In the summer of 2023, I typed five words into a textbox that changed my life.
“give me a report bruh”
I was a software engineer working on an AI feature in the Google Ads org. The products this reporting tool supported generated billions in revenue per year. My job was as boring as it sounds.
I hit the blue “Generate” button. And a report appeared with our most commonly added metrics. Skipping the entire process of manually drag-and-dropping five chips from a sidebar that had hundreds if not thousands of options.
It was a stupid prompt with a simple output. But it worked.
I was shocked.
This newsletter is sponsored by Figma Make, a new all-in-one coding tool that turns that app idea you’ve always had into an interactive, clickable prototype in minutes.
I recently played around with it and built this ‘Should I Quit My Job?’ calculator in just a few minutes.
You can try Figma Make completely free.
Fast forward through 2023. We switched from PaLM 2 to Gemini, and the accuracy of the AI report builder shot up like a rocket. This shift prompted me to actually start using Gemini to help me write unit tests, debug code, and understand massive context from files I'd have spent many minutes tracing.
Sometimes, it hallucinated. Or wrote some nonsense. Or went in circles.
But a lot of times, it just worked.
Every successful interaction left me thinking the same thing: Where the hell was this when I was pulling all-nighters in school after procrastinating on my project?
I’d spent years learning to code. Four years of school. Four years in corporate. And this tool was already doing in seconds what took me hours.
Back then, when I asked Big Tech friends if they were using LLMs at work, the most common answer was a flat “no.” Most weren’t even aware of ChatGPT or Gemini beyond the name.
That was over two years ago.
“Did you guys get hit?”
Since my experience in 2023, the software engineering field has been disrupted. And I don't use that word lightly.
In January of 2023, I remember waking up to a Messenger bubble on my iPhone.
“Did you guys get hit?”
I was in a group chat with three college friends — all Googlers. One of them was the first to break the news of the initial round of layoffs. Another had checked their email and found out their role had been terminated.
In our next 1:1, my manager told me she'd "never seen layoffs" in her 11-year tenure at Google. By the second major round in early 2024, an acquaintance on my sister team had a name for it: "Sundar's Annual Christmas Present."
Layoffs are commonplace now across tech.
But the downstream effects are where the real damage is.
When I called my father last week, he told me that the university where he teaches has slashed its CS acceptance rate nearly in half.
Besides me, my software engineering friends are still employed. But the psychological safety across all knowledge work is at an all-time low. CS graduates face 6.1% unemployment. Teams are getting offshored. Layoffs constantly loom. Everyone feels it.
And yet… I think that for a certain type of person, this might be the best time to be alive.
危机
There's a word in Chinese: 危机 (wēijī). It directly translates to "crisis."
The first character 危 (wéi) means risk. The second character 机 (jī) means opportunity.
Every crisis is danger and opportunity fused together.
And here’s why I believe it’s relevant to you.
AI advancement isn’t going to stop.
If Google pauses, Meta doesn’t. If the US pauses, China doesn’t. AI is the new arms race both in capitalism and in military power. The incentives are too massive. The consequences of losing are unfathomable.
And so there’s never been a more dangerous and yet more opportunistic time to be alive.
Because here’s the pattern I keep seeing.
The New Paradigm
In university, we took a class that used C.
C is a low-level language. It’s used to build operating systems like Linux and Windows. You have to manage memory and do precise and tedious work.
Because of that, C became the foundation. Many higher-level languages, including Python, were built on top of it. One line of Python can trigger thousands of lines of C that you don’t have to think about.
When C built the world, programmers were craftsmen.
They worked close to the machine. Every instruction was explicit. Precision and manual control were the job.
Then “higher level” languages like Python arrived.
The craftsman didn’t disappear. Their skill abstracted up a layer.
They became designers who focused on structure and behavior, while the language handled the repetitive mechanics.
Now AI coding is here.
And the designer won’t disappear either.
The work abstracts up again.
They become architects deciding what should be built, setting constraints, and correcting outcomes. A friend of mine calls Claude Code and Codex his interns.
The entirety of knowledge work is going through this shift right now. We are all gaining access to 24/7, on-demand interns.
They’re not A-listers by any means, but they have a broader knowledge set than any living human and can do certain low-level tasks like no other.
As these interns turn into full-timers, and the full-timers cost less and less to spawn and maintain, one human turns into a full team, or a whole organization, or an entire company.
So the skill that's appreciating isn't implementation. It’s taste.
What This Means For You
If the paradigm is shifting, what do you do about it?
The way I see it, there are two responses people currently have:
Wait. Keep your head down. Hope the storm passes. Stick to “the old ways”. I understand this instinct because I lived it for over a year at Google before I left. Many people I know even in Big Tech still don’t use AI (startups are a completely different story). But inaction compounds.
Or ride the wave. You don’t have to love every aspect of what’s happening. But you learn the tools. You build something small. You develop a feel for what AI can and can’t do. Not from reading hype articles or content, but from using it yourself.
I think the second group is going to be disproportionately rewarded. Not because they’re smarter, but because they moved first and compounded skills+taste faster.
If you want to start, the bar is lower than you think.
Three Levels of AI Proficiency
Level one: Automate. Take one annoying, repetitive task from your job and hand it to AI. Maybe it’s summarizing meeting notes, drafting a follow-up email, or making sense of a messy spreadsheet (I recently pasted 1,400 rows of survey data into Claude and had a full analysis in minutes).
Level two: Learn. Find a pdf of a book you love and upload it into Claude or ChatGPT and have it respond to you as if it were the author. Start using AI not just as a search engine, but also as a thinking partner.
Level three: Build. Tools like Figma Make and Lovable now let you turn an idea into a working prototype in minutes. No coding required. I built a “Should I Quit My Job?” calculator just playing around with it. A year ago that would’ve taken me days as a software engineer. The gap between “I have an idea” and “I have a thing” is collapsing.
This is what gets me excited about all this: maybe your favorite authors become interactive mentors you can talk to anytime. Maybe AI handles the mundane parts of your job so you can focus on the parts that actually require you. Maybe it connects you with compatible people who share your beliefs and values. (OK, that last one's admittedly a bit Black Mirror but I need a girlfriend.)
What made me leave corporate wasn't just the search for authenticity and meaning. It was also my belief that the future of work is fundamentally changing. That this is the most leveraged time in human history and that every ounce of extra compounding matters. And so clinging to a career where I couldn't excel due to lack of fulfillment and interest would hurt me more than any short-term financial cost.
Whether the future I bet my career on pans out is yet to be seen.
But if I’m still sending you these letters years from now, I guess you’ll know!
For Terry
Last month, during a family trip to Atlantic City, I spent a few hours with my 8-year-old nephew at an arcade.
Terry loves video games (a trait his father and I share). He couldn't care less about AI or careers or any of this. We played games together on machines that'll probably be long gone by the time he's my age.
I kept thinking: by that time, the world will look nothing like it does now. The rules I followed — get good grades, get into a good school, get a good job — might be completely irrelevant.
If I could tell him one thing, it’d be this:
The world you’re growing up in is uncertain. It’s changing faster than we can keep up with. Mom, Dad, and Uncle Jim don’t know what’s right and what’s wrong.
Trust yourself to figure it out.
But I didn’t say that to him. He’s 8. He was too busy beating his high score.
He was already living it.
So I said it to myself.
And now I’m saying it to you.
To figuring it out,
Jim
P.S. If this piece hit something in you, here are a couple of my other letters that go deeper on what I touched on today:
I Quit My $350K Job at Google and My Life Got Worse. I'd Do It Again.