I wrote this a couple of days ago, and I became interested in trying to explain it myself, so I "back tracked" SupaBase back to its source trying to make sense of it, figuring at least I could learn something in the process.
First things first, for as long as I have maintained Hyperlambda, the primary argument against it was that, and I quote; "Hyperlambda is a single person programming language, and the risk is simply too large for us to use it".
Ignoring the fact that it's got 21x as many downloads as SupaBase, if you're counting NuGet downloads, a pattern rapidly started to emerge.
SupaBase is entirely based upon PostgREST. In fact, PostgREST is arguably 49% of their value proposition according to their own website. The other 49% is PostgreSQL. PostgreSQL of course is a super mature database, and some would argue the best RDBMS on the planet, so let's ignore that part for a moment, and consider it a mature thing and move on to PostgREST.
According to the PostgREST GitHub repository there are 184 contributors to the project. However, if you start analysing it in more depth, you rapidly understand that out of these 184 people, 183 of them did roughly 50% of the work. The rest was literally done by one single human being.
If you click the above link, you will also notice that PostgREST is created in Haskell. According to the TIOBE index, Haskell is used by 0.37% of all software developers in the world. In fact, there are almost 3 times as many software developers working in Classic Visual Basic than there are software developers coding in Haskell. For crying out loud, FoxPro is more popular. FoxPro haven't been actively maintained by its owner for a couple of decades.
Classic Visual Basic, which was made obsolete in 2001, is almost 3 times as large counting their "community" as Haskell
Even Lisp, Julia, and Lua are larger than Haskell, 3 programming languages 99% of all junior software developers probably haven't even heard about.
More interestingly, if you ignore "begriffs" in the following link, you will see that for all practical concerns "steve-chavez" is more or less single handedly maintaining PostgREST. Begriffs seemed to quit working on PostgREST some 10 years ago, more or less "handing over" the project to Steve Chavez.
Ignoring begriffs, and counting only those with more than 10 commits to the PostgREST GitHub repository, Steve Chavez is responsible for 43% of all commits done towards the repository since the project was started in 2015.
Great job Steve! 😂
This means that the corner stone project SupaBase is 100% dependent upon is created almost exclusively by a single human being, Steve that is. Psst, you should ask for a raise Steve. Implying that a 1 billion dollar company, is 100% dependent upon a single human being, implying if Steve quits, and there's a bug in PostgREST, there won't exist a single human being on earth capable of (rapidly) fixing that bug.
AND it's implemented in a programming language that's the 38th most popular programming language on earth, with less software developers daily using it than FoxPro, where only 0.37% of the world's software developers could in theory replace Steve if he was to quit.
If you put 300 software developers into the same room, ONE of them could (in theory) replace Steve!
Conclusion
A single human being is responsible for having built most of the value proposition in a 1 billion dollar company single handedly alone. Pretty amazing if you ask me ...
More interestingly, is that from a risk assessment point of view, there are literally zero differences between the risk associated with SupaBase, and the risk associated with Magic Cloud.
The only real difference here being that Magic Cloud has 21 times as many downloads, and it therefor could be argued it has 21 times as many "real users", and that SupaBase (apparently) is worth 1,000 times as much according to venture capitalists, because, and I quote the venture capitalists themselves here ...
There is simply too much risk associated with investing in a single software developer project!
PS, impressive work Steve, however you should seriously ask for a raise, because you built SupaBase more or less alone!
Facts are, ignoring the technology, the only difference between SupaBase and me is that at least I'm honest about creating a single software developer project, and the fact that Steve had better marketing ...
Top comments (5)
Wow, love how you've unearthed this information. It's totally captivating, especially on the language side.
I'm curious, what drove such massive adoption of Postgress... and more so what makes devs decide differently from VCs with risk perspective? In other words, if a VC is not willing to put their money into a project based on its associated risk, what made it so that developers were?
(Real) devs don't care. As long as they've got access to the source code, they don't care how popular it is, only if it solves the problem. Mediocre devs will care about popularity, since they're not looking to actually solve problems, they're just looking out for themselves, and wants to have whatever is most popular and more likely to be giving them job interview opportunities on their CVs.
VC dudes of course, are pure psychopaths, and would invest in sjit if they could gets them a 10x ROI ...
... which is why everything that's VC funded, if it's not sjit already, inevitably turns into sjit over time, because the VC firm will dictate how the CEO runs the project, resulting in the development process turning into a "popularity contest" ...
Indeed, the VC is very able to destroy future value for the sake of extracting more present value. His motives are often the antithesis of those who do make things.
It was not all that uncommon even last decade. When I developed the original fire engine red wickr client, I personally did over 50% of the code of that, and it had a few million public users.
I'm not remotely surprised. Even though there exists technically 27 million software developers in the world today, I suspect less than 100 of us are responsible for more than 50% of the working production code ...