From academia to startups
Why I love coding for no-code
My personal journey to working on no-code. How it can be a deeply technical area. Why I love this space, and why you might too.
I never thought I’d end up working on no-code.
No-code software is meant to be simple, easy, and intuitive. It’s tools like Excel for mom-and-pop users. There shouldn’t be anything too deep and challenging for geeks like me. Oh boy, was I wrong!
Let’s see how I got here…
As a teenager with wide-ranging interests and far too meagre attention span, it was painfully obvious my time on this earth would be insufficient to learn everything worthwhile, so I would have to prioritise certain fields of knowledge over others. With my fascination for the abstract, I would rather read about linguistics than learn a language – or better still, study mathematics (the language of nature).
No-code sometimes feels like this.
No-code tools are usually quite flexible in that they are platforms that give the user lots of freedom in what to do – although not necessarily how to do it. This guardrail is the tradeoff that makes them easy to use.
Instead of tackling a specific instance of a problem, you need to abstract away and solve for the most general case. A bit like a mathematician, you go from the particular to the universal – you build up your intuition by playing with examples until you notice an underlying pattern.
These days, I put myself in the users’ shoes, anticipate the interactions they are going to have with my software, what might delight them and what is prone to frustrate them, and I strive to write code that handles this variety of tasks with elegance. Well, that’s what I keep telling myself anyway!
As a smug college student, I was obsessed with logic, epistemology, and decision theory.
After all – as my argument supposedly went – learning how to reason, how to learn, and how to make optimal decisions is so foundational it should trump everything else. Maybe there’s some merit to that claim, but at the end of the day, who cares if your epistemic modal logic can be modelled as graph with a transitive accessibility relation?
No-code feels the opposite of this, and in a good way.
People who make decisions in the real-world must deal with all its messiness.
It won’t be some theorem that saves them. But a piece of software that might make them a little bit smarter, a little bit more knowledgeable, more capable of drawing conclusions from their data… well, that might just help people make better decisions, and in turn make the world a better place.
After university, I gravitated towards finance. Too many Nassim Nicholas Taleb books or something.
Jokes aside, working as a developer on a quantitative trading desk was fun. But it made me aware that our software – so crucial to profits that it had to be delivered under the tightest time constraints – inevitably ended up being frail, wasteful, and disposable. If Wes McKinney hadn’t donated his time to writing the popular library pandas for data science, we all would’ve been 10x less productive at our jobs.
There’s enormous leverage in building reusable solutions.
No-code tools are exactly that. But they aren’t just for yourself, or for your team, or for your company. Or only for technical people who can code. They are accessible to everyone.
The impact is so much greater. In solving things once, getting them right with full generality, and letting others reuse your solution, you are enabling people to focus on the problems they truly care about – whether it’s saving the world from climate change or merely increasing profits for a trading desk.
Maybe not all no-code development interests me.
I have never touched a front-end UI component.However, some areas can be quite challenging – giving you the kind of intellectual vibes you get from working on libraries, frameworks, compilers, developer tooling, etc. – while having the potential to reach the broadest audience possible.
There’s something uniquely satisfying in that.
If you’d like a taste of the research problems we are working on, subscribe for our upcoming blog series Seamless & Collaborative, where we show how ideas from Git, collaborative editing, and differential dataflow are key to the future of data.