I Promise I'm Not Insane

Will you be disappointed if we don't spend this week's newsletter on the angle of the sun? I know, I promised, and I'm sure you were all wildly eager—edges of your seats no doubt—to see these images. Some of you really understood my concern last week too, like on a visceral level, and I appreciated it very much.

Will you be even more disappointed if we instead had a quick about AI? What if I promise to tuck some of the photos way down at the end?

I know, I'm sorry. I am SORRY. I don't want to talk about AI either. It's somehow more enjoyable to take you on a tour of the deepest, most embarrassing recesses of my psyche with its labyrinthine passages of anxiety and fun obsessions and cat photos than it is to discuss AI. And yet!

We begin with a confession: I finally used Claude a few weeks ago. You might remember last year when I mentioned a conversation I had with my doctor, and I told myself I would at least try to use one of these GenAI tools. Then I didn't, because it felt gross. Then, a few weeks ago, I had to write a business proposal. So I used Claude, and in two hours I was finished doing this task that normally would have taken me days or even weeks (including procrastination time). It was genuinely helpful! I hate myself for even saying this, but it's true! I typed in all the pertinent information, and Claude took care of everything I'm not great at. It gave me the structure of the proposal, including sections I might have forgotten to include. It did all the formatting. It helped me figure out things like the timeline and fee structures. In other words, it acted as a sort of assistant by doing the boring grunt work that my brain loathes, while letting me do the actual thinking work that my brain enjoys. I had to clean it up a lot and edit the language, but again, that's work I like doing.

This past weekend, I spent some time with my dad, who as you know has been programming and doing information architecture since the 1960s. He's been using Claude lately too, and his experience—while more technical—is similar to mine. It's a very helpful assistant, as long as you feed it the correct information, prompts, and specifications. When Claude is good, which is maybe 30% of the time, it's incredible, but the other 70% of the time, it's absolutely terrible. But my dad said something that really struck me. He told me about some technical science classes he took in high school at Brooklyn Tech in the 1950s, during which he had to write specification documents. "Everything I learned about computers and programming I learned first in those classes," he told me. The way he uses these GenAI tools is essentially identical to the work he learned in the '50s. Rather than doing the technical work himself, he's instructing the AI assistant how he wants the work done.

Now, you might say to me: Duh. This is the whole problem, Leah. We're training these AI tools to do and take our jobs. Ok, but what comes next? Bear with me. What if the people who possess the natural language skills are actually those best situated to make our way through this crazy, shitty moment in which we find ourselves?

I know. You're going to look at me like I'm crazy, or you're going to get mad at me in the comments or on BlueSky the way some people did the last time I wrote about AI. I'm not defending GenAI, nor am I defending the bullshit the tech industry is putting us through, or even the creative industry. Our entire economy is being upended and oriented around a set of tools that do things like this, a real screenshot I took earlier today.

This shit doesn't work! What are we doing here?

Or around people who do things like this: I received a newsletter from a fairly well-known influencer (I am going to be nice and not mention her name publicly) that was clearly, and I mean CLEARLY, written by GenAI. The opening paragraph was a dead giveaway, and the eerily hollow tone continued for the entirety of the piece. This is a newsletter on which many, many people commented things like "beautiful writing, wow, you're so talented." This is a newsletter people pay for.

Right now, we're in a really horrible, scary moment. All of these tools are built on stolen material and everyone's losing their jobs. I am not a booster for Claude or ChatGPT or any of these companies. I'm not cheering on the industry. I refuse to go on LinkedIn and talk about how #meaningful #thework is. But I am a realist. No matter how hard we want to push back, this technology is here to stay, and I'm at least 20 years away from retirement age. Is there a way to look at this the way my 81-year-old father does? What if this isn't the doom wave of the future but a reminder to look to the past and use what we already know?

Or maybe I'm just crazy from worrying too much about the angle of the sun.

Until next Wednesday!

Lx

PS - I know this technically went out at 12:02 Wednesday night, which gives it a Thursday timestamp, but that's because my browser kept crashing, so I've given myself a 5 minute grace period.

PPS - So much for the title, right?