About 15 years ago, I've developed an analog and very haptic multiplayer simulation game for students in a class setting, typically for innovation courses. "Investors" would give money to "startups" and get shares. Each investor and startup received a briefing with goals (all slightly different) and had to reach these goals. Students love it - they learn a lot about how these markets really work; it's competitive, fast, and fun and there's lots of adrenaline since we also play for real money.
However, the game has many limitations because of its drastic simplifications due to the analog nature.
So, I thought several times about digitizing it, since it would be actually more real and even more interesting if digital. However, building this simulation just two years ago would have been a major investment (weeks of coding work), also because of its near real-time character.
But things have changed... I asked Claude and ChatGPT for suggestions on how to build such a thing. ChatGPT's output was far too complex. Claude came up with a simple proposal that ran almost immediately.
It still took about 15 hours of work with Claude, but the simulation is ready to play for the next class - including TLD, dynamic rules, penalties and bonuses, etc. Of course, it helped that I had built distributed networking apps myself many years ago: Claude's original suggestion was not at all optimal in terms of network load and synchronization of the clients.
Still. It's spectacular what you can do with relatively little effort in a short time. Not sure whether this would be possible without any programming / software design knowledge. But for someone like me, who did software for a living some 30 years ago, it's a dream come true.
So, if you need a one-day German course about startup investing including a simulation part, please let me know. We're ready to go :-)