The Best Way To Learn Generative AI is...

... actually using it, not just reading about it.

Hey y’all!

What a week! There have been a lot of advancements and industry news this week but you probably get those sound bytes from other sources (and there are plenty) so you won’t usually find those in my humble newsletter. Besides, the focus here is practical use and applications than anything else so I’ll dispense with burying the lede!

You see, the best way to learn about generative AI is to actually use it, try it, and experiment with it in your own life. You’ll hear me harp on this point here and there but it’s like learning anything new — just go and do it.

So much of what artificial intelligence is — in its present state — is test-driving the variety of applications and tools and services in an effort to give you the results and outcomes you want. Even this newsletter is a personal exercise of doing more in the space so I can feel better equipped to leverage this new technological form in my personal and professional life.

I was encouraged by this tweet just this morning that reminded me that we all start at “zero” and that the only thing that’s stopping us, more often than not, is ourselves:

My mother has no idea I have a newsletter. ¯\_(ツ)_/¯

So wherever you’re at with your experimentations and use of genAI you can believe that you’re not alone as everyone is starting from zero but those who will reap the most rewards and benefits of it are the ones who actually go out and simply start using it. Find a simple application first and then expand!

For instance, I’m constantly on the hunt for a better avatar / profile pic for this newsletter as well as my X account and although many tools offer similar workflows, not all of them produce the results I want. So, every time I see a tool that offers some of this “text-to-image” production I give it a go and see if it works. Recently folks have been talking about StockImgAI so I decided to boot up an account and see if the results were to my liking.

Not unhappy with this!

Here’s the prompt I used to generate the above:

A korean asian 30-year old woman with a purple streak in her hair, portrait style, hyperrealistic, in a cyberpunk city, slightly smiling with the sun behind her wearing a black top with a sun streaks, slightly smiling looking at camera.

I’m not unhappy with the result! I could now upscale this via another tool if I wanted or iterate on this one (although that would cost me some money for additional credits).

🚨 — One quick “hack” would be to just use a different email address so you can have “unlimited” iterations but it would be a bit onerous over time. But, it would work.

Here’s the full output via StockImgAI:

I forgot to add a “small gold necklace and earrings”.

Since I had forgotten a few, small elements that I wanted to include — gold necklace and gold earrings — I also tried the exact same prompt via Stable Diffusion XL (for free!) on my macOS box and got some nice results too. It didn’t take too long to produce although the necklace is a bit large.

I can fix that, of course.

I love rendering this stuff for free.

Here are the two original outputs:

Actually, this is pretty close.

I like the purple and wind!

The point is this: If you keep experimenting and trying new tools you’ll eventually develop personalized workflows and expertise that you once didn’t have. And over time you’ll discover that you might even be an “expert” amongst your peers (not that that’s the goal, but would be fun to be a source of useful information for them!).

It’s like anything else that we learn from scratch: We go out onto the “field” and we start throwing the ball around and, after time, we learn how to throw it proficiently and catch it without dropping. We learn the rules and then we can play in larger, cooperative games. But, you have to step out there and go do it.

And I hope this newsletter can help. Thanks for subscribing and please share with others so that you can encourage them to get in this exciting game too! Cheers.

✌(-‿-)✌

— Summer