I Got Laid Off After 7 Years in Big Tech

I won’t talk about anything internal here. When I mention Netflix, it’s about my own job or something the company has said publicly, and the sources are linked at the bottom of this page.

Big companies are saying it out loud

Is AI actually being used at big companies? Yes, and they’re not shy about it. In November 2025, Netflix engineering leaders gave a public webinar with Anthropic on scaling AI-powered productivity across more than 3,000 developers. In its Q2 2026 letter to shareholders, Netflix wrote:

“We are increasingly leveraging these tools to deliver higher quality output more quickly and at a lower cost than traditional methods.”

That was about making shows, not writing code. But you can hear the direction, and it isn’t one company. Across tech, engineers are being pushed to use AI more and more.

At work, it felt like drowning

AI writes code faster than the blink of an eye. Somebody still has to review it, and I couldn’t review it as fast as it was being written. For me, reviewing AI’s code is harder than reviewing a person’s. Not because it’s messy, but because there’s so much of it: it often writes far more code than I would to solve the same problem, and every now and then it writes something I just don’t get, because I would have done it completely differently.

Then came the part that really got to me. I wasn’t writing code from scratch anymore, and every time I got better at using AI, it felt like I was teaching it to do my job. That’s where my head was when the layoff came. I still don’t know if AI was the reason. But I know how it felt.

On my own, it’s kind of incredible

After the layoff I started building on my own, and that’s where it got weird. I started looking at apps I’d happily pay for every month and building just the features I care about, usually in under an hour. Since then I’ve built:

• A YouTube outliers tool, in one week, that replaced the main thing I used vidIQ for. It’s how I found the idea for this video.
• This website, in an afternoon.
• A Blender plugin, in half a day.
• A computer-vision workflow that splits my daily recordings by clapboard and puts that information into the file names.

So when engineers call all of this “AI slop,” I don’t fully agree. The Go code I’ve seen it write is clean, and the more precise your instructions, the better the code you get back. Could I be wrong to trust it? Sure. But here’s what I think changed: now I’m the one choosing what gets built. I’m not trying to keep up with AI anymore. It’s keeping up with me.

What I actually believe

If your whole job is writing code, and a tool can do it faster than the blink of an eye, that job is close to being eliminated. What’s left is figuring out what needs to be built, and why. That’s still you.

I keep seeing posts saying AI means more engineers are getting hired, not fewer. My own experience says otherwise. A year ago I got several LinkedIn messages a day from people who wanted to interview me, and I wasn’t even looking. Now it’s a few a week. I could be wrong about why, but it lines up.

If you’re mid-level: stop resisting, add guardrails

Panicking won’t do anything. If you’re still resisting AI at work, I’d stop. Learn to use it to do your job, and put guardrails around it. Here’s the setup I’d start with:

1. Coding skills. Public skill repos keep Claude or ChatGPT writing clean code inside your rules. Good places to start are Anthropic’s official skills and Superpowers, which adds planning, test-driven development and code review.
2. A review skill that runs before any human looks. Mine checks clarity first: clear names, functions that say what they do, well-organized code and modern, idiomatic patterns.
3. Bots on every pull request. Security checks, your guidelines, all of it, running automatically so it isn’t all on your eyes.

That covers the code. The other half is building the skills AI can’t copy: your creative thinking and your product thinking.

If you’re a junior: build your own to-do list

Early in my career it worked like this: get a small task, finish it, get the next one. But the engineers who grow work differently. They find what needs doing, do it, and then figure out what’s next. They build their own to-do list.

For me, that started when I went to the people who used what I built and asked what they were struggling with. Their answers told me what to build next. Not the maintenance work, like keeping frameworks upgraded: anything tedious should be automated, so build that into your AI setup too.

So if you’re wondering whether it’s too late for you, my advice is the same. Stop waiting to be told what to do. Ask what problem you’re solving, and who you’re solving it for. Then keep asking how it could be better, and make it better. All of that relies on you.

Is it okay that AI does my job?

Maybe every time I got better at AI, I was teaching it to do my job. And honestly, that’s okay. There’s a lot more you and I can do besides write code. In the next video I show what I mean: I build an AI agent in Go, from scratch, and let it review my own code.

Views are my own. Nothing here is based on any employer’s confidential information.