Skip to main content Skip to footer

Fintrano - the game!

At the weekend, Iain started playing with AI coding tools. A few days later, this retro arcade game was born.

We're entering a new world of AI coding tools, and this week I decided to see how far I could get using Google Antigravity. It didn't take long before I developed a game I was happy to share.

I don't believe that you can judge the quality of code by the speed with which it was developed. In my experience there's generally a negative correlation! In the case of Fintrano, we spent months talking to potential customers before we wrote a line of code, and as a result we developed a powerful system which is very easy to use. 

But the world moves on. Just like a car body repair shop that needs to stay up-to-date as electric cars become common, we have to be ahead of the game and adapt to the new tools that are coming our way. The first step is to test those tools and make sure that we understand their capabilities and limitations.

Probably the best-known of these - at the time of writing certainly - is Anthropic's Claude Code. I hadn't had a chance to give it a look before I read a blog by a coder who said that he was getting great results with Google Antigravity, so I decided to give it a try by building an arcade game that could run on mobiles or laptops.

I was lucky that Out Run (the 1980's arcade game on which this is based) has obviously been built and well-documented in the past because all I did was point Antigravity at the Out Run Wikipedia page and describe how I wanted it to be configured (using the Fintrano Blue brand colour, for example).

What came out basically worked first time. There were a lot of tweaks to make, to the speed, how obstacles rendered, collision detection and do on, but the basic game engine that renders a 3D-like road, a car you control, and obstacles you avoid just worked like magic.

What came out basically worked first time.

I added things like different graphics if the car was turning, a leaderboard of scores, and even sound effects by prompting them into existence. I have a feeling that I was able to get better results more quickly because I understand a bit about graphics and sound files, but although I hand-finished some of the graphics (not too much, because wanted to keep this a predominantly AI project) I literally never edited a sound file.

(Any genuine game developers out there would say, "Yes, it shows!")

I even added the ability to create branded versions for customers and friends - for example this Car Doctor version that includes their branding in the billboards and share screen. That went from an idea to finished work in under an hour.

I can't state enough that this is not production-ready code. If I was a game designer/developer I'd have a long list of ways that this could be improved. And I think part of the reason I could deploy this quickly and successfully is that I'm already very familiar with that side of things, due to my background in web software.

But it is a sign of things to come, and it gave me a lot to think about.

About the author

Iain

Iain is a software developer with over 25 years in the industry. He runs Prominent Media, a software company based in Milton Keynes. Fintrano is his brainchild.