"You're Fired!" (or not?)

Hello again my friend,

You know I’ve been talking about (and using) AI for years. Since 2022, when ChatGPT was first released, we’ve been on this journey together with excitement, skepticism, and, in many cases, fear.

It’s the same for most experts. I remember interviewing founders of AI media companies years ago (like Pete from The Neuron, for example), where he described the issue with nervous laughter: “sooooo… when do the robots take over and we all lose our jobs?”

It feels like pretty soon.

There is a lot of credible commentary online suggesting that roughly half of all people in white-collar jobs (especially entry-level) may face losing their jobs within five years (the CEO of Anthropic said this).

I’ve talked about getting a job in 2026, and crushing your first 90 days. But I never addressed the underlying fear: how do we keep a job when AI (or a 21-year-old who’s good at using AI) could rip it from us? Today, we complete the trilogy.

This post will cover:

I get how this is scary. But please remember that it’s also the biggest opportunity of our careers.

Let’s lock in.

Matt Shumer, CEO of an AI company, HyperWrite, wrote (unironically) an article that’s going viral right now, suggesting we ring the alarm. He’s saying it’s time that we (the people who know and use AI) tell our loved ones, who don’t use AI, that our fears have arrived and our jobs, careers, and livelihoods are at risk. For real.

He talks about how big the advancements are, and how the rate of advancement is just getting faster and faster. How someone who tried AI and said it sucked last year may not believe how good it is now. That AI is building the next generation of AI.

But there’s also an opportunity. Matt’s clear that right now is a kind of “Goldilocks” period: it’s hot, enough people are really using AI so we can learn from each other, but not too hot. Remember, most people haven’t caught on.

His post is enough to make someone worry. He advises things like getting your finances in order and rethinking how you talk to your kids about their futures. I recommend reading it.

I’ll be going in a very specific direction around keeping our jobs.

But before that, we need to cover how jobs are changing.

To a certain extent, every knowledge-based industry will have humans. While ChatGPT, Grok, and other AI models can read and advise on how to edit a contract (for example), the humans are still necessary to sign off on everything.

But with this technology, we’ll see these industries shift where the top 10% will get much bigger, and the bottom 90% are most at risk. Roles like:

Anything with people applying their knowledge through a computer, generally speaking, will feel the pain.

Firms that provide this service as their specialization will feel it as well, if they don’t start using the technology soon (like today).

Take customer service for example. When I worked in customer support, I would answer maybe 100 - 200 tickets a day. The AI we use at beehiiv now handles around 2,000–3,000 a day.

Yes, some are escalated to a human. But the point is obvious. It’s cheaper, doesn’t sleep, and does 10x the work. Just not quite at the same level, yet.

This is the important part.

Until last month we were in a golden era. Someone who used to work full time, earning a full-time salary, would use AI to make it a half-time job with full-time pay. Maybe they took on a second (or third) job, who knows.

But that era is changing because of how roles are shifting.

If you’ve read my newsletter for a while, you know I’ve already called this out several times. The idea that someone manages people exclusively (as in they just go to meetings, follow up, and write emails to their colleagues, and that’s their job) is also fading away.

What’s replacing it is the player-coach model. The manager will oversee a team, but also do the work. The support manager will manage their team, but still respond to customers. The sales manager will also make sales, not just hype up the team. You get the idea.

Yes, every business will still need people leadership in charge. But leadership and management are two different things.

If your job is all talking to your direct reports and pushing messages around internally, in my opinion, your position could be at risk (unless you have a “C” in your title).

In my field for example (product management) there are typically three roles that work together. A product manager, a designer, and a software engineer. The shift I’m describing is a standoff between the three where it’s more of a merger than anything else.

The product manager who learns how to code. The engineer that learns how to design. The designer who develops product management skills.

Between them, the one who can add another role to their responsibilities will be in the greatest demand.

If you thought firing one person, doing one job, who knows everything about it was hard, imagine firing someone who knows two.

The same is true for most industries.

Using AI is not just about doing more in the role anymore (although that’s part of it). It’s about taking on horizontal responsibilities. Not moving up into management, moving to the side and adding adjacent responsibilities. And this doesn’t necessarily mean they’ll have to do both. But the fact is that having a product manager who can also code makes it much, much easier for them to work with engineers.

Just like how a marketing guy working at a hardware store benefits from knowing about the tools, how to use them, and the different needs of a local handyman vs a major construction company. He’s not necessarily going to be in the store explaining everything, but the fact that he could be means he’s better equipped to do the marketing job.

It’s the same principle when working on teams, and why expanding to the side can be more beneficial than moving up when looking to keep your job.

The irony is that the AI that reviews your resumes won’t look for this yet. That’s what makes it so hard. If the posting says “product manager” then the AI will seek out specialists that have that one thing in spades.

So having all your experiences related to software development or design when applying to be that product manager actually hurts you more than it helps you when AI reviews the resume. But it benefits you when a human who gets it reviews it.

My article doing a bit of the work upfront to make your application stand out makes more sense now.

But again, this is about keeping a job, not getting one. So what can someone genuinely worried about their job security do moving forward?

Right now, most people who use AI (and remember, the general population doesn’t), use it like Google. Ask a basic question, get an answer, move on.

There’s value in the ongoing use in extended conversations having it as a thought partner. But there’s something bigger. It’s the magic moment when you say “hey, can you just do this for me?” and the AI does it.

Side note: consider paying for ChatGPT, Claude, or any of these more advanced models. The free ones are alright, but not as powerful, nor as useful. It’s like $20 to get started with one.

It’s hard, but to get better at anything, we need to get the reps in. It may feel demoralizing at first, because most people suck at first, but that’s just part of the game.

AI is like the ultimate training partner. But there’s a trick to this most people don’t mention.

Here’s a breakdown of what I was doing before, why it didn’t work, and what I’m doing now.

My initial approach:

But it didn’t work because it was extremely boring and none of the work was really sticking with me.

Based on more recent advice from our CTO, I instead thought of something I wanted to build, and worked backwards from that.

The real value is in that last bullet. The point of AI is not just about asking how to do something. You’ll also want to test what it can do for you.

It’s like going to a school and paying for a teacher to walk you through a final project, except skipping the money and time. And then going to the teacher and saying “hey, can you do this for me?” and them actually doing it.

I train almost every day, but of course, you can go at your own pace.

If you’re a salesperson, and want to get into marketing, talk to AI about creating an actual ad campaign for your product. If you don’t have one, pick an existing product or create your own.

Again, it’s not just about asking quick questions.

By the end of two weeks, there’s a good chance you’ll have a full campaign done, and ready for what would effectively be a portfolio.

And if you’re looking for a new role, the bonus is to take the company you want to work for, and use AI in this way to create a project for them as part of your application. You’ll stand out.

Limit testing is another video game term, (the APM reference in my last post was very popular, so here’s another).

In games like League of Legends, you and the other players select characters with different stats and abilities, and fight. As the game progresses, there are moments when you could win that big fight, but we know it’s close. Most wouldn’t risk it thinking: “I’ve been playing for 20min I don’t want to lose on 50/50 odds like a coin-flip.”

People who took the risks learned the limits of their characters. Just like how you can learn the limits when working with AI.

Most fights were not actually 50/50, it was just too close to tell that it was 52/48. The ones who risked it, eventually noticed the details. That 1% here, that 2% there, and when/how they had advantages most people weren’t experienced with (because most never risked it).

AI works exactly the same way. It may not be able to do the entire job now, but it’s getting there. Maybe it does a bad job sometimes. That’s okay. You can always send a message like this to it in another conversation and improve.

You might be wondering: “what’s the point of training to learn new skills if you’re just going to ask AI to do it for you every time?”

That’s actually the entire point. If you can have AI do the entire job, you should. Of course, you’ll have to learn the concepts and push the buttons, but that’s fine.

There’s no way I’m going to become an “actual” developer (if you will), within one year. The team at beehiiv are by far the best in their field, many of them are what we’d call 10x engineers already. I’ll just never be a developer on their level. And I don’t have to be.

It can feel a bit contradictory.

On one hand, we have CEOs of AI companies warning us that the time many of us and our families should fear has arrived. That our jobs and livelihoods are on the line.

On the other hand, there’s me telling you to train with AI for an hour a day and expand your skills to the side, and it’ll be okay.

Easy right?

No, of course not. But it is doable. Other people have done it, so you can do it.

I’ve been talking to my friends and family too. But not to warn them that their jobs are on the line. It’s to explain the incredible opportunity that most people just haven’t caught on to yet.

For every concerned person afraid of losing their job, there’s another who could potentially double their income. I’d love for us to be on the winning end of this.

Thanks again for reading. Let me know if you have any questions.

All the best.