Hi everybody - an observation that keeps playing out for me is that AI makes it ridiculously cheap to try a weird idea and see what happens. You’ve probably seen it yourself as well.
One such idea that I brought to life recently is an interactive Who’s Online feature, and this one features a twist. Instead of being a flat list or overlays on a map, I decided to build this involving a claymation-style spinning earth, clouds, pins representing live users, and...flying paper airplanes.
The video of this in action can be seen below:
The video doesn’t do this justice. Go see this spinning globe live for yourself and interact with it to see all of the little details. There is also a more traditional table view below it just in case tables are your jam! 🍯
You can also see this and my brief writeup on X / Twitter and LinkedIn. If you don’t mind, please share it with your network. The extra reach would help me out a lot. More importantly, my ego has grown accustomed to being fed on a regular schedule.
Wouldn’t have happened without AI
This interactive Who’s Online visualization idea is something I've had at the bottom of my backlog for years:
I had completely forgotten about this idea until a few days ago when I stumbled upon it by accident.
In the olden days (like 6 months ago), to bring this idea to life, I would have cracked open a book on WebGL and learned about meshes, shaders, and all sorts of techniques for creating a performant 3D scene that dynamically updated with data:
Realistically, though? I would never have even started. I simply don't have the massive amount of time to learn a whole new area of programming, especially when I have no idea if this idea is even worthwhile. Reading the description of what I wanted, it doesn't exactly scream Winning idea!, right? There are higher-priority P0 activities with a more predictable payoff (like writing a top-requested tutorial) that would be a better use of my time.
This is where AI played a crucial role. In this new world, the cost of bringing a seemingly complex idea to life is dramatically lower. I gave my agent my criteria for what success looked like, and I let it run overnight. I woke up and saw more or less a version of what I had imagined all those years ago. The agent got many things right and many things wrong, so I spent a few more hours tweaking the output. That is not a bad ROI. Then I pushed it live. That’s it.
What my path to getting this item shipped raises this point: we're no longer constrained by the cost of bringing an idea to life. At least, AI dramatically reduces the cost of producing a credible first experiment. It makes it much cheaper to find out whether an idea is worth pursuing. For example, if I had to discard this idea, no biggie. The time lost would have been only a few hours as opposed to the multiple months the traditional path would have taken.
AI Upends the Traditional Team Dynamics
If we take a step back and put this into the context of the software industry and the roles many of us have within it, it turns everything upside down. We are no longer bound by the traditional roles where:
PM writes a spec
UX helps the PM think through the feature in detail
Eng builds a prototype to validate
Eng then collaborates with the PM and TPM to create a final plan
Eng begins delivery of feature
TPM helps project manage (status updates, stakeholder management, etc.)
Domain-specific disciplines like Marketing / DevRel / Legal / Sales / etc. get involved at various stages of the product ideation and delivery
???
Profit
Wherever your center of expertise is, you plant your flag. From there, you use AI to either fill in tasks handled by other disciplines or provide evidence that helps bring them into your proposal. If nobody believes in your crazy idea, AI lets you build enough of it to make the case. You may still need buy-in to ship it, but you no longer need buy-in to prove it can work.
For this feature, because I greatly influence what happens on kirupa.com, I didn’t have to ask for permission to build it. If I had to push for a low-pri P3 feature in any other setting (like in a team at work), fuggedaboutit. Most teams operate under a mindset honed over decades, the scarcity mindset. That is a topic that requires a dedicated post of its own, so we’ll get to that in the future.
Till Next Time
What is a low-pri crazy item that you were able to validate as a good or bad idea because of AI? If you have something to share, feel free to reply back or post online in the forums or on x / twitter (tag @kirupa). I’d love to hear it.
Cheers,
Kirupa 😇