Originally, I wanted to entitle this post 危"机" — opportunity in crisis.
Our web host is shutting down its operations in Singapore, and we have to migrate our website. Suggestions by the other two frontier labs' AI models was to either choose another existing web host or move the site to WordPress.
However, our new Coworker aka Claude had other ideas. Long story short, it not only helped to revamp the website, but also taught us how to migrate the whole site to another service provider. It also went on to help us build a website builder so that we're never held hostage by a service provider again.
From building a new website, to making a website builder, all as a result of being forced to migrate — what a fine depiction of 危"机"!
The experience of working this through with Claude, especially with Anthropic's Fable 5, is worth recounting and warrants a separate post altogether. I mean, I can still relive the excitement of the moment when Fable 5 says "I've been waiting for that question"... This is surely one fable that needs to be shared.
But I really want to share the key insight that I have drawn from this experience first.
And that is: we have to know what good looks like. Only then can we enable it towards the output we desire.
And how do we know? And to add further nuance to that question: how do we know we know, and how do we know if we know what we don't know? How we answer these questions will affect our interpretation of how good looks like.
Why is this important? Because a low baseline of what good looks like will deliver a poorer output, compared to a higher standard of what good looks like.
Can the machine know what good looks like? Yes and no, depending on what it uses as a gauge. A simple example would be: if all it can reference is AI slop to know what good looks like, your output is likely to be AI slop as well. So relying on the machine to tell you what good looks like may not be the ideal solution.
So to answer my own questions, I believe education and experience, or what eventually develops as domain knowledge, would give one a good idea of what good should look like.
If not, proper research into the subject matter can lead one there. And from there, we should strive to understand how to achieve what good looks like.
But why bother, you may ask? Because we are in this nascent era of our very own Renaissance age. Where in the past, a PC to most of us just meant either an overglorified typewriter or a gaming device, computing is now becoming actually personal because it can effectively help us build and design solutions for ourselves. What we are limited by is no longer our capability but rather creativity and imagination. I daresay this remains the last bastion of our humanistic nature, so we should keep developing our creativity and expanding our imagination.
To echo the words of the great George Bernard Shaw:
"Some men see things as they are and ask why. Others dream things that never were and ask why not." — George Bernard Shaw
Not only do we ask why not, we can actually aspire to build these dreams and turn them to reality.
But to do that well, we need to truly know what good looks like.