I want to talk about the subject that Clint and I discuss over breakfast and that I avoid in conversation, but really want to push back on. It’s about the kind of comment so many people make - AI is going to take our jobs.
As many of you know, I’m 50 and I’ve been job-hunting for a couple of months. So far, no interviews, only generic emails telling me “While your resume is impressive, we’re going in another direction” or whatever.
First of all, don’t fluff me with flattery about my resume. I know you didn’t read it.
AI did.
Also, what am I, stupid?
It’s so insulting to receive this falsely flattering email the next day telling me how amazing my resume is when the first email said you were inundated with resumes and your response time might be slow.
At least be honest.
In your rejection email, tell me that your AI assistant scanned my resume and didn’t like that I have far too many years of experience or not the right keywords to meet its algorithm.
We’re all trying to play the damned game - at least tell me what the rules are. You’re missing out on a lot of great potential employees. And you’re just pissing me off.
So then I think: Maybe they don’t really want employees. Maybe they are trying to hire me based on my AI algorithm-beating skills. Maybe they’re trying to get to my AI through me.
Fuck, I don’t know.
What I do know, though, is that AI isn’t taking our jobs. It’s changing them. And we need to change, too. The problem is that this was mostly just a feeling I’ve had since I started the job search. And then, this morning, in one of my nerd newsletters, I found this article:
I don’t know what it was about the title that intrigued me, but I opened it and read the first few lines:
AI was supposed to replace human labor.
It did the opposite.
For the first time in history, humans are cheaper than software.
As I read this article I thought, “Yes! This is the kind of information I’ve been looking for to back up my opinion about AI actually creating more jobs instead of removing them.”
If this is cognitive bias, so be it. The result is what matters. Read on.
As I kept reading the article, I realized that I understood a majority of the jargon.
Ho, ho! What’s this? I understand tokens, agentic, context windows, etc? Huzzah!
Sure, I’ve worked in the tech industry since I was in my mid-twenties, but these weren’t terms that we used. The kind of jargon I learned was “I’m just going to spin-up another virtual PC” or VPN, GUID, sharedrive, and binary.
There was a time when we still used the word analog, as in, “I still wear an analog watch instead of digital to keep up my skill of being able to read a clockface.” It’s likely a necessary skill for the apocalypse, I’m sure of it.
The reason I understand much of the jargon in this article is because I’ve been busting my ass over the past couple of months to learn as much as I can about AI. Or, at least, as much as is relevant to a person like me who isn’t particularly interested in becoming an engineer or developer. I want to understand the words so I can talk to those nerds, but I don’t want to be the builder. I want to be the teacher.
As the teaching type, I listen to the limiting or generalized things people say and try to figure out ways to gently push back. You see, about 20 years ago, I realized that I really loved the part of my job where I got to teach nurses and doctors how to use Electronic Health Records (EHR) - though back then we called it Electronic Medical Records (EMR). (EMR turned into EHR because Health is more all-encompassing than Medical.)
When I realized that I love teaching people how to do tough stuff, I began to regularly say, “I prefer to teach people how to fish instead of just giving them a fish.” For those of you unfamiliar with the original saying, it goes like this: Give a man a fish, you feed him for a day. Teach a man to fish, you feed him for a lifetime.
And it applies to all kinds of things.
As a yoga teacher, I can’t tell you how many times I’ve heard this: “I can’t touch my toes so I can’t do yoga.” My answer: so, bend your knees.
“I can’t learn…”
Insert whatever new thing a person’s afraid to start because of whatever limiting fear they have. Fear of being a newbie. Fear of asking questions. Fear of potentially never understanding or learning it fully. Fear of being wrong. Fear of being in a classroom full of younger people.
First of all, stop saying the word “can’t”. What you really mean is “won’t”. Remove “can’t” and insert “won’t”, then think about why you won’t learn the thing, and then tackle that part. The thing you’re afraid of can be learned. French, yoga, tap dancing, EHRs, cell phones, AI. If you want to learn it, you will. You can.
So. As I read this article, it occurred to me that many people might start to read it and then stop. Their eyes will glaze over at the first word or acronym they don’t understand.
Ah, ha! This is a perfect learning opportunity!
Go read the article. Here it is. It’s dense and a little self-impressed, but the points are valid. When your eyes glaze over, go to step 2.
Ask it to summarize the article for you, in plain language, and flag anything that’s jargon — “tokens,” “context window,” “agentic,” “evals,” whatever. Then ask it to explain those terms like you’re smart but new to this, because you are smart, and you are new to this, and both of those things can be true without it being an insult.
Ask it where to go to learn the concepts you didn’t know. Not a 40-hour bootcamp. Something small. I’d point you toward Andrej Karpathy’s free “Deep Dive into LLMs” video on YouTube for how this stuff actually works under the hood, and DeepLearning.AI’s “Generative AI for Everyone” on Coursera (or DeepLearning.AI for free) if you want something structured and short.
I read the article, had the inspiration to suggest all of this to you, and then had my Claude AI summarize and suggest things for me.
Below is the summary Claude made for me:
The piece’s whole point is this: AI didn’t replace human labor. It made humans cheaper than the software. Companies are burning enormous money on tokens because almost nobody in the building actually knows how to direct AI well. The author’s estimate is something like 1 in 100 people can articulate a process clearly enough for AI to actually execute it. Everyone else just generates what he calls “loops” — AI calling itself over and over trying to fix a task that was never defined properly in the first place. Which is, if you sit with it for a second, just expensive confusion wearing a labcoat.
The article’s actual closing argument is that the next massive economic opportunity isn’t more AI infrastructure — we have enough of that — it’s people who can translate a real business’s actual know-how into something AI can reliably run.
By pasting the link into GPT and asking for a summary and suggestions, you are now miles ahead of most of the people in the country.
What I highly encourage everyone to do is at least take a course in prompting AI. Learn how to ask AI for what you want. The irony is that by learning how to prompt AI, I’m learning how to communicate better with humans.
I’m learning to not assume that someone knows what’s in my head or my life experience. I’m learning how to describe well and thoroughly. I’m learning how to be bold and ask for exactly what I want.
This applies to the job hunt, to landlords, and to life partners. Salary, exit timeframes, and “dessert” ~ahem.
I’m not writing this from some serene, enlightened place where I’ve made peace with the job search and the studio move-out deadline and the fact that my resume apparently vanishes into an ATS black hole more often than not. I’m writing this from the middle of it, tired, a little scared some days, and deeply unwilling to let a hype cycle convince me that two and a half decades of hard-won, self-taught expertise is suddenly worthless because a chatbot can write a paragraph.*
*Claude wrote the last paragraph. I asked it to sum up my post in my voice taking into consideration my humor and sass, and all it knows about me. That’s prompting. These are the things that are going to help us move forward with AI. It need humans to train it and to manage it into for it to assist us.
The other truth about AI is that it will only take us over if we let it. I could have had Claude write my entire post - I did ask it to, just for funsies. I read the output, and it was okay, but I will always prefer my own messy, authenticness. I’m a writer. I love to write. I need to write. I’ll be damned if I let AI take all of that from me.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.