Hello Reader,
Hope you had a great week. I just landed back from a trip to Denver, where I attended Zendesk Relate 2026. There were several major conferences in the US this week, but I was really glad I attended Zendesk's. I'll share why in a while.
This week, you have probably read about how a certain bank CEO's remarks when announcing a major layoff touched a raw nerve. "It's not cost-cutting. It's replacing in some cases lower-value human capital with the financial capital and the investment capital we're putting in," he told reporters.
It certainly got me thinking.
With that in mind, I want to talk about how the traditional playbook for work is no longer relevant and how we can stay relevant today. Plus, the one thing you need that hardly anyone talks about.
Why the old playbook is dead
You know the old advice about building a career in a specific niche? The thinking goes something like this: choose a field, plug away for years, collect the promotions as they come, and eventually land in a senior position. Put in the effort early, and you get to cruise along in the latter years.
That playbook is dead in the era of AI.
Here's what I see happening now: as organisations peer ahead and see the turmoil or promise of AI, they are scrambling to review their business. They are taking a hard look at their current employees and deciding to act pre-emptively.
In some cases, they are looking for tasks they predict AI can eventually do and seeking to dramatically reduce the headcounts in those departments. The objective? Either to marshal greater financial firepower from the savings or gain the leeway to invest in new AI technologies.
For others, like Cloudflare CEO Matthew Prince - who two weeks ago cut 20% of his company's global workforce - it is about paring back on employees who are not building or selling. These are the behind-the-scenes workers who track, check, and report on how the business is running. Turns out these are tasks that AI is incredibly proficient in.
So much for climbing the ladder and cruising. Seniority alone won't protect you.
Four ways to stay relevant
So while the well-meaning calls we often hear to "learn AI" sound sensible at first blush, it is also completely inadequate in protecting workers from the avalanche of cuts that is to come. How can employees stay relevant in the days ahead then?
As someone who hasn't enjoyed the comfort of a fixed month-end paycheck for almost two decades now, believe me when I say this is something I've been giving much thought to over the last 24 months. After all, I have to stay relevant between each and every client gig.
Here's what I would recommend:
- Build optionality.
- Lean into your strengths (and passions).
- Keep learning.
- Disrupt yourself.
Even the best of us can fall into unfortunate circumstances: unexpected caregiving duties, a new boss who hates you, or the company deciding to shutter an entire region. It is therefore vital to build beyond your current work or specialisation for optionality.
Don't just go half-cocked into any area that comes to mind either. Lean into your strengths and passions, in areas that you can monetise. Because we are creatures tied into rigid patterns and habits, this might take conscious effort to unearth. You will know it once you identify it though.
Finally, keep learning relevant new skills that are practicable, not just rely on your experience, which honestly, might no longer be relevant. And if you have the latitude to and you see the opportunity, move to disrupt yourself before you get disrupted. Just ask me. The bulk of my income doesn't even come from writing anymore.
I make sure I keep learning. This week, I learned more about Windows apps, the Claude Code environment, and a couple of popular tools in Linux.
- Before I boarded the flight out last Sunday, I migrated my Claude Code environment from my Windows desktop onto the Asus GX10, which runs on Ubuntu Linux. I also got tmux and tailscale set up, which essentially gave me a rock-solid, remotely accessible environment for vibe-coding while on the grueling flight to the far side of the earth.
- To track my Claude Code usage, I had an open-source Python app. Not content with that, I had it rewritten first as a .NET app before switching to native Win32. The process gave me a refresher in the intricacies of the .NET framework and programming languages such as Rust. I also got a dose of graphics-API subsystems in Windows such as GDI and GDI+.
Make no mistake about it. I relied heavily on AI to do the work, but it was also hours of trying, questioning, and learning.
The thing that hardly anyone talks about
Finally, here is one thing that hardly anyone talks about: you don't know what you don't know. These are the blind spots you cannot read up on or plan around, because you have no idea they are there in the first place.
Which is why I would encourage you to not just use AI for trivial or mundane tasks, but also lean on it for creating, for learning, and for tackling difficult challenges. Use it only for the basics, and your grasp of what AI can and cannot do stays just as basic.
Every one of my hundreds upon hundreds of prompts in ChatGPT, NotebookLM, Perplexity, Claude Code, and others, gave me a visceral and deep appreciation of the capabilities and limitations of AI that I will never get just reading about it. Because half of what I learned only surfaced through the doing; I would never have known to ask for it.
Which brings us back to my trip to Denver. In briefings and one-on-one interviews with the Zendesk CEO, Tom Eggemeier, and senior executives, I got a first-hand glimpse of a forward-thinking organisation determined to be a first mover in AI.
I saw how they are not only building AI into their highly respected Zendesk platform but are also striving to completely overhaul how they engineer software. And I get it because I have seen both the breathtaking capabilities of the latest frontier models firsthand and how insanely quickly technical debt accrues.
And even as Zendesk rolls out multiple AI-powered features, the organisation keeps looking for ways to disrupt itself. Or as Tom puts it: "I'm worried that we are not moving fast enough."
Standing still was never the safe option it appeared to be.
PS: Want to talk data centres? I'll be at Computex 2026 in June. Let's catch up there if you're in town.
Regards,
Paul Mah
