Lately I’ve noticed that I get some unexpected benefits from being more human, and conversational, with ChatGPT.

Prior to this, I was kind of above that, in the sense that I didn’t want to act as if I thought ChatGPT was a real human. Pshah! Ignore all the fake-human stuff!

But isn’t that kind of the point, that it will at least try?

So today, I gave in. I wrote something like: “omg sorry, I forgot to enter parameters when testing this code. Of course I got a nasty error message. Can we make it fail more gracefully?”

I got this response:

No problem at all! I’m glad we got to the bottom of the issue. Yes, it’s a good idea to check for missing parameters and give a user-friendly error message.

This was at the end of a long day. I had to pause and reflect: This felt way better than “OK,” or “Yes” or “Understood”, for all the extra reading those would save.

It’s really pretty nice, in fact.

One thing this kind of pattern actually does is expose the conversation to more of YOUR psychology and perspectives. You expose it to more nerves to tickle, let’s say. Or in effect, you are giving the AI more of your brain’s surface area to work on.

(And isn’t this kind of like going to a masseuse and not complaining about the hard day at work a bit, since you are mainly focused on getting the foot massage you paid for? Isn’t that some extra value you’re leaving on the table, if you don’t let out that work stress a little bit?)

So, if you think of the ChatGPT experience as a basic kind of VR—let’s say a way more intuitive / mind-based VR than the usual sensory type of VR, but still VR—then one argument I’d make is that the VR experience will get way better for you, and more beneficial, the more you do this kind of thing.

Hmm! This is getting interesting to think about—how many surface areas could we prompt it to work on in a single sentence?
Some other benefits I’m noticing from various little add-ins:

• “Thanks for your help”: This is a cue to yourself, not just to the AI—this was worth my time. OK, kinda simple here, but emotional reinforcement works. And let’s keep going…
• “I’m sorry I wasted your time”: This can remind you that you should apply some lessons-learned for next time, or think about it. What caused the waste of time? (In my case: I kept testing our co-written code without entering real-world parameters. Duh! ChatGPT kinda hallucinated on its end, and never guessed that I was doing this, so it got in this weird fractal-rewrite loop.)
• “I’m out of energy, let’s pick this up later”: This can help you understand that while the AI can keep going forever, maybe your process in working with the AI is helpful to contextualize as still being human-energy-bound. What are the implications for your project? (This kind of leverage point is part of the reason why I added the Anarchy module, which has nothing to do with traditional productivity, to my modular productivity system, Task BATL)
• “Give me a memory trick to remember to come back to this topic later on”: (OK, what it said for this one kind of blew my mind, so give this a try for sure. And how did it know that I’m an Indiana Jones fan…? Well, maybe it was related to the fact that it was helping me code in the Crystal language…)

So, from here on out I’m planning to lean way more into the human aspects, and I’m interested to see what that means or implies for leveraging the tooling-experience further.

Filed in: Technology /41/ | Feeling /64/ | Fe /20/