The swamp of brilliant ideas

Share
The swamp of brilliant ideas
Photo by Håkon Grimstad / Unsplash

When I left Microsoft, I went back to Turkey to turn my website into a proper business. I succeeded. It's still the most popular Turkish social platform in the world. But, obviously, I had more ambitions. I'd been getting ideas to do everything and anything.

I started keeping a list for these ideas. I thought I would later visit them and develop them into full products, but my first priority was "don't forget this!". So, I kept writing down ideas on a spreadsheet.

I've had the ideas on that list so long that some actually got outdated. Either other people would have finally done it, or the technology it relied on would become obsolete. I started scratching over those that did, but the list kept growing.

Years passed. Whenever I visited that list, it made me depressed. There it was, something I jotted down three years ago, but never acted on. Something else I wrote down with excitement months ago, but don't feel as excited today. The list started to turn into "the list of things I failed to achieve", like the opposite of a success story.

What was wrong? I didn't know, but I thought that if I developed the idea to a certain point, at least more than a single-sentence explanation, it could push me further to realize it.

That's when I started writing design documents for ideas. If a new and exciting idea hits, I would immediately open up Google Docs, and start writing a design document for it. That was helpful indeed. At least, I was able to find blind spots in the ideas, or find out why they wouldn't work while writing about them. Some of them would still feel great so I'd leave their design document around. I would visit them one day for sure.

This year, I migrated from Google to Obsidian for my documents. That gave me an opportunity to go over my old archive. Design documents, some are pages long, some are barely a page. The same feeling of not having been able to achieve so many things hit me again.

It's not like I'm dormant obviously. I create new stuff all the time. I write apps, write blog posts like this one, I draw, I even wrote a book which was also a success. Yet, I can't get over the feeling that those dusty design documents gave me.

Why didn't I visit them? Maybe, as Louis CK once put it, writing down the idea, and imagining my idea becoming a success made me feel satisfied enough that I didn't have the need to implement them.

Or, maybe something else...

Regardless of why lists or design documents don't work, your ideas can turn into a swamp that pull you in. You can dream about their success, their implementation, the challenges all along the way, but you can never go ahead and implement them.

Me looking at my lists

The escape from that swamp is simply not clinging on the ideas too much. Ideas come all the time, "show me the code", as Linus Torvalds once said. One finished novel is always better than a thousand great novel ideas.

So, writing a longer description of your idea isn't enough. Creating a full-blown design document isn't enough. I think it's only enough if you develop your idea into a prototype that can be experienced by others. Otherwise, you're wasting time even thinking about the idea.

When you start implementing your idea, you start facing the reality of costs. Your use cases are challenged by the mere realization that they may not be worth to spend that much money on. And finally, when you have a working prototype, you also face your idea the first time in person, and it may not look as good or be as useful as you thought.

That's probably why I loved Borland Delphi as a development environment back in the day. Because it let me design and develop simulatenously. Both acts would be intertwined, and that way, I could benefit from the behavior and the actual looks of my app to reshape my idea, and my renewed ideas could turn into results almost instantly, almost like a dialogue between your ideas and your code. It compiled and ran extremely fast too, we're talking about sub-second build times on CPUs and RAMs of 30 years ago; no joke.

Unbelievably, we lack this kind of fast iteration tools today. Microsoft even dropped the support for design time viewers in its latest generation of GUI tools because they thought runtime previews were enough. But, it made the design process slower as you now need to be able to recompile your app to iterate over your design. It's not always possible, so, you're stuck when the build's broken. You can't test your ideas anymore. Not the build, the process is broken.

What we have is layers and layers of abstractions with varying levels of overhead. Every layer requires vast knowledge about its domain in order for you to understand its characteristics.

So, AI is here to fix all that. That's what we hear. Instead of reducing complexity, we have created a monster that can tackle that complexity, and add more layers on top of it every iteration. It's like Titanic pushing its own iceberg in front to avoid crashing into other icebergs. Yes, AI makes realizing an idea simpler. But, I expect that the goalposts to move too. With AI making creation easier, even making something won't be enough to have it truly realized. Our expectations of a finished product will mature too: we will want better UX, leaner and faster implementations, and more importantly solid reliability. Your single-shot prompt app just won't cut it anymore. It can only be part of the garbage pile if you don't work on it. Your ugly-purple homepage isn't going to be different than an unrealized idea but with you having had spent perhaps thousands of dollars on top.

Since we have even more tools at our disposal now to make stuff, our minds should be more focused on realizing than imagining. More importantly, we must think extra hard on what it means to actually realize something other than blindly accepting what AI gives us. Otherwise, we'll just have moved from the swamp of brilliant ideas to the swamp of brilliant AI-generated prototypes.