Because I Want It to Exist

It's been a while since I have listened to a talk and found a real connection with the speaker and what they were sharing. This streak was broken a couple of days ago when I listened to a talk by Ryo Lu at Cursor's Compile 26 conference.

Published on: Sun Jul 05 • Written by Schalk Neethling

AI is a divisive topic. While a lot of what I will talk about here is related to AI and engineering, I want you to try and look past it for the moment, because the topic I want to touch on is broader than AI.

“I wish this existed.”

That is how Ryo started their talk, and I was hooked. On any given day, I have several ideas running through my head. These could be ideas that have been simmering for some time, a new idea that popped out of a conversation, or, seemingly, out of nowhere.

However, you may have heard the phrase:

“Ideas are cheap, execution is everything.”

The thing is, the lack of execution is often not by choice. It is because life happened. A lack of time, resources, incentive, self-doubt, or that nagging fallacy that often plays out in our heads: “Why bother? Nobody cares.”

The longer an idea remains just that, an idea, the more we start to impose judgment on it. Whether that judgment is accurate or not is not the point. The point is that you are analyzing and judging something that does not yet exist. I do wonder how many great ideas have gotten lost to time for exactly this reason.

For me, time is at the center of all of this.

The greater the time investment, the more those individual concerns compound. The more time is needed, the more resources are likely required. The more time invested, the higher the incentive must be, or at least, this is how it feels. The longer it takes, the more those nagging questions start to compound.

“Why bother?” “Can you even do this?” “Is this even possible?” “There is probably a reason this does not already exist.” “This is such a poor idea.”

And so on. As each of these questions festers without an answer, self-doubt gets fed a little more, until the idea becomes yet another side project that never gets finished. A domain you registered that never gets used. Over time, I have personally felt how this loop becomes the thief of ideas and creativity.

But as Ryo states, AI is collapsing the time between ideas and reality. AI is introducing a new loop. A loop where ideas become reality in hours, and sometimes even minutes, and then turn into clay you can reshape, play with, look at, ask questions about, share, change completely, or even discard.

Because of this collapse, we can get to something tangible and real more quickly, and start to see and experience the things that are wrong. There is a well-known quote that goes something like this:

“No plan survives its first contact with reality.”

I bet many of you reading this are nodding along, as you have encountered this many, many times. The design in Figma that looked just so, not quite working out when it meets the nuance and wild abandon of the web. Where data would come from, and the shape it would be in, not meeting expectation. How data would flow from backend to frontend, and whether we even have all the data we need, remaining an open question until that second, awkward check-in.

Then we actually try to use it, only to find that while all the data exists, and it looks beautiful, the experience is simply not what we want to ship to our users. In fact, we are not even sure that it solves the problem we set out to solve.

And here we stand again. What did we lose? Time. What did we ship? Likely, very little. And what did we accumulate? Frustration. And to round it all out, what did we move ever closer to? The chance that all of this work, all of this time spent, will simply be discarded, unused.

The new loop does not so much change the stations, but the time between them.

Because of the time compression, and what we can learn from each turn through the loop, we can now repeat this loop many more times, learn, build stronger intuition, hone our craft, and build better solutions.

Mediocrity and the Danger

This collapse is not all positive if we are not careful. We can lose ourselves in the machine. As Ryo says, “I do not want to live in a world where machines build the software, and humans just watch.”

This is something we need to actively guard against, because if we idly stand by and allow this to happen, we just might be surrounded by software that looks fine, works fine, but is lifeless and boring. Software that is broken in subtle ways, software that excludes groups of people, software that is just good enough that most people do not complain.

Ryo names this directly: “this does not lead us to a renaissance, but into a world of slop.”

I do not want to live in a world of slop. Now more than ever, I want all of us to demand more of ourselves and others.

On caring

Ryo draws a distinction here that I think is the missing piece of the whole picture: output versus material. Output is what the machine hands back to you, the generated code, the assembled screen, the draft of an idea. It is complete in the sense that it runs, but it has not yet been decided upon. Material is what output becomes once a person actually looks at it, questions it, and chooses to keep, reshape, or discard it. The same block of generated code can be output in the morning and material by the afternoon, and the only thing that changes in between is whether someone engaged with it.

This is the piece that guards against the mediocrity I described above. Nothing about generating output demands judgment; that is precisely what makes slop so easy to produce at speed. Material demands judgment by definition, because it does not exist until someone has applied it. This is the step where we take a step back and ask:

Skipping this step is the easiest way to end up with software that looks fine but was never truly considered. It is also, I think, the one part of the loop that AI cannot do for us. The machine can generate output all day. Only a person can decide whether it deserves to become material.

There is a fear running through much of the resistance to this shift: that if we are not making every small piece ourselves, we will lose the ability to truly care about the outcome. I do not believe this is true. Everyone finds their own place in the process of bringing an idea into existence, and how close a person stands to the underlying pieces does not decide how much their contribution matters, nor how much they care. Care, taste, and meaning are not properties of proximity. They are poured into the act of making itself, wherever in the process that act happens to fall for you. The world and our lives are still fueled by dreams and ideas, and only we get to choose whether AI empowers that or destroys it.

This new material allows more of us to collaborate across disciplines, allowing us to continue to break down the walls between product, design, planning, and engineering. It allows us to truly step out from under the waterfall. It allows us to finally deliver software and services that enable and include everyone.

In conclusion

As mentioned, a lot of this is about AI, how we work with it, and how I think about it. However, I do not believe it applies to AI only. It is about us. It is about how we choose to move forward from here.

Care more now than you have ever cared, because there is room to care. Push the boundaries of what is possible, because time to experiment is no longer the constraint it used to be. Be curious about how and why something works, because exploring the answer is now within reach of more people. Open the doors ever wider, and invite more people from more diverse backgrounds, perspectives, and levels of experience, because we can now all contribute. Let down your guard and take the time to find your place within all of this. Your place might even be to oppose all or some of it. That, too, is a valid place to stand.

Above all, dream. Make it real. Share it with the world. Shape it. Make it better. Make it meaningful.