In last week’s post, I wrote about my lunch with AI maestro Scott Schang; and how AI had helped me choose a suitable restaurant for our meet up.
I teased about the conversation we’d had and promised to fill in some of the details this week. There’s way too much to go into a single post, so here’s an overview. With more to follow over the coming weeks …
In an earlier post, I noted that Scott was brilliant at getting Chat GPT’s custom GPT builder to do all sorts of amazing things. Things that on the face of it seem impossible.
I’d seen the results of Scott’s genius; but I hadn’t fully understood how he’d been able to achieve the results he had.
Over lunch, he very patiently and very generously started to explain to me his methods and techniques.
This was enough to set a hare running in my brain. It lead to an hours’ long back and forth with ChatGPT over the following two days. I started to dig deeper and really understand more fully what Scott had described.
And this is one of the things I just love about AI. It’s ability to meet you where you are and help you learn from there.
I’ve spent years designing and delivering training to adults across all manner of topics and at all sorts of levels of complexity. One of the hardest things to get right is the pace and the level of detail. Too slow and too easy and people get bored. Too fast and too complex and they get overwhelmed.
It’s rare to get a group of adults who are all at exactly the same level of learning need; so, pitching at the right level of detail and at the right pace is incredibly tricky. You’ll almost never satisfy the needs of every single person in the group.
But AI is truly personalised, one-to-one learning. It meets you where you are at and responds in a way that is generally pitched just right. It doesn’t get offended or irritated if its responses are a bit off. You can easily tell it to simplify or level up its explanations.
Better yet, it never gets tired. You can just keep going for as long as you like; break when you want to; pick up again minutes, hours or even days later. It’s still there waiting to help.
In short, my hours long lesson in understanding how Scott has pushed the envelope with a relatively simple GPT-maker helped me to thoroughly re-evaluate the capabilities (and usefulness) of said GPT-maker.
It also helped me to understand how Scott’s approach can be re-purposed and refined slightly for use within custom GPT set-ups that are more technically complex.
This felt like a real breakthrough moment.
I’m planning to write much, much more about all this in the coming weeks.
