You want to know what I've learned about leveraging large-scale AI functions in my work? The key thing?
The best ideas still come from my own brain.
And that's not to say that I'm standing on high claiming I'm brilliant. Not at all. But as I've leveraged, specifically, ChatGPT for I don't know, maybe a year now, to brainstorm and analyze and sort of dismantle and reinterpret concepts from technical to simply analytical, there have been a lot of times when I've gotten a pretty solid result. Whether I'm trying to develop something that's more color science oriented, working on the retouching tools that I develop for myself and ultimately sell, or maybe just analyzing an itinerary for an event and breaking it down in ways that make sense, GPT has been reasonably helpful.
Think about it. What can a computer do that it is usually very, very good at that humans tend to be weaker at? Well, large-scale analysis of lots of complex data. That's an oversimplification, but basically, yeah. So if I have an idea to create a tool or function in Photoshop, and I think it could possibly work, but I need to check some specifics about certain calculations, or dig through Adobe documentation quickly and find references and things, GPT has been hugely useful there.
But the truth is, even as I sit here doing a vocal dictation into GPT so it can type this out for me, what I can tell you openly and honestly is that GPT's ideas, when it comes up with "fresh" ones, have not been of the caliber that I need to run my business and do the work that I like to do.
So in a year of leveraging GPT for various business things, mostly in the conceptualization space and the analysis space, like I said, I've realized that my own ideas have been the most fruitful. Either they're the most satisfying and fulfilling, or they're the most beneficial business-wise and have resulted in great products.
The human idea that looks forward is extremely high priority still, even in an AI world.
If you feed a great idea into an LLM such as GPT, a lot of times the adjacent ideas it can add and provide will augment what you're trying to do, or maybe clarify a specific process or step. And the more technical you leverage it for, the better. (I mean, many people are using it for programming and coding and things like that, and there's a benefit.)
That's not to say there isn't misuse, overuse, and a ridiculous amount of AI promo flyers flooding social media. That's true. But despite those promo flyers, and all the AI music out there that's wasting capacity on Spotify and other platforms, these dead, empty vessels of AI-created slop, somewhere there are still humans everywhere around the internet with ideas. New ideas, fresh ideas, and new interpretations of older ideas.
I try to think of GPT as the corny stereotypical android house assistant from countless sci-fi shows and movies from the last several decades. It'll help with the technical things and organizational things sometimes. But, I promise you I am living proof (pun intended) that your own ideas are always going to be superior to trying to let any type of large-scale AI tell you how to run your shit.
Do not discount yourself. GPT can tell you a detailed historical record of any given region, in any given country, in any given time frame in the history of civilization in the blink of an eye, never mind absurd mathematical calculations and more. Just let computers do what they're good at, but continue being the human that brings ideas from nothing into something.