Build captain fractal using Colab

Thank you for tuning in to this audio only podcast presentation. This is week 130 of The Lindahl Letter publication. A new edition arrives every Friday. This week the topic under consideration for The Lindahl Letter is, “Build captain fractal using Colab.”

We are getting closer and closer to the world envisioned in Star Trek and countless other science fiction universes where you could just ask the computer to complete some sort of task or build for you. We are starting to see what would be like an AutoGPT or a plugin style version of ChatGPT with agents that work to complete some sort of request. It’s not like they have a full built in chron system with scheduling and routines. Really actionable virtual assistants are going to show up here very quickly. I’m pretty sure Google is not very far off from being able to allow their voice assistant to activate more things within routines and schedules with a real degree of agency. I’m actually fairly surprised they have not just launched that feature already. That ecosystem has the connectivity to be able to turn off and on lights or other home connected devices and connect to so many other things. What I am saying here is that you very well could have an agent do a variety of things very soon.

You can imagine that it would be fun to just ask the ChatGPT agent to write and do something with fractals. To that end I spent some time asking by prompt both Bard and ChatGPT to produce code for a Colab notebook. Strangely enough Bard never produced any code that was executable within a Colab notebook. When advised I was wanting code that would execute in a Colab environment the ChatGPT model did spit out code that would execute properly [1]. I’m adding notebooks that work to my GitHub repository on an ongoing basis. 

Beyond messing around with the generative models it seemed like a fun idea to look around and see what people were doing with fractals in Google Colab notebooks. The notebooks I found were produced a while ago, but that made them no less fun to play with. 

Fractal Art With Python by Dr. Mike Lam of James Madison University:
https://colab.research.google.com/drive/1tSBkON1Uj0NCfYEgELdXxC-PS9eOaUQG?usp=sharing#scrollTo=p02VqSeizpUN

Fractal Generation with L-Systems by Paul Butler
https://colab.research.google.com/github/paulgb/notebooks/blob/master/source/l-systems/Fractal%20Generation%20with%20L-Systems.ipynb  

I ended up wanting to just look at Python code for fractal building and manipulation. That is where I’m working with things right now in the process [2]. My base test for these new generative models relates to how well they can actively generate and manipulate fractals. I’m really starting to dig into how to use these generative features to build things in terms of code. Right now it is hard to get them to produce working code for a specific ask which makes me wonder if building out a full program will really be a reasonable thing to expect any time soon. Right now you really have to have a solid idea of what you want and be able to evaluate the output.

Footnotes:

[1] https://github.com/nelslindahlx 

[2] https://towardsdatascience.com/creating-fractals-with-python-d2b663786da6

What’s next for The Lindahl Letter? 

  • Week 131: Bulk imagine improvement scripting
  • Week 132: Synthetic data notebooks
  • Week 133: Automated survey methods
  • Week 134: Make a link based news report automatically
  • Week 135: Saving some notebooks every day

If you enjoyed this content, then please take a moment and share it with a friend. If you are new to The Lindahl Letter, then please consider subscribing. New editions arrive every Friday. Thank you and enjoy the week ahead.

Leave a Reply