Many people oftentimes associate ChatGPT and the gang with plagiarism. Need an essay churned out in two hours? Want to write a difficult email to your boss, but not actually have to write it? Who’s your chatty!
As unfortunate as these use cases are, a world of fantastic (and ethical) possibilities using Generative AI’s abound. Many of these come from simply taking advantage of this technology’s ability to recognize images we snap from our phone.
Below are four creative and unique ways to do this. And they might just spark a few more creative ways that you can come up with!
1. Curious about the World Around You?
The natural world is full of curiosities. But knowing exactly what you are looking at is often reserved for botanists, ecologists, and the like. When my 8-year old son, for instance, asks me what kind of spider just wiggled out of a hole in the fence post, the best I can often muster is “that’s a creepy crawly.”
Not that I advocate taking pictures of spiders at close range (or really any range) but by snapping pics from your phone, you can upload them to ChatGPT 4 and ask it all kinds of things. What kind of tree is that weird barky thing in my backyard? Or what is that weird green stuff growing on the side of my fence?
And unlike apps, say a tree identifier, that simply identify tree genus and give you a quick description, you can dig with ChatGPT 4 around any detail about the tree you’d like. (If I cut it down, are creepy crawlies likely to sprout from its trunk?)
Our physical world is also full of wonders that aren’t natural. For instance, I was driving with my family when we spotted a Tesla that was somewhere on the color spectrum between Pepto Bismol pink and loud purple. We had a spirited debate about exactly what shade it was (I was in the fuchsia camp).
So, I uploaded a couple of pics and ChatGPT 4 said, based on the lighting, it was most likely magenta. That certainly didn’t put an end to our debate — opinions are opinions, after all — but it did inject another voice, one more rooted to an objective color standard (at least that’s my position in this ongoing family debate).
The idea here is to turn your curiosity into learning — with a pic and a prompt.
2. How’s My Musical Form?
Many of us play (or at least attempt to) a musical instrument of some kind. We might have a teacher, but given the cost and the general hectic day-to-day of life, many use online videos to improve. The thing is none of these videos are able to give you that one all-important thing: feedback.
While Large Language Models (LLMs) like Claude 3 and ChatGPT 4 are currently unable to analyze audio samples, one useful thing they can do is analyze your form as you play an instrument. For those wanting to improve and avoid injuries down the road, knowing the proper posture and technique is, ahem, instrumental.
Speaking from personal experience as a piano player, ChatGPT 4 was able to point out in a screenshot that my elbows were a little high and I should not lean too far forward. I’ve heard similar things from teachers I’ve had over the years (though that advice did not come on the cheap.)
So, whip out that bassoon, bagpipes or piccolo and let Claude 3 or ChatGPT 4 help you level up your technique.
3. How’s My Fitness Form?
As many musical people as there might be out there, they likely pale in comparison to the number of the people who exercise (even if it is only in those first two weeks after the New Years). So, this is a use case that I’m sure will resonate with many.
For me, I’ve never been stellar in the yoga studio. (I’d tell you a joke about how I can’t touch my toes — but that’d be a stretch!) I recently had ChatGPT give me feedback on my downward dog. I was actually surprised that it wasn’t as bad as I thought. But it did give me some good pointers.
(To save myself from public embarrassment, I’ve opted not to include a screenshot of me in downward dog).
Essentially, you can take any stretch or exercise (are you getting the most out of those dumbbell curls?) and get feedback from an LLM. And for many of us, this kind of feedback is invaluable–and a fraction of the cost we might pay a personal trainer.
4. How Can I Be More Like Monet?–-just ask Claude
A fantastic way to use Claude 3 and ChatGPT 4 is for feedback on artworks. This was one area that came to mind right away because, unlike music, which LLMs can’t quite yet parse, their image recognition abilities are highly sophisticated.
The only thing is I’m not much of a visual artist, so I wasn’t quite sure if I could speak to this. That’s when it dawned on me: my wife is an avid watercolorist, a hobby she took up during the pandemic. I was able to take pictures of artwork (with her permission, of course!) and get feedback.
Even better, I was able to take pictures of many of her artworks and asked Claude 3 if it could see any progression in the execution of her pieces. Its feedback matched her own timeline, her own evolution. My wife loved this feedback and found validation in it for all the hours she had spent painting (not to mention the online courses she took that had helped her become increasingly — per the LLM — more accomplished.)
Putting it Altogether
So whether you want to overturn a rock in your backyard to wonder just what kinds of critters live underneath it or whether you want to render said rock and critters in watercolor, Generative AI can help you answer what, how, or how well. And for curiosity bugs and budding artists alike, that is certainly a game changer.
Comments