The AI Apocalypse Isn’t What the Clout Chasers Told You
Every time I talk about AI capability, someone inevitably throws out the same lazy line: “AI is going to take all our jobs.” And every time, I find myself wishing people would look at a single chart, just one, before firing off their next viral hot take. Because the story isn’t as simple as the clout-chasers make it out to be, and frankly, their performative doom is getting old.
Let’s
start with capability. I showed a chart recently, one of those almost comically
dense ones where the presenter says “don’t worry about the details, just look
at the black line.” That black line represents human-level performance. And
every colored line beneath it represents some AI system inching, then
sprinting, toward that line. Over the years, these lines get steeper, almost
vertical, because models aren’t just improving; they’re accelerating toward
human-level capability faster each cycle.
But
here’s the important part: those lines represented old-school AI, the
kind that recognizes speech, or handwriting, or plays chess… one task, one
domain, very smart but very narrow. This is the AI most people still think
about when they confidently announce their opinions at panel discussions after
reading two Medium posts.
What
we’re facing now isn’t narrow AI. It’s general-purpose AI, systems like LLMs
that can take exams across multiple fields at PhD-level performance. And as of
late 2024, they were already there. Not “PhD in one field.” PhD in everything, law,
biology, economics, philosophy. All of it. And that chart is outdated by a
year. I don’t even want to imagine where the dot sits today.
This
isn’t just software iteration. This is physics. Around 2010, the slope of what
we thought was Moore’s Law suddenly snapped and shot upward. Why? GPUs.
Originally designed to draw prettier shadows on video games, then hilariously
repurposed for crypto mining, they’ve now become the fuel for deep learning.
Once GPUs switched from rendering dragons to crunching text, images, audio, everything,
that’s when AI crawled out of the lab and started passing bar exams.
A
year ago, people were still whining about hallucinations. Today, the frontier
is AI writing production-level software. As a software developer myself, I
didn’t see that one coming. Ironically, one of the first “at-risk” professions
in this supposed job apocalypse is the very profession building the apocalypse.
But
here’s where nuance matters, and where the clout-chasers lose the plot
entirely.
When
you look at exposure to AI across jobs, something surprising happens: the higher your salary, the more your
tasks are exposed to AI. Which should make a lot of
white-collar fearmongers pause before lecturing Uber drivers about automation.
We used to think robots meant factory floor automation. Instead, AI came for
the lawyers, consultants, accountants, and designers first. Even I, an
accountant by training, am feeling pretty validated for never practicing.
The
IMF chart I referenced makes this painfully clear. Architecture, engineering,
science, insurance, computer programming, these are the most exposed.
Meanwhile, landscapers, veterinarians, and construction workers are chilling at
the bottom, practically immune in the near term.
But
the real insight isn’t exposure. It’s complementarity.
Every job has three components:
- The part AI can’t do (green)
- The part that complements AI, tasks
where AI makes humans more powerful (blue)
- And the dangerous part, tasks AI
duplicates, where you could theoretically replace the human outright (red)
Sort
jobs by the red bars, and guess who’s suddenly on the chopping block? Office
and business support. Information services. Legal and accounting. Middle
management. If your job mostly involves moving information from one place to
another, congratulations: a machine can also do that.
The
irony is thick. We used to mock low-level workers as being replaceable. Turns
out the least replaceable
people in the economy today are landscapers and transport workers in countries
like the Philippines. Even with self-driving tech, drivers might be augmented, not
replaced, for a long time. Meanwhile, whole floors of office workers generating
slide decks? Those jobs have a giant, blinking “caution” sign above them.
Which
brings me to the real problem with doomsday narratives. They treat “jobs” as
monolithic. Humans don’t get replaced. Tasks do.
We
don’t have switchboard operators anymore. We don’t have ice cutters. We don’t
have human “computers”, people who literally did math for a living.
Tell
someone in 1900 that their prestigious job as a human computer would eventually
become a handheld device, and they’d laugh you out of the building. People
cling to titles; technology disrupts tasks.
Today,
the saddest job I can think of is the elevator operator. A human whose entire
job is pressing the button you could
press, inside a four-story building. And yet, in the same city, we have smart
elevators where you can’t press anything because the system already knows where
you’re going. That’s the absurdity of technological transition: legacy habits
coexisting with futuristic systems.
So
no, AI won’t “replace everyone.” That’s not how economies evolve. What we will do is reassign,
restructure, and rethink the tasks humans add value to.
The
real question isn’t “Will AI take our jobs?”
It’s: Which parts of your job do you insist
on doing manually, even when a machine does it better? And why?
In the end, intentionality matters. Human-in-the-loop matters. We decide which tasks we keep, which we delegate, and which we redesign entirely. But if there’s one thing I’m certain of, it’s this: the loudest voices predicting total human obsolescence are usually the ones who understand the least about how work actually works.
And as the charts show, if the clout-chasers bothered to look, this future is a lot messier, stranger, and more interesting than their soundbites ever admit.
About Me:
Dominic “Doc” Ligot is
one of the leading voices in AI in the Philippines. Doc has been extensively
cited in local and global media outlets including The Economist, South China
Morning Post, Washington Post, and Agence France Presse. His award-winning work
has been recognized and published by prestigious organizations such as NASA,
Data.org, Digital Public Goods Alliance, the Group on Earth Observations (GEO),
the United Nations Development Programme (UNDP), the World Health Organization
(WHO), and UNICEF.
If you need guidance
or training in maximizing AI for your career or business, reach out to Doc via https://docligot.com.
Follow Doc Ligot on
Facebook: https://facebook.com/docligotAI
