linkedin info without a login (and I dont want my real account banned!) but if someone figures it out please reach out to me! I think this goes viral if its a simple webapp, no download required.
Writing This Post
I am terrible at writing.
The way this post was written was basically:
- I dictated my thoughts into a document
- fed that to GPT
- asked it to clean up the flow
- Did a pass and killed some of the AI smell leftover
It took a few revisions but the result is okay. Was it obvious? Let me know