Jake McKeon
A warningApril 2026 · 6 min read

If you're an employee, you need to read this.

This isn't a prediction. It's already happening.

If I soften this too much, the people who need to read it will nod, close the tab, and go back to what they were doing. That's the exact outcome I'm trying to prevent. So I'm going to be blunt.

The core warning

If your job is done on a computer, AI will do most of it.

Not 10%. Not 20%. I think 80 to 90% of the execution in most computer-based roles is already replaceable. And not by average output, by world-class output. The AI isn't replacing a mid-tier performer. It's doing the job to the standard of someone genuinely excellent at it.

85%

of execution already replaceable

5-10x

output per person with AI

Feb '26

the inflection point

The simple maths on that is frightening. If AI handles 85% of execution, one person directing it can do the work of five to ten people. That's not speculation. That's what I'm doing right now inside my own business.


The mistake most people are making

Most employees think they understand AI because they've used ChatGPT.

That's like saying you understand the internet because you've sent an email.

The real shift isn't the chatbot. It's the layer behind it: automation, workflows, APIs, systems that chain tasks together and run on their own. That's where the leverage lives. And 95% of people have never touched it.

10%

of what's actually possible is visible in a chat window


Who this is really for

This isn't for everyone. It's for people who rely on a salary. Who aren't in control of the business they work for. Who assume their role will still exist in three to five years in roughly the same shape.

It's for the people I know. Friends and family who are providers, who have mortgages and kids and commitments, who are counting on a paycheck from a company whose strategic decisions they don't control.

The dividing line

Largely safe

Tradespeople. Work happens with your hands, not on a screen.

Pay attention now

Screen-based roles. Computer work. White-collar execution.


What changes in the next five years

Most computer-based work becomes substantially automated. Teams get smaller, output per person increases dramatically. And salaries for the majority of white-collar roles start to compress, not because companies are evil, but because the economics change.

The people who get hit hardest won't be the lowest performers. It'll be the middle. Solid, reliable people executing repeatable work. The kind of roles that make up most of corporate employment.


Why this happens faster than you think

At Coconut Bowls, when performance slipped or a role wasn't working, it still took months to make the call. There's emotional weight. There's legal process. There's the hope that things will turn around. I once held on to a team that needed restructuring for 12 to 18 months longer than I should have. It probably cost me a million dollars.

That friction no longer exists in AI-native businesses. Scaling down is instant. Scaling up is instant. When the economics tilt, the response is immediate.

0

Sick days

0

Underperformance

0

Management overhead

0

HR

I'm not saying that's good or bad. I'm saying it's the reality of how decisions will get made.


The flip side: this is also the opportunity

If you lean into this early, you don't become less valuable. You become dramatically more valuable. Because AI still needs someone who knows what to do with it. Someone who understands the business, who can direct the systems, who can judge the outputs. That person, inside any organisation, becomes essential.

“AI won't replace you. A person using AI will.”

Attributed to Santiago Valdarrama

The opportunity is that the bar for becoming that person is actually quite low right now. Most employees aren't doing the work. Which means if you put in 5 to 10 hours a week for the next 12 months, you'll likely be in the top 10% of AI-capable people inside your company.

The window

5-10 hrs/week

for 12 months = top 10% of AI-capable people at your company


What to do, starting now

If you do nothing else, do this.

01

Invest your own time, not your employer's

This is the most important point in the whole article. This is your leverage, your career, your future earning power. Treat it like going to the gym. Minimum 5 hours a week, preferably 10 or more. Your own time, your own money.

02

Start with your own job

Right now, in April 2026, download Claude Co-Work ($20 a month) and feed it everything you do in your role. Then ask it: what can we automate together? Start with the stuff that bores you. Reports. Emails. Admin. Content. The repeatable work. That's where the wins come fast.

03

Build one simple system, not just prompts

Don't stop at asking AI questions. Create something repeatable. A workflow, a template, a chained automation. That's the difference between using AI and operating AI. That's where the real value is.

04

Become the AI person at your company

Every company will need one. Be early. Because once it matters internally, the person who already knows what they're doing wins the role by default. You're not competing with anyone. You're just paying attention sooner.

05

Start a side project

This is where it clicks for most people. Doesn't matter if it makes money. The goal is to build something end to end, by yourself, assisted by AI. That's how you go from user to operator. If you're looking for a framework on what to build, I've written separately about asymmetric upside and why self-employment is the best bet of all.


The uncomfortable truth

This article will upset some people. That's fine. Ignoring this is worse.

I'm not writing this to scare you. I'm writing it because I'm watching it happen in real time inside my own business, and I know the people I care about aren't seeing it yet. The gap between people who use AI well and people who don't is going to widen fast. And the window to get ahead of that gap is open right now.

The default option, stay the course, hope it doesn't come for your role, is the worst option.


The real takeaway

You don't need to become an AI expert. You don't need to learn to code. You don't need to quit your job and start a company.

You just need to understand how your role changes when execution becomes effectively free. And position yourself as the person who directs that execution, not the person who performs it.

The game now

AI won't replace you. Someone using AI will.
Make sure that person is you.