Used the new Google Gemini 2.5 Pro to create a computer graphics simulation to generate 3D plants with a technique called L-systems! Claude and Gemini were the only ones who could do it. This was 2 weeks of homework in computer graphics class in college!

this can easily be done in blender so seems like trained on some source code of blender as well

Can Machine Learning beat the market? Check out this post on my free Substack where I share code and commentary for an XGBoost model and a Random Forest model that both deliver powerful performances.

You can imagine the pain of now trying to invent programming homework or small software projects for university students. Most of them can be solved with a single prompt with good programming LLM (Claude 3.5/3.7 Sonnet, o3-mini-high, DeepSeek V3-0324/R1, Grok).

this interface looks nice, but o3-mini-high was definitely capable of L-system rendering. maybe not a single session or as comprehensive of a preset grammars setup tho

Yes I should’ve clarified. It can do a simple L system but I struggled to get it to do it to this level of depth and took a few cycles back and forth. Partially it’s the lack of the large output window

Wow. It’s also interesting how hard but fun projects just stick with students. Reimplementing class projects is what I do too when learning a new language

That sounds insane! L-systems are such a cool way to model plants. Gemini 2.5 Pro sounds like a game-changer for this kinda stuff. Wish these tools were around back when I was grinding through CS homework! 🚀👏

But you could have generate Ghibli memes dude! LOL. Thanks for something more productive.

looks great!

Wow

Fascinating! L-systems always remind me of how blockchain protocols grow organically too - both creating complex structures from simple rules.