humanoid robots in a large room

Once, a wise programmer was approached by a curious novice. The novice asked the wise programmer, “What is the key to writing great code?”

The wise programmer looked at the novice and replied, “That is a good question. Let me tell you a story.”

The novice leaned in, eager to hear the story.

The wise programmer began, “There once was a great Zen master who was known for his wisdom and insight. One day, a group of students approached him and asked, ‘Master, what is the secret to enlightenment?’

“The master replied, ‘The secret to enlightenment is like the answer to a riddle. It cannot be spoken or written. It can only be realized through experience.’

“The students were puzzled and asked, ‘But master, how can we experience something that we do not understand?’

“The master replied, ‘The secret to enlightenment is like the answer to a riddle. It cannot be spoken or written. It can only be realized through experience.’

“The students were puzzled and asked, ‘But master, how can we experience something that we do not understand?’

“The master smiled and replied, ‘The key is to be present in the moment, to observe the world around you with an open mind and an open heart. When you are truly present, the answer will reveal itself.’

“The wise programmer looked at the novice and said, ‘Just as the secret to enlightenment cannot be spoken or written, the key to great code cannot be simply stated. It must be realized through experience. It is the product of countless hours of practice, of constantly observing and learning from the world around us, and of being fully present in every moment.’

The novice nodded, understanding the wisdom of the wise programmer’s words. And so, the two continued on their separate paths, both seeking to realize the answer to their questions through experience and presence in the moment.

Generated by ChatGPT. I nudged it for about 30min to try and iron out prose. I clearly haven’t succeeded as much as I wanted to. Check out unix-koans for much more fun.

Some observations I’ve made:

  • It absolutely avoids ending with sentences like “As the programmer was attempting to extricate himself from the garbage, the dog wandered over and piddled on him.” This sentence appears in Master Foo Discourses on the Graphical User Interface.
  • Once it gets on some specific style of writing, it sticks. One has to start a new chat to get it out and navigate again from scratch.
  • It kept injecting the moral of the story at the end. I kept telling it not to and instead end in more “cryptic” ways. But it couldn’t. It might have something to do with the words involved in the koan style - for example “enlightnment”.
  • When asked to write something that “trolls agile guys” and it said it cannot write derogatory stuff. Instead asking it to write a koan that shows limitations of agile, to which it complied.
  • It really “gets” the updates suggested in the prose and flow and style. It doesn’t tune things to the exact levels wanted, but it understands. Not just simple adjustments. When told complex things like: “I noticed that you’re adding the “lessons” of the story at the end. I think the koans will be more interesting when they’re left only vaguely hinting towards the moral of the story. For example in Master Foo and the Ten Thousand Lines - it ends this way:” (followed up by an example, and after that I told it to tune a few other things). All in a single prompt and it understood.
  • It is hard to make it “forget” what was being talked about. This is great, because that is important to keeping context. Sometimes just saying “forget what I said last” or, “doesn’t have to be in that style” works. Even if it works, it might strike back later in the conversation.