By SGN | Updated 13 Oct 2026
Key Takeaways: Career resilience in the AI era
- AI produces “the average of everything ever written.” The scarce skill isn’t generating output anymore — it’s judgment: knowing which version is worth keeping.
- Skills tied to a specific tool (like prompt engineering) have a short shelf life. Skills tied to people, such as selling, negotiating, earning trust, don’t.
- A forced career pivot rarely comes with a salary match. Holding out for one will cost you more than taking the reset and negotiating back up.
Start by developing ‘taste’, because AI has none.
You’ve seen this already: every Instagram food video that opens with “hidden gems, I’ve been gatekeeping this.” Same template, different avatar. Once you can hear the template, you can’t unhear it. That’s what a world without taste sounds like, and AI can now produce it on an infinite scale, for free. It’s trained on the average of everything ever written, so left to itself, it produces the average — competent and completely forgettable.
Upskilling in the age of AI isn’t just about learning the latest tool. The more durable skill is knowing what good looks like: which ideas are worth pursuing, which work feels original, and which output should never leave the drafts folder.
When anyone can generate a hundred decent versions of anything in ten seconds, the scarce skill is knowing which version to keep and which ninety-nine to kill. AI will happily present all hundred with equal confidence because it has no real opinions. Someone in the room has to have one.
I see this in my own writing. AI cleans up my dictation and fixes my grammar, and it does that well. But it has no idea which premise is funny, which angle is fresh, or which sentence to cut because it sounds like everyone else on LinkedIn.
That judgment is now the whole job, and it can’t be downloaded. It’s built the slow way, by consuming good work and making plenty of bad work until you can tell the difference – anything that takes years to build doesn’t get commoditised overnight.
In that case, how can you tell if AI will replace your job, or simply change what you’re paid to do?
If your job focuses mainly on producing a first draft of something: minute-taking and transcription, copywriting, translation, basic graphic design and deck-making, bookkeeping and data entry, summarising reports, compiling research.
Honestly, junior-level coding too; I built working websites without knowing how to code myself. All of these were respectable, paid skills five years ago. Today AI does the first 80% in seconds, and the last 20% is closing in fast.
The test I’d apply to each job is simple: does it produce output, or does it make judgment calls? AI is frighteningly good at output, but it’s still bad at deciding what’s worth producing. This includes making a judgment call between whether the result is any good, and who needs to hear about it.
Simply put, the skill of minute-taking is dying, but having the skills to know which two lines in those minutes will matter next quarter is very much alive.
How can we tell whether an emerging skill has lasting career value, or is just a passing trend?
My rough sorting rule is this – if the skill is tied to a specific tool, it’s trendy: this month’s hot AI app, the growth hack for whichever algorithm. Learn them if they’re useful today, but treat them like milk.
The durable skills are tied to people: selling, negotiating, managing a team, earning enough trust that a client hands you a budget, telling a story that makes a room care, and deep domain knowledge from a decade of real experience.
Prompt engineering is my favourite example of a trendy skill. In 2023 it was a job title paying north of US$300,000; companies genuinely hired people for it. Today the models mostly prompt themselves. That’s the shelf life of a skill built on a tool: one product update.
Whose roles are at risk of being reshaped by AI, and what should they do?
Create real work, as much as possible, while it’s still affordable. The subsidised pricing won’t last, but the skill stays with you.
My whole AI journey is this playbook in action. I tried building agents, websites, scripts, decks, music, video — small, cheap experiments, and I genuinely didn’t know which would stick. The music went nowhere. But the video experiments led to a commissioned piece for Jabra, and the websites led to client work. Each experiment might have cost me a few weekends, but a wrong master’s degree costs you two years and $60,000.
The same thing happened with my content. For years I ran what was essentially a topic account, everything about HR and careers. Once I’d exhausted that terrain, I pivoted to writing as myself, about whatever I found interesting. Some followers left; the fans stayed. No course could have told me that in advance — I only found out by shipping the pivot and watching what happened.
So, the answer to panic-learning is to make the experiments small enough that failure is cheap. Try something for a month. If nobody bites, shut it down and move on, exactly like a startup would. The commissions and the client work are the market telling you which experiment to double down on.
What should you do if AI might have reduced the market value of your role, and your next job offer pays less?
Your last salary was what the old market paid for your role, and that market is gone. We understand this instinctively with the property market. A seller who insists on their neighbour’s 2022 peak price watches the unit sit for a year, then quietly cuts.
Do the math on holding out. Say you were on a $10,000 monthly salary and an offer comes in at $8,000. Rejecting it to keep hunting for the full match costs you $10,000 a month while you wait. Hold out for a year and you’ve burned $120,000 to avoid a $24,000 annual pay cut. The spreadsheet says take the job, keep earning, and negotiate your way back up after you’ve proven yourself. After all, nobody’s final salary is their first offer.
The other realistic move is to let go of the idea that money must arrive in a single paycheck. That was my own path. I went fractional, stacked a few smaller engagements instead of hunting for one big title, and each piece of work kept compounding into the next. While the title shrank, my options grew.
Finally, how important is LinkedIn job search today? What should jobseekers do differently?
Treat it as the place where hiring decisions get warmed up, long before any job listing exists.
Mechanically, do these four things.
Fix your profile so a recruiter’s search actually finds you. They type keywords into a search box, and “passionate professional seeking new challenges” is invisible. Say what you do and the tools you use, in plain words.
Write. Post about your domain two or three times a week: what you learned, what went wrong, what you’d do differently next time. You’re demonstrating judgment in public, for free. When your CV eventually lands on a hiring manager’s desk, the reaction you want is “oh, I’ve seen this person’s stuff” and not “who is this?”.
Comment on the posts of the people you want to work for — real observations, not just “great post!” Do this consistently and you become a familiar name at zero cost. LinkedIn is the only networking event that never closes and lets you attend in pajamas.
And the reason the game is worth playing: a job listing gets a thousand applicants in a week, but the feed has no queue.
Lastly, don’t underestimate the power of building a strong personal network. Sometimes, when it comes to a job, it’s not about what you know but who you know. Most jobs on LinkedIn average more than 100 applicants – you’re just joining the snaking queue that doesn’t seem to move.
However, if you have spent time and effort to join networking events and happen to know the hiring manager from these events, you might just get a free “express pass” to skip the queue – joining a global community like Singapore Global Network is one way.
TLDR: How jobseekers can approach career growth in a structured way
Run small, cheap experiments
Build judgment, not just output
Treat LinkedIn as a stage, not a job board
Fix your profile for search, write consistently about your domain, and engage the people you want to work for. By the time you apply, you want to be recognised.
Reset your salary expectations to the market
Your last drawn salary was priced by a market that may no longer exist. Take the current offer, prove your value, and negotiate back up rather than paying the holding-out cost in lost months.
About Adrian
Adrian Tan writes and speaks on the future of work, drawing on more than 20 years running recruitment, career coaching, and outplacement businesses in Singapore. He is the Fractional CMO of Marketing Sumo.
Connect with him here.
