Welcome to the 14th edition of Work in Beta.

In this edition, we look at what Anthropic's new workplace data actually means for your role and give you a 15-minute exercise to figure out which parts of your job are most exposed to AI.

Also, if you are looking to build your individual or organisational system with AI, scroll down to the bottom of the newsletter to know more and connect with us.

So, let’s dive in!

Anthropic published something that changes the AI-and-jobs conversation. Not an opinion piece. Not a prediction. Actual data from real workplace usage of Claude across ~800 occupations.

The headline finding: in Computer and Math occupations - programmers, data scientists, QA engineers - 94% of tasks could theoretically be done by AI. Only 33% actually are.

That gap exists across every knowledge work occupation. And it's the most important number in your career right now, because it won't stay this wide forever.

Everyone's asking "will AI take my job?" Wrong question. The right one: which parts of your job are quietly becoming worthless while you're still doing them manually?

Forget the old automation narrative: robots replacing factory workers, AI coming for entry-level jobs first. The Anthropic data shows the opposite.

If your work happens on a laptop - writing, analyzing, summarizing, formatting, researching - you're in the exposed category. If it requires physical presence -cooks, mechanics, bartenders, lifeguards - you're not. Here's how we'd put it: the real divide isn't white collar vs. blue collar. It's screen work vs. physical work.

And here's the part most people miss: workers in the most AI-exposed roles earn 47% more than those with zero exposure. AI isn't starting at the bottom of the pay scale. It's starting wherever the work lives on a screen.

Companies aren't cutting roles - they're quietly expecting more from fewer people.

The tasks AI handles well today - drafting, summarizing, formatting reports, first-pass research - don't disappear from the company. They disappear from your job description. What's left is the judgment, the relationships, the decisions. That's great if you've been building those muscles. Not great if 40% of your week was the stuff AI just ate.

The data backs this up: overall unemployment in AI-exposed roles hasn't budged. No mass layoffs. No crisis. But that doesn't mean nothing is happening, it means the shift is structural, not dramatic. Roles are being hollowed out, not eliminated.

There are early signs that companies are hiring fewer young workers into AI-exposed roles. Not because they stopped needing the work, but because the entry-level tasks that justified those positions are shrinking. Drafting, data entry, first-pass research - the tasks that used to be a junior person's entire job.

And it's not just juniors. Mid-level roles aren't disappearing, but they're being redefined. The people who use AI effectively absorb more scope. The ones who don't become hard to justify at their salary.

That's the preview at every level: when the tasks filling your calendar become automatable, the calendar gets reassigned. To AI, or to someone who already uses it.

The overall numbers look fine. No unemployment spike. Sounds reassuring.

But think of it like a thermometer that only reads whole degrees. You won't see 98.6 tick to 99.4, but that doesn't mean the fever isn't building. The data can only catch big, sudden shifts. Slow, gradual erosion? It wouldn't show up yet.

The junior hiring slowdown is the first decimal point moving. The question is what the next few readings show.

Here's what to do with this information. Not a career overhaul. A 15-minute exercise that shows you exactly where you stand.

Step 1: List your 8-10 core recurring tasks. Not your job title - your actual tasks. What fills your calendar week to week.

Step 2: For each one, ask: "Could AI do 80% of this today?" Sort into three buckets:

Step 3: Count the split.

In our experience, people who do this exercise are surprised by how much of their week sits in bucket 1, and they're still doing all of it manually. That's the gap the Anthropic data is measuring, playing out in your specific role.

The instinct is to avoid AI - learn nothing, change nothing, hope the wave passes. That's the worst thing you can do. The roles that shrink fastest belong to people who don't notice which tasks have already moved.

The right response: deliberately shift your time toward bucket 3 while using AI for bucket 1. The goal isn't to work less, it's to spend your hours on the work that actually requires you.

The strategic move most people miss: become the person who builds the systems that handle bucket 1 tasks for your team. Not just using AI yourself, designing how your team works with it. We wrote about this shift from doing work to building systems in our March edition. That's not a threat to your role. That's a new skill and it sits squarely in bucket 3.

Your value isn't in the tasks you complete. It's in the judgment you apply and the decisions you make. The task map shows you where to invest.

The Anthropic data tells one clear story: the gap between what AI can do and what it is doing is massive. That gap is your window - not to panic, but to position yourself on the right side of it.

Map your tasks. Know which bucket each one sits in. Start moving.

Knowing how to prompt well is roughly 5% of what it means to actually work with AI. The other 95% - context architecture, workflow compression, thinking behaviors, tool orchestration - is where your workday actually changes. Not "I got a better first draft." More like "I built a full client proposal in one sitting that used to take my team three days."

That's what we work on with professionals and teams through Work in Beta.

If you're curious what the other 95% looks like, reach out to us here.