The Welcome Post
How I left Netflix to Launch a Startup Yet Again
To Begin Again
After another stint at Netflix, this one longer than anticipated at 8 years, I quit. Again.
I’m starting up a new business in a space I’ve been obsessed with for over 20 years now. Why am I obsessed with Semantic Layers and their ability to maximize analyst happiness? Doesn’t matter. I guess I grew to love what I’m good at.
We have lots to discuss in this new era of AI. Follow along as we build a new B2B business in the Business Intelligence space. This is not going to be the usual startup blog slop. Get the good, bad, and ugly. I’ve been a founder before and it’s not all fun and games. Subscribe and follow our journey.
This Time is Different
This will be my third time building a Semantic Layer from scratch. But this time I think I truly cracked it. It just happens that my last two and half years at Netflix was spent honing in on the right product. This was my second attempt and I was able to resolve the many issues with off the shelf BI tools like Looker, MicroStrategy, Tableau etc when it comes to self service.
Some people try to deploy Semantic Layer technologies as a panacea when the right approach is to target a much more narrow use case. I will say more on this in future posts.
Why I Left Netflix
If you’re in the tech biz working for any of the big names, you know Netflix’s has the golden handcuffs of golden handcuffs. (maybe Meta now for AI stuff) Needless to say it’s extremely hard to give all that up for no salary :). You have to be a little crazy.
Hard to do anything risky if you don’t have a little crazy in you.
Anyways, as I alluded earlier I did build a BI tool with a powerful Semantic Layer at Netflix. It is currently deployed in production for several projects. In fact, I kicked off a sort of internal race to build Semantically enabled solutions. When I started working on it people were just getting reintroduced to Semantic BI tooling. In fact, it was once the default way for analysts to access data at Netflix. But, abandoned due to many reasons worth a post of its on. Long story short, the leading offerings were too hard to implement and maintain as the company grew.
Back to the main plot. I made a wild choice to win the race. I went against the standard nodejs, python, or java stack and chose Ruby on Rails. It was once popular at Netflix but now banned. Persona non grata!
Despite the end users loving and adopting the tool faster than ever, its success was being politically limited. Teams of many more developers were pushing promises while my solo developed solution was already in production. My choice of Ruby on Rails being off the paved path was the argument against further adoption.
It is a fair argument but we all need to veer of paved paths at some points to discover new possibilities. With the the Ruby on Rails renaissance we are seeing, I’ll be proven right for eventually. Nevertheless, this was also my cue to leave.
On Startups
Startups are hard. Startups are fun. But, Startups are hard.
This time I’m going to launch an LLC and delay funding til I’m satisfied that we are at or near product market fit. Maybe we’ll never raise. We are in a new tech epoch with AI. Lots to learn and innovate in terms of how you build a business.
I’m talking to several potential cofounders. Almost had a person quit and join me for no money just equity only to vanish suddenly. The ride starts immediately and before you know it your going down and then up. It’s thrilling!
With AI maybe Solo is the right move.
Why Ruby on Rails
If you made it this far, you probably guessed I love Ruby on Rails. My philosophy in building data tools for the past 20 years aligns well with Rails philosophy. Namely “complexity compression”. Having worked on internal tools at large enterprises it feels like complexity in every domain is increasing. Whether its deployment tools, data tools, web development frameworks, it doesn’t matter. In fact, I think large corporate management teams are incentivized towards complexity. Even with AI they will chose the complicated solutions over the simple ones at every turn. I bet there is a Dilbert comic strip showcasing this.
In anycase, Ruby on Rails since version 7 is experiencing a well deserved renaissance. Hotwire with Stimulus is a great answer to building modern Web UI’s. Even still you can easily add React or some other frontend framework. I think we will see many more startups choosing Ruby on Rails in the coming years.
At Netflix, I was just a bit ahead of trend. In time, my former colleagues will come around.