In partnership with
Hello! 👋
It's Thursday, 27th August 2026. Hello and welcome back to Bold Efforts!
For most of my career, getting better at work meant learning how to do more of it myself. You became faster at analysis, better at writing, better at making decisions, better at managing projects. Things that once took a day eventually took a couple of hours.
That was what experience looked like. You accumulated judgment, pattern recognition, shortcuts, and instincts.
While it is still true, AI changes this in a strange way.
The better you understand your work, the easier it becomes to see which parts of it a machine can do. A recruiter who has spent years learning how to source candidates can now automate much of the sourcing. An analyst can generate a first version of a model in minutes. A marketer can produce ten campaign variations before lunch. A software engineer can create working code without typing every line.
All of this is useful. But what happens when you automate the work you spent years becoming good at?
For a long time, professions gave us fairly stable identities. You were a designer because you designed things. You were an analyst because you analysed things. You were a recruiter because you found and assessed people. Your value was closely tied to your ability to perform the activity itself.
That connection is starting to weaken.
A great recruiter may spend less time finding candidates and more time figuring out what a company actually needs, judging unusual profiles, convincing great people to move, and improving the system that finds everyone else. A great analyst may spend less time assembling information and more time deciding which questions are worth analysing in the first place.
The work does not necessarily disappear but it moves. And I think that movement can be surprisingly uncomfortable.
Most of us like becoming good at things. There is satisfaction in competence. Something that once intimidated you becomes easy. Other people start relying on you for it. Eventually, it becomes part of how you think about yourself.
Then a machine arrives and does some of it almost instantly.
The natural response is to protect the task. We say the AI is not quite as good, that it makes mistakes, that it cannot understand the nuances, that there is something special about the way a human does it.
Often, that is true. But it can also distract from the bigger change.
The valuable skill may no longer be doing the task perfectly. It may be understanding the work deeply enough to know what matters, what does not, what can be automated, what still requires judgment, and how all of those pieces should fit together.
Recently, I wrote here about how AI-native businesses may be built. You perform a service manually, learn how the work actually happens, identify the recurring decisions and bottlenecks, and gradually compress more of the service into software and AI.
I increasingly think the same thing may happen to individual careers. You perform the work. You learn the system. Then you start compressing parts of your own job.
The obvious fear is that if you keep doing this, eventually there will be nothing left for you to do. I am not sure that is what happens. Once one layer of work disappears, another often becomes visible.
This may also be why traditional job descriptions are starting to feel a little dated. Most jobs are bundles of tasks. A company needed a collection of things done, so it created a role and hired someone to do them. But when the cost of performing individual tasks falls, those bundles start to come apart.
Someone in marketing starts automating research. Someone in operations starts building internal tools. Someone in strategy starts prototyping products. Someone in recruiting starts designing AI workflows. The boundaries between jobs become less obvious because the work itself becomes easier to rearrange.
Perhaps this is also why generalists suddenly seem more useful again. A person who understands the business, can talk to customers, work with data, use AI tools, make sensible decisions, and build simple systems can now cover a remarkable amount of ground.
Not because they are the best in the world at every task. Because increasingly, the task itself is not where all the value sits.
There is something exciting and exhausting about that.
If the tools keep improving, there may never again be a point where we can simply say that we have mastered our jobs. The job itself keeps moving. The thing you spent five years learning may become something you supervise rather than perform. The skill that made you unusually valuable may become widely available.
And then you have to move again.
I think this is one of the less discussed parts of the future of work. AI is not only changing productivity. It is changing our relationship with competence itself. And which is why we will see more and more 20-somethng year-old experts only increasing in number because no skill will last decades and you have to keep learning new things.
We may have to become less attached to the tasks we perform and more attached to the problems we solve. That is probably harder than it sounds.
But the unsettling part of the future of work may be that the definition of being good at our jobs keeps changing underneath us. Perhaps that has always been true. AI is just making it happen much faster. Thank you for reading.
Best,
Kartik
I write Bold Efforts every week to think clearly about where work and life are actually headed. If you want these essays in your inbox, you can subscribe here.