Month to month, week to week, day to day, almost hourly, my biggest challenge with AI has very little to do with the technology itself.
It’s my own thinking.
The models keep improving. Their capabilities expand. Things they struggled with last month become routine.
But my mental model of what they can do doesn’t automatically upgrade along with them.
It looks like this: You try something with GPT. It handles X beautifully but falls flat on Y.
So you file that away: ๐๐๐ ๐ช๐ด๐ฏ’๐ต ๐ท๐ฆ๐ณ๐บ ๐จ๐ฐ๐ฐ๐ฅ ๐ข๐ต ๐ .
A week later, an update makes it excellent at Y.
But nobody tells you. You don’t think to try again. As far as you’re concerned, that limitation still exists.
๐ก๐ผ๐ ๐บ๐๐น๐๐ถ๐ฝ๐น๐ ๐๐ต๐ฎ๐ ๐ฏ๐ ๐ต๐๐ป๐ฑ๐ฟ๐ฒ๐ฑ๐ ๐ผ๐ณ ๐ฎ๐๐๐๐บ๐ฝ๐๐ถ๐ผ๐ป๐ ๐๐ผ๐’๐๐ฒ ๐ฎ๐ฐ๐ฐ๐๐บ๐๐น๐ฎ๐๐ฒ๐ฑ.
It gets worse for people who use AI occasionally. Some are still making decisions based on limitations they encountered two years ago.
Here’s the uncomfortable part: How much potential are you leaving untapped because you’ve already decided something can’t be done?
How would you even know?
The bottleneck is increasingly at the interface between what AI can accomplish and what the human thinks to ask it to accomplish.
So what’s the solution?
๐๐ผ๐ป๐๐ถ๐ป๐๐ผ๐๐๐น๐ ๐๐ฝ๐ด๐ฟ๐ฎ๐ฑ๐ฒ ๐๐ผ๐๐ฟ ๐๐ต๐ถ๐ป๐ธ๐ถ๐ป๐ด.
Revisit old assumptions. Challenge conclusions you reached six months ago. Retry things that previously failed. Experiment with approaches you once dismissed as impractical.
Make a habit of taking something you’ve decided AI can’t do and seeing whether that conclusion still holds.
And occasionally, ask AI to challenge your assumptions about AI itself.
In the old days, technological revolutions were rare events in an individual’s life. Seeing a Model T drive down the street for the first time might have been a once-in-a-generation shock to your understanding of what was possible.
Today, AI can deliver another such shock before you’ve fully absorbed the last one.
๐ฌ๐ผ๐๐ฟ ๐๐ต๐ถ๐ป๐ธ๐ถ๐ป๐ด ๐ฐ๐ฎ๐ป ๐ฏ๐ฒ๐ฐ๐ผ๐บ๐ฒ ๐ผ๐ฏ๐๐ผ๐น๐ฒ๐๐ฒ ๐๐ต๐ถ๐น๐ฒ ๐๐ผ๐’๐ฟ๐ฒ ๐๐๐ถ๐น๐น ๐ฐ๐ผ๐ป๐ด๐ฟ๐ฎ๐๐๐น๐ฎ๐๐ถ๐ป๐ด ๐๐ผ๐๐ฟ๐๐ฒ๐น๐ณ ๐ผ๐ป ๐ต๐ผ๐ ๐บ๐๐ฐ๐ต ๐๐ผ๐’๐๐ฒ ๐น๐ฒ๐ฎ๐ฟ๐ป๐ฒ๐ฑ.
RW
.
.
.