The Democracy of Doing Things
Debating whether it's ok that we have tons of AI tools for doing so many things.
In this post:
We constantly debate whether to do something on our own or ask for help.
AI tools give us the power to design our own images and logos and not hire professionals.
Is it bad that we’ve democratized access to something once reserved for specialist designers? I don’t think so.
I approach many tasks with a simple but vexing question: should I do this, or should I ask someone else to do it? The possibilities are endless.
Perhaps I’m cleaning data as part of a broader project analysis. I’ll dump that on someone else. Or maybe I’m finalizing slides for an executive pitch. I won’t outsource that work because I want the slides to be authentic to my story-telling prowess.
Maybe the task is purely domestic. My lawn is overgrown and needs a mow. Do I do it? I was quite the lawn artist as a teen tasked with lawn upkeep. But today’s time might be better spent elsewhere. For now, I’ll outsource.
Suppose I need a logo for a project (i.e. The Teardown). But I’m not a designer. I don’t have vector illustration chops. I don’t have a degree and I don’t have the required software.
Should I design the logo or outsource the design to someone else?
Not that long ago, many people asked someone else to help. A designer charged a nominal fee (say $500) to create a logo. But, we probably didn’t imagine and finish the logo through one conversation. You had requirements for the logo, and of course, the initial draft might’ve looked very different from your vision. So, you traded emails and calls with the designer, eating away at the value of the $500 to the designer. Meanwhile, the question lurked as time passed: was outsourcing worth it?
That cost-benefit question swirls around your brain when making decisions about things you want to do. How much does it cost to ask someone else to do something vs. what can you do, and how much does it cost you. Time and complexity are inputs.
I’m weighing this type of question now. I want to design a logo and simple website but am balking about whether to build the site myself or give the work to someone else. External estimates seem quite reasonable to me.
But I’m wondering whether I can build these assets with some help. I’m speaking about AI, of course.
If I do and don’t pay someone to help, am I assaulting the art of design? Am I screwing over an entire category of professionals?
John Gruber and Ben Thompson discussed this broad problem during their paywalled March 28th Dithering podcast titled Ghibli and Democracy. The episode title referenced the very recent explosion of ChatGPT 4o-generated images styled after famed Japanese animation house Studio Ghibli’s work. See it here. And - that’s me:
The internet was ablaze with opinions about what seemed like plagiarism. Did OpenAI train ChatGPT 4o (4o) on Studio Ghibli images and animations? We will obviously not know the answer to that question. But how 4o generates images so closely resembling Studio Ghibli work is an ongoing debate.
There’s a question about plagiarism, sure. However, Dithering’s hosts touched on the existential question surrounding this example: is it bad/wrong/etc. that we (you, me) now have a tool to create stylistically-similar images?
I’m not sure. What do you think?
I’m happy about a few core outputs from the last 2(ish) years of AI product releases: I write code faster. I clean messy data faster. Questions about nearly any topics are answered in less time, with fewer hopes between my curiosities and their resolutions.
Also, in writing this newsletter, I use AI to create images that align to the themes of the post. I don’t ask designers, don’t spend much money (other than monthly AI subscriptions), and spend very little time crafting those images.
You might look at that last paragraph and say: it shows.
Your images suck! You’re allowed to think that. But I optimize around writing posts and putting some basic necessary visual breaks in between blocks of text comprising those posts.
What would it cost me to hire a designer for every post, or buy images from a library? More than I’ve spent in pure monetary terms. But also more in time-cost, and in logistic-cost. I don’t see my choices as bad, evil, or usurping an entire industry of people. Newsletter images are easy. A brand’s professional image and visual appeal is a different story.
So many products are out there with the goal of democratizing doing or access to something. Perhaps access to financial products. Maybe better-than-average grammar, spelling, and editing on written word. Better tools for creating and distributing informal audio (e.g. podcasts).
And yet, with AI, we are very polarized. We’re mad when AI takes something away from people. But we’ve turned that dial on ourselves forever. Algorithm-first underwriting removed loan underwriters from smaller homogenous lending efforts. Microsoft long ago sidestepping professional editors with spelling and grammar tools in its Office suite. Nearly anyone can record and distribute a podcast today and not hire an expensive (relative to near zero) producer.