October 29, 2025 • 40 mins
The source provides an alarming overview of the ongoing Artificial Intelligence (AI) job apocalypse, asserting that AI is rapidly eliminating human labor across numerous sectors, unlike previous technological shifts. It details the mechanics of mass obsolescence, showing how exponential efficiency curves and drastic cost reductions in AI are displacing jobs in fields such as truck driving, customer service, creative industries, and white-collar professional roles like radiology and legal research. The episode quantifies the resulting human toll with stark statistics and personal anecdotes, noting severe economic fallout like skyrocketing income inequality and projected fiscal crises due to collapsing tax revenue. Furthermore, the source critiques the policy vacuum, arguing that current proposed solutions like retraining, Universal Basic Income, and soft regulations are proving woefully inadequate to address the accelerating job destruction and subsequent societal catastrophe.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Welcome back to the Deep Dive. We are here to
take the stack of source material you shared with us,
the research papers, the internal memos, the corporate projections, and
will really extract the critical knowledge you need. Today. We
aren't discussing potential future disruptions. We are diving headfirst into
what these sources characterize as an accelerating economic crisis.
Speaker 2 (00:22):
Yeah, and the premise here is pretty aggressive, almost terrifying,
you could.
Speaker 1 (00:26):
Say, it really is. It demands immediate attention. The core
argument artificial intelligence isn't just augmenting human labor. The sources say,
it's obliterating.
Speaker 2 (00:36):
It obliterating, that's the word they use, and on a scale,
and importantly, at a speed we just haven't seen since
what the Industrial.
Speaker 1 (00:43):
Revolution exactly two centuries ago.
Speaker 2 (00:45):
And that speed, you know, that's the critical difference. It's
what makes us Deep Dive so urgent for you, the listener.
I mean, think about it. The initial warnings just a
few years back, they projected maybe seventy five million jobs
displaced globally by AI.
Speaker 1 (00:58):
Which sounded like a lot at the time.
Speaker 2 (01:00):
It did, but the sources point to this grim acceleration.
The updated twenty twenty five World Economic Forum forecast, it's
drastically bleaker. We're now talking eighty five million jobs globally
displaced by twenty twenty.
Speaker 1 (01:15):
Seven, eighty five million by twenty twenty seven. That's a
massive jump in projections over a really condensed timeline.
Speaker 2 (01:22):
It is it suggests the catastrophe, or well the shift
is unfolding much faster than anyone predicted.
Speaker 1 (01:28):
And what's crucial here, and the sources really hammered this home,
is that this isn't just theory anymore. It's not some
future fear. They argue, quote, the apocalypse is not coming.
Speaker 2 (01:38):
It has arrived, right, So we need to ground this,
make it real. What are the specific current corporate actions
that show this is already happening? Where's the reality check?
Speaker 1 (01:46):
Okay, let's look at some big names.
Speaker 2 (01:48):
We can start with the giants. Yeah, look at IBM.
Back in twenty twenty three, they announced they were freezing
hiring for about seven eight hundred back office.
Speaker 1 (01:55):
Roles, right, the administrative stuff.
Speaker 2 (01:57):
Exactly, and they specifically cited AI's capacity to handle those
routine tasks. That's not waiting for displacement. That's a preventative
strike against labor.
Speaker 1 (02:04):
Okay, IBM freezing roles. What else.
Speaker 2 (02:06):
Even more dramatically, you have Klarna, the huge Swedish fintech company.
They used AI to actively shrink their existing workforce by
fifty percent. Fifty percent, how after successfully deploying AI for
customer service interactions. And this wasn't, you know, just letting
people leave naturally, this was systemic replacement.
Speaker 1 (02:28):
That distinction feels really important workforce shrinkage versus just not hiring.
Speaker 2 (02:33):
It's vital. It means the company is actively taking salaries
benefits off the balance sheet entirely. It's a direct substitution.
Speaker 1 (02:41):
So where's the most compelling evidence that AI is really
out of the lab and hitting the mainstream.
Speaker 2 (02:46):
Well, it's interesting. You see it in two sort of
very different sectors simultaneously, transportation and high end diagnostics.
Speaker 1 (02:53):
Okay, you claim.
Speaker 2 (02:53):
So transportation, we now have autonomous Waimo fleets. They're logging
millions of driverless miles in places like Phoenix, San Francisco,
dense urban area, deplacing.
Speaker 1 (03:02):
Ride share drivers, delivery drivers.
Speaker 2 (03:04):
Thousands of them daily. And at the same time, you
look at the medical field, a really high skill domain,
right radiology.
Speaker 1 (03:09):
Yeah, it requires years of training.
Speaker 2 (03:11):
Google's DeepMind AI now routinely surpasses the diagnostic accuracy of
human radiologists.
Speaker 1 (03:17):
Wait, surpasses, not just matches surpasses.
Speaker 2 (03:20):
So when the machine is cheaper, faster, and more accurate
than a highly paid specialist, will the economic floor for
that job just disappears instantly?
Speaker 1 (03:30):
Okay, So our mission today for you listening is to
give you that necessary shortcut through all this evidence. We're
going to dissect the mechanics. First, how AI actually destroys
this value, focusing on this concept of cheaper cognition. Then
we'll try to quantify the human toll. Look at specific
industries that are already, as the sources say, in freefall, trucking,
(03:52):
customer service, even white collar jobs.
Speaker 2 (03:54):
Yeah, the place is feeling the immediate impact.
Speaker 1 (03:56):
And finally we have to expose the policy vacuum, the
frame total failure of governance to keep up with this
accelerating shift.
Speaker 2 (04:03):
Which you know might be the most dangerous part of
all this, the lack of response.
Speaker 1 (04:06):
Okay, let's unpack this core mechanism. Then it's one thing
to say AI is smart, right, but we need to
understand how it delayers this well economic annihilation. The sources
keep coming back to this phrase cheaper cognition.
Speaker 2 (04:19):
Right, if you think about the last say fifty years
of the digital age, it was all defined by Moore's.
Speaker 1 (04:24):
Law, cheaper transistors, more computing power.
Speaker 2 (04:27):
Exactly, cheaper processing. But the next decade, maybe longer, it's
defined by AI providing cheaper, faster cognition.
Speaker 1 (04:34):
Cognition like thinking.
Speaker 2 (04:36):
Thinking, writing, analyzing, diagnosing, translating the fundamental output of millions
of white collar jobs, specialized jobs. So when that fundamental
output becomes nearly free, I mean, what happens, the economic
justification for paying millions of salaried workers just evaporates.
Speaker 1 (04:53):
Why pay a specialist analyst one hundred thousand dollars a year.
Speaker 2 (04:56):
When in LLM, a large language model can perform the
synth sist of maybe one thousand analysts simultaneously and it
works two hundred and forty seven, No benefits, no salary,
doesn't need a desk, and the source of stress.
Speaker 1 (05:08):
This isn't just a marginal improvement. It's not like getting
ten percent cheaper.
Speaker 2 (05:11):
No. No, it's a structural collapse in costs. That's why
the speed is so terrified.
Speaker 1 (05:14):
And this is where it gets really interesting. We have
a data point that just crystallizes the economic difference. I
want you, the listener, to really pay attention to this
cost collapse for routine language processing. Get rewind just a
bit twenty nineteen. If you wanted, say, one thousand words
of professionally drafted, human quality text generated via an API.
Speaker 2 (05:36):
Like for marketing copy or a basic.
Speaker 1 (05:38):
Report exactly, the cost was roughly sixty cents zero dollars.
Sixty cents.
Speaker 2 (05:43):
Sixty cents seems reasonable.
Speaker 1 (05:44):
Now, fast forward to twenty twenty five. Open AI's latest
model OH one preview. It delivers equivalent often actually superior
output for wait for it, zero point zero zero zero
one two dollars from sixty cents.
Speaker 2 (05:57):
That's a ninety nine point nine to eight percent price drop.
Speaker 1 (06:00):
Six years, six years. That's staggering. That's not evolution, that's
instantaneous collapse, right, I mean, think about previous disruptions. Getting
in ninety nine percent costroduction took decades, like manual typesetting
to digital printing back.
Speaker 2 (06:12):
This speed is the thing, the novel destructive element. Absolutely so,
if your whole livelihood depends on routine language generation, drafting
legal summaries, synthesizing earnings reports, writing ad copy, your economic
floor just vanished. Almost one hundred percent drop, and it
happened while you were like still negotiating your annual rais.
Speaker 1 (06:32):
Now what's truly frightening, as the sources make clear, is
that this pattern, this radical cost collapse, it's not just text.
It's universal.
Speaker 2 (06:40):
Yeah, it's happening across basically every cognitive domain. At the
same time, the economic incentive for capital earners, for businesses
to switch is just overwhelming.
Speaker 1 (06:49):
Now give us some examples beyond text.
Speaker 2 (06:51):
Okay, take image generation. A human illustrator might charge what
four hundred dollars or more for a commercial grade image
plus licensing fees revisions standard practice mid journey V six.
It produces equivalent, sometimes even more creative, high resolution output
for about zero four cents per image four cents four
cents versus four hundred dollars.
Speaker 1 (07:11):
So the decision to hire a human it's not just
about saving a bit of money anymore. It becomes frankly
an economic absurdity for many routine tasks.
Speaker 2 (07:19):
Okay, images, what else coding? Coding? Yeah, gethub copilot. It's
now reportedly responsible for writing forty six percent of developer's code,
almost half almost half. It cuts the intellectual heavy lifting
for basic programming dramatically, and this directly led firms like
Accenture to report stunningly a thirty percent cut in demand
(07:40):
for junior programmers.
Speaker 1 (07:41):
So the entry level cognitive jobs are getting hit first.
Speaker 2 (07:44):
The first casualties. Absolutely, And if your job involves a
sophisticated data analysis, spreadsheets, modeling, andthropics, claude three point five
sonnet model performs the same deep multi variable analysis four
hundred times faster than a high highly skilled human analyst
four hundred times.
Speaker 1 (08:03):
The human just can't compete, not on speed, not on cost,
not on scale. So this efficiency isn't just a threat,
it's becoming an economic mandate for any business that wants
to stay competitive very much, which leads us directly to
what the source is called the substitution flywheel. This explains
the mechanism for mass job loss, why it scales so
incredibly quickly. It's a critical three step process, right.
Speaker 2 (08:23):
The flywheel starts with the first step, initial substitution. This
is the bit we see most clearly.
Speaker 1 (08:29):
One AI system comes in and.
Speaker 2 (08:30):
Immediately replaces let's say ten human jobs, a team, a department.
Speaker 1 (08:35):
Maybe okay, step one, but the.
Speaker 2 (08:37):
Real danger, the reason it scales is the second phase,
scaling via cloud deployment. Because AI systems are fundamentally software.
Speaker 1 (08:45):
You can copy them infinitely deploy them anywhere exactly.
Speaker 2 (08:47):
That one AI system that replaced ten people, it can
be instantly distributed globally through the cloud, instantly replacing ten
thousand human jobs, often without moving any physical infrastructure or
needing any local train. It's centralized and incredibly.
Speaker 1 (09:02):
Swift ten thousand from one system. And then step three compounding, ah.
Speaker 2 (09:07):
The compounding phase. This is the part that means the
human worker, well, they can't realistically catch up even with retraining,
because the AI system improves itself through reinforcement learning feedback
loops gets better every single day, widening the performance gap
between the machine and the human.
Speaker 1 (09:24):
While the human needs years to learn a new skill.
Speaker 2 (09:27):
The AI needs milliseconds to update its model based on
new data. It's a constantly accelerating gap.
Speaker 1 (09:33):
Do we have a specific example of this flywheel in action?
Something concrete?
Speaker 2 (09:38):
We do, and it's a chilling one. The sources pulled
it from an internal document leak. Apparently Amazon they're fulfillment centers.
Speaker 1 (09:45):
Okay, Amazon warehouses.
Speaker 2 (09:46):
A single AI system was rolled out its job optimize
logistics and crucially labor scheduling, and the result the elimination
of a staggering one hundred and thirty thousand seasonal warehouse.
Speaker 1 (09:59):
Jobs one hundred and thirty thousand in the.
Speaker 2 (10:01):
Fourth quarter of twenty twenty four alone, just like that,
mass obsolescence in action, one centralized diploma decision, wiping out
thousands upon thousands of jobs instantly right before the peak
holiday season. Two.
Speaker 1 (10:13):
Wow, that illustrates the scale perfectly. Okay, so we've established
the economic mechanism, this relentless drive for cheaper cognition amplified
by this substitution flywheel. Now let's pivot. Let's move from
the abstract cost collapse to the specific tangible industries that
the sources describe as being already in well free fall.
(10:34):
Let's start with transportation. This one feels particularly significant, maybe
politically sensitive, just because of the sheer scale.
Speaker 2 (10:39):
Right absolutely, the job title heavy and tractor trailer truck drivers.
It's the single largest common job classification in the US
how people, about one point nine million workers. The scale
of the potential devastation here is just enormous. It hits
the American middle class right where it lives.
Speaker 1 (10:58):
So what are the projections, saying they're damning.
Speaker 2 (11:00):
McKenzie's analysis, which the sources reviewed, estimates that one point
four million of those truck drivers will be unemployed by
twenty thirty.
Speaker 1 (11:08):
One point four million. At one point nine million, Yeah.
Speaker 2 (11:11):
That's almost three quarters of the entire long haul workforce
wiped out in just seven years from now.
Speaker 1 (11:16):
Can we see this happening already? Is there real world evidence?
Speaker 2 (11:19):
Oh? Absolutely, it's not hypothetical. Companies like Aurora and Kodiak,
they've already logged over two point five million autonomous miles,
mostly on key logistics corridors, particularly in Texas, and.
Speaker 1 (11:30):
The big shipping companies UPS FedEx.
Speaker 2 (11:32):
They're rapidly adopting these systems. UPS specifically was cited in
the material for cutting ten thousand drivers in twenty twenty.
Speaker 1 (11:39):
Five, ten thousand drivers right.
Speaker 2 (11:40):
After deploying two simples driverless technology on the high volume
I ten corridor. That's the big one running across the
entire South Okay.
Speaker 1 (11:48):
So that's the long haul freight side. What about cities,
ride chairs deliveries, that's.
Speaker 2 (11:53):
The other shoe dropping. The sources point out the urban
dimension two every week, Waimo and CRU's Robotaxi fleets just
in San Francisco are now completing about one hundred and
fifty thousand fully driverless rides.
Speaker 1 (12:05):
One one hundred and fifty thousand rides a week live less.
Speaker 2 (12:10):
And every single one of those is a fair a
shift that a human uber or lift driver isn't getting.
The problem is, these autonomous systems aren't just niche fixes.
They're a universal threat across the whole transportation sector, delivery vans, buses, taxis,
tractor trailers, everything.
Speaker 1 (12:27):
But hang on, what about the classic argument that AI
creates new jobs. You know someone has to maintain these
complex autonomous trucks, right, AI repair technicians.
Speaker 2 (12:35):
The sources looked at that. They do acknowledge some new
rules are created, but the number is negligible compared to
the displacement. Think about it. An autonomous truck driven optimally
by AI probably needs less maintenance than a human driven one.
It drives more consistently, less wear and tear.
Speaker 1 (12:52):
Fewer sudden breaks, smoother acceleration exactly.
Speaker 2 (12:56):
So the sources estimate for every one hundred drivers displaced,
you might need only three new specialized sensor technicians or
maybe remote operators, three new jobs for every one hundred lost.
Speaker 1 (13:07):
The net result is still catastrophic.
Speaker 2 (13:10):
Job loss, overwhelmingly so.
Speaker 1 (13:11):
And this has ripple effects, doesn't it. It's not just
the driver who loses their job. There's an economic multiplier effect.
Speaker 2 (13:17):
Precisely. You have to think about the whole infrastructure that
supports one point nine million truck drivers stopping, resting, refueling every.
Speaker 1 (13:25):
Few hours truck stops.
Speaker 2 (13:27):
We're already seeing devastating consequences in those crucial truck stop towns.
The source material mentions Breezewood, Pennsylvania, classic example, right off
the turnpike. They're reporting a forty percent decline in business
revenue already forty percent?
Speaker 1 (13:39):
What does that mean locally?
Speaker 2 (13:40):
Think about the local economy just vanishing, the independent mechanics
who rely on human driven trucks, needing repairs, the massive
fuel sales, the taxes that support state roads, and the
local diners, the motels, the small retail shops build entirely
around serving those transient drivers.
Speaker 1 (13:59):
So it hits me canics, waitresses, gas station attendance.
Speaker 2 (14:02):
Entire communities face potential extinction when the trucks don't stop anymore.
Speaker 1 (14:06):
Okay, if the trucking industry is facing a kind of
physical collapse customer service seems like it's undergoing an instantaneous
auditory replacement.
Speaker 2 (14:15):
Yeah, that's a good way to put it. The scale
here is also huge. In twenty twenty three, about two
point eight million Americans worked in call.
Speaker 1 (14:21):
Centers two point eight million, and AI is.
Speaker 2 (14:23):
Impact by twenty twenty five. So basically now, AI voice
agents have already captured and estimated sixty percent of all
Tier one support functions.
Speaker 1 (14:30):
Percent of the basic inquiries, the simple.
Speaker 2 (14:32):
Stuff, exactly. And the corporate rush to automate is just overwhelming.
Why because the cost savings are immediate and they are massive.
Speaker 1 (14:40):
You mentioned Clarna earlier.
Speaker 2 (14:41):
Right, Their AI is handling two point three million chats
every single month. That's the equivalent output of about seven
hundred human.
Speaker 1 (14:49):
Agents, seven hundred salaries gone gone.
Speaker 2 (14:52):
And Airbnb they won't even further. They aggressively replaced twelve
hundred support staff how using eleven Labs hyperrealistic voice clones.
The tech is apparently so good, so realistic, that the
substitution is seamless. The average customer calling in probably can't
even tell they're talking to an AI.
Speaker 1 (15:11):
Wow, and it's not just simple chants or questions, is it?
Speaker 2 (15:14):
No? AT and T is using Google's Dialogue Flow that's
their AI platform, to resolve a staggering eighty percent of
complex billing disputes.
Speaker 1 (15:21):
Billing disputes those often need negotiation, maybe some.
Speaker 2 (15:24):
Empathy, apparently not anymore. Eighty percent resolved without any human
intervention whatsoever. It eliminates the need for human discretion and
tasks we used to think required it.
Speaker 1 (15:33):
But the sources they shift focus pretty dramatically when talking
about the global consequences here, right, it's not just us
call centers.
Speaker 2 (15:39):
No, and this is crucial. They highlight the global human cost,
specifically in the massive outsourcing hubs.
Speaker 1 (15:45):
Think about the Philippines, Yeah, a huge center for BPO
business process outsourcing.
Speaker 2 (15:50):
Exactly, home to one point two million BPO workers. These
are the outsourced global call center jobs. This industry is
a cornerstone of their national economy, supporting millions of Filipino families.
Speaker 1 (16:03):
So what happens when that economic base gets pulled out
from under them so quickly?
Speaker 2 (16:07):
The trauma is profound. The sources site WHO Data World
Health Organization reporting that suicide rates and former call center districts,
particularly in Manila, have spiked an astonishing three hundred percent
since twenty twenty three, three hundred percent since twenty twenty three.
This isn't just an economic data point anymore. It highlights
an immediate humanitarian crisis caused directly by the speed and
(16:29):
scale of this displacement.
Speaker 1 (16:30):
That's a visceral outcome. It really shows that yanking the
economic engine out of a region so suddenly it leads
to systemic societal failure, not just individual job losses.
Speaker 2 (16:41):
Yeah, and for a long time, you know, white collar
workers felt relatively safe. The assumption was AI would automate
the repetitive, maybe glue collar stuff, but not our jobs,
not the thinking jobs.
Speaker 1 (16:50):
They were profoundly wrong, weren't they.
Speaker 2 (16:52):
Completely The sources argue that AI's assault on creative and
professional industries is particularly well cruel because it targets what
they call the very essence of human identity, are unique
skills of creation, of abstract reasoning.
Speaker 1 (17:07):
So where's the evidence of this white collar devastation.
Speaker 2 (17:11):
It's becoming clear. Journalism, for instance, The New York Times
is already using AI to generate seventy percent of its
routine stock listings and company earnings recaps.
Speaker 1 (17:20):
The bread and butter financial reporting.
Speaker 2 (17:21):
Exactly functional, high volume writing that no longer requires a
human journalist. That single application eliminated one hundred and eighty jobs.
Speaker 1 (17:30):
There, okay, journalism, What about more purely creative fields animation?
Speaker 2 (17:34):
Animation took a huge hit. Disney reportedly laid off four
hundred specialized animators. Why because generative AI tools like Stable
Diffusion V three could cut the storyboard time for complex,
high budget projects from three weeks down to just three hours.
Speaker 1 (17:48):
Three weeks to three hours, the.
Speaker 2 (17:49):
Time to market advantage, the cost savings for shareholders. It's
almost impossible to ignore that kind of efficiency gain.
Speaker 1 (17:55):
And even for those creatives who managed to keep their jobs,
the economic ground is shifting under them, isn't it?
Speaker 2 (18:01):
It absolutely is. The economic leverage is just vanishing. The
sources note that forty two percent of professional copywriters lost
their jobs in twenty twenty four, forty two percent, and
the survivors, the few survivors, they're often reduced to becoming
AI prompt engineers basically people who supervise the models tweak
the inputs. But and this is key, they're earning on
(18:24):
average forty percent less. Why less because their core value
is no longer seen as their unique creative genius it's
shifted to curation workflow management. Their fundamental skill has been
devalued almost overnight.
Speaker 1 (18:37):
That erosion of the middle class firewall. That feels like
the most alarming long term trend here.
Speaker 2 (18:42):
It really is.
Speaker 1 (18:43):
Let's look at the data the sources provide comparing human
and AI accuracy in these highly skilled, traditionally safe fields.
The numbers are pretty staggering.
Speaker 2 (18:51):
Yeah, this data really destroys that old assumption that humans
are always necessary for quality control or nuanced judgment. Often
the machine is simply superior quantifiably.
Speaker 1 (19:00):
Okay, lay it out for us.
Speaker 2 (19:02):
All right, consider these stats pulled directly from the source
material radiology field requiring years of intense medical training. Right defo,
Google Health's AI achieves ninety four point three percent accuracy
and diagnosis the human baseline ninety one point two percent
higher accuracy from the AI significantly higher. That single performance
gap immediately puts an estimated one hundred and ten thousand
(19:25):
radiology jobs at risk globally.
Speaker 1 (19:27):
Wow. What about law legal research?
Speaker 2 (19:30):
Specialized legal lms like Harvey dot AI are boasting ninety
six percent accuracy and synthesizing case law and predicting outcomes
human paralegals and junior associates around eighty eight percent accuracy.
That potentially risks four hundred twenty thousand jobs in the
legal support sector.
Speaker 1 (19:46):
Four hundred and twenty thousand. And finance financial analysis.
Speaker 2 (19:50):
Bloomberg GPT train on decades of financial data achieves eighty
seven percent accuracy and forecasting market trends and generating complex reports.
The average human analyst around eighty two percent. That threatens
about two hundred and fifty thousand jobs.
Speaker 1 (20:03):
So the machine isn't just cheaper, it's often better at
the core tab That's the.
Speaker 2 (20:06):
Harsh reality, and that quantitative superiority leads directly to the
kind of corporate actions that you, the listener, really need
to understand are happening.
Speaker 1 (20:13):
Now, where do we see this most clearly?
Speaker 2 (20:15):
High finance is a great example. Investment banking. Goldensachs obviously
a powerhouse. They cut twelve hundred investment banking analysts in
twenty twenty.
Speaker 1 (20:24):
Five, twelve hundred analysts.
Speaker 2 (20:26):
Why after deploying their own proprietary large language model specifically
trained to generate pitch books.
Speaker 1 (20:33):
Pitch books that's like the core deliverable for junior analysts exactly.
Speaker 2 (20:38):
It used to be the cornerstone of starting a finance career.
Hours and hours of grinding out slides. Now the AI
does it.
Speaker 1 (20:46):
So what happened to the remaining analysts. Did they get promoted?
Speaker 2 (20:48):
No, they weren't promoted. They were effectively demoted. Their role
was reduced to being AI validators, meaning they just check
the model's output. They're policing the machine, not doing the
primary analysis, and their pay reportedly cut down to junior
associate levels. The entire value chain got flipped upside down.
The machine does the work, the human verifies. Okay, so
(21:09):
we've seen the mechanics the industries. We really need to
shift the focus now to the systemic failure that this
level of job destruction causes when millions of jobs just
vanish almost simultaneously. The economic fallout is sort of twofold,
isn't it. You get collapsing tax revenue on one side
and explosive inequality on the other.
Speaker 1 (21:28):
But first, the sources really emphasize we can't just talk statistics.
These numbers have to be made real exactly.
Speaker 2 (21:35):
This is where the abstract economics just crashes into lived reality.
These aren't just numbers on some nber spreadsheet. These are
people's livelihoods collapsing. Right now, the sources give us some
specific stories to focus on.
Speaker 1 (21:47):
Let's hear who are we talking about?
Speaker 2 (21:48):
Okay, First, there's Maria. She's a forty two year old
heavy truck driver down in Laredo, Texas. She was, you know,
the backbone of her family, paid off her house, driving
one hundred and ten thousand miles a year, solid middle
class job. Then Aurora's driverless trucks hit the I thirty
five corridor, her main route. Her dispacter basically told her, look,
(22:10):
we love you, Maria, but the AI doesn't take bathroom breaks,
it doesn't.
Speaker 1 (22:14):
Need sleep, just like that job gone.
Speaker 2 (22:16):
Job gone. Now she's reportedly on food stamps, stuck with
this expensive commercial driver's license that's becoming rapidly worthless.
Speaker 1 (22:23):
God, that's brutal. Who else?
Speaker 2 (22:25):
Then there's Rejesh. He's twenty nine, a call center team
lead over in Manila in the Philippines.
Speaker 1 (22:30):
Right the BPO sector we mentioned.
Speaker 2 (22:32):
Yeah, he was a dedicated middle manager doing well overseeing
a team of forty agents. His entire team, all forty
plus him, displaced basically overnight ow the US client they
served switched to an eleven labs AI one that spoke
perfect localized tobolog better than some of the agents. Probably
two thousand people in his immediate district lost their jobs
in that.
Speaker 1 (22:51):
Wave, two thousand. What's Rejesh doing now?
Speaker 2 (22:53):
Now? He applied to like three hundred jobs, but they
all demand AI literacy, and he asked, pretty poignantly in
the source material, how can I use AI literacy to
feed my kids tonight? The skills mismatch is immediate and devastating.
Speaker 1 (23:08):
Yeah, retraining doesn't help when you need income now. And
there was a white collar example too.
Speaker 2 (23:12):
Yes, Emily, she's thirty five, a copywriter living in Brooklyn.
Speaker 1 (23:15):
Good creative professional.
Speaker 2 (23:16):
Doing well, pulling in one hundred and twenty thousand dollars
a year at a pretty prestigious ad agency. Comfortable, upper
middle class life. Now her client base, they prefer the
twelve dollars output they can get from mid Journey for
images and Claude for text for their campaigns.
Speaker 1 (23:31):
Twelve dollars versus her.
Speaker 2 (23:32):
Salary exactly, So she's left scrambling. The sources say she's
ghostwriting LinkedIn posts for fifteen dollars an hour.
Speaker 1 (23:39):
From one hundred and twenty k to fifteen dollars.
Speaker 2 (23:41):
An hour, the value of her entire skill set built
over a decade destroyed.
Speaker 1 (23:45):
These stories, they really highlight the speed of the displacement,
don't they.
Speaker 2 (23:50):
That's the key, and it's why the government systems just
can't cope. The US Census Bureau reported, get this, two
point one million workers filed for unemployment, specifically citing AI
displacement as the reason. In twenty twenty five.
Speaker 1 (24:01):
Alone, two point one million sighting AI. How does that
compare to the year before.
Speaker 2 (24:05):
That figure is triple what it was in twenty twenty four. Triple.
The system is just being completely overwhelmed.
Speaker 1 (24:11):
Okay, so jobs disappear, this fast productivity technically soars, But
who captures the wealth? This is what the source is
called the one percent capture right, or the inequality bomb exactly.
Speaker 2 (24:23):
There was a critical paper cited from the NBER, the
National Bureau of Economic Research, published in twenty twenty five.
They meticulously tracked where the labor savings generated by AI
adoption actually.
Speaker 1 (24:35):
Went in the findings.
Speaker 2 (24:37):
The findings are stark. A staggering seventy eight percent of
all generated labor savings accrue directly to capital owners shareholders.
Speaker 1 (24:44):
Basically seventy eight percent, what about the workers?
Speaker 2 (24:46):
Another fifteen percent goes to the relatively small group of
high school workers needed to supervise and maintain these new
AI systems, the AI supervisors as they call them.
Speaker 1 (24:56):
Okay, seventy eight plus fifteen, that's ninety three percent. What's
left for the displaced.
Speaker 2 (24:59):
Worker a minuscule seven percent, just seven percent trickles down
to the millions losing their jobs, and that's usually just
through meager severance packages or frankly highly ineffective retraining programs.
Speaker 1 (25:10):
This concentration of the winnings, it feels unprecedented. We see
Nvidia's market cap soaring to three point four trillion dollars.
That's up twelve hundred percent since twenty twenty two.
Speaker 2 (25:20):
It's astronomical. And you also have figures like Sam Altman's
twenty twenty five compensation package cited in the Sources one
point two billion dollars in stock one point two billion. Well,
millions are losing jobs exactly.
Speaker 1 (25:32):
This explosive wealth creation for a tiny few is happening
at the exact same time as sharp economic decline for
the vast majority of blue collar and even mid skill
white collar workers.
Speaker 2 (25:42):
But isn't the counter argument always? While these founders, these investors,
they took massive risks to build this technology. Why are
the sources so focused on the wealth transfer aspect?
Speaker 1 (25:52):
Because the transfer isn't just about a few massive CEO packages,
it's systemic. The data clearly shows that median wages for
non college educated man who make up a huge chunk
of the displaced workforce in areas like trucking and manufacturing,
have fallen six point two percent in real term since
twenty twenty three. Adjusted for inflation. They are earning less.
Speaker 2 (26:11):
So the economic benefit is just incredibly polarized, which brings
us to the G and E coefficient, the measure of inequality. Right.
The Genie coefficient measures income inequality. Zero is perfect equality.
One is maximum inequality, where one person has everything. The
US figure hito point four to nine in twenty twenty five.
Speaker 1 (26:30):
Zero point four to nine. How does that compare historically?
Speaker 2 (26:33):
That's the highest it has been since nineteen twenty nine.
Speaker 1 (26:35):
Nineteen twenty nine, the year the crash.
Speaker 2 (26:37):
The sources draw a very deliberate parallel there. Nineteen twenty
nine led to the Great Depression, which eventually resulted in
the New Deal. Massive systemic changes to protect labor regulate capital.
Speaker 1 (26:48):
And the argument is AI is the modern accelerant driving
us back to that nineteen twenty nine level of inequality.
Speaker 2 (26:53):
Exactly, but crucially without the political mechanisms or maybe the
political will for a new Deal style response this time around.
Speaker 1 (27:01):
And we see companies making structural choices that reinforce this
hoarding of the gains, don't we. The Walmart example was striking.
Speaker 2 (27:07):
Yeah, Walmart's big AI deployment in their e commerce division
reportedly saved them two point one billion dollars annually in
labor costs. Just labor costs two.
Speaker 1 (27:15):
Point one billion. What could they have done with that?
Speaker 2 (27:17):
Well, the sources calculate that money could have paid forty
two thousand warehouse workers a decent fifty thousand dollars salary each,
forty two thousand jobs funded, But they did do that now. Instead,
the company used that labor saving to help fund a
massive twenty billion dollars stock buy back, returning the capital
directly to shareholders.
Speaker 1 (27:37):
The corporate incentives are just structurally geared toward capital now,
not labor, it.
Speaker 2 (27:42):
Seems that way, and this also creates another looming crisis,
a fiscal crisis for the government. Household displaced workers, well,
they aren't paying income tax, they aren't paying payroll taxes.
The irs, according to the sources, projects a potentially catastrophic
one point one trillion dollar federal tax revenue shortfall by
twenty thirty.
Speaker 1 (28:00):
One point one trillion dollars just from labor displacement and
wage stagnation.
Speaker 2 (28:04):
Purely from that effect. And think about states like Michigan
or Ohio where maybe twenty twenty five percent of the
workforce is concentrated in manufacturing and transportation. The two sector
is getting hit hardest and fastest for them. This tax
revenue collapse isn't just a problem, it's potentially existential.
Speaker 1 (28:19):
Okay, So if the problem is this systemic, this massive,
this immediate, what's the policy response been. This is where
the source material gets frankly, deeply disheartening. It paints a
picture of total failure. Governance just isn't keeping pace.
Speaker 2 (28:35):
Not even close.
Speaker 1 (28:36):
Let's look at the common solutions people propose and why,
the sources argue they've basically failed already.
Speaker 2 (28:41):
All Right, where do we start? Retraining? Is usually the
first thing people mention, right.
Speaker 1 (28:45):
Yeah, the classic answer to industrial change when one industry declines,
you retrain for the new one. For decades, the advice
was simple, learn to code.
Speaker 2 (28:54):
That mantra it's now completely obsolete. Why because, paradoxically, coding
is arguably one of the first high value cognitive jobs
that AI successfully ate.
Speaker 1 (29:04):
We saw the GitHub data earlier right.
Speaker 2 (29:06):
Githubs twenty twenty five report sixty one percent of all
new code is now AI generated. You can't effectively retrain
millions of people into a field that is itself rapidly
becoming automated. It makes no sense.
Speaker 1 (29:17):
Okay, so learn to code is out. What's the new advice?
Learn to manage AIS. Become a prompt engineer.
Speaker 2 (29:23):
That's the new buzzphrase. But the problem is scale. It
just doesn't scale to absorb eighty five million displaced workers globally,
or even the millions displaced in the US.
Speaker 1 (29:34):
How many prompt engineering jobs are there?
Speaker 2 (29:36):
Roughly the sources estimate only about one hundred and eighty
thousand specialized prompt engineering or AI oversight jobs worldwide. Compare
that to say, fourteen million displaced coders alone, let alone
truck drivers, call center agents, analysts.
Speaker 1 (29:51):
The math just doesn't work.
Speaker 2 (29:52):
It doesn't The source material really emphasizes the difference here.
Past disruptions, like say, the shift from farming to fact work,
created large new classes of jobs that could absorb millions.
This disruption, it seems to be eliminating entire classes of
jobs and replacing them with highly centralized, machine managed systems
that only need a tiny handful of human supervisors.
Speaker 1 (30:14):
Okay, so retraining looks like a dead end for the masses.
What about universal Basic income UBI? That's often floated is
the big solution for technological unemployment?
Speaker 2 (30:22):
It is, and it's politically attractive in some ways, but
the sources just point to the immediate financial reality. Yeah,
it doesn't work right now?
Speaker 1 (30:29):
Why not? The math?
Speaker 2 (30:30):
The math is brutal. Andrew Yang's proposal, which was kind
of the benchmark, one thousand dollars a month for every
American adult, would cost roughly three point two trillion dollars
a year just for the US.
Speaker 1 (30:41):
Three point two trillion. And how much tax revenue is
AI realistically generating that could fund this?
Speaker 2 (30:47):
Current projections cited put AI related tax revenue at only
about one hundred and eighty billion dollars annually. The gap
is just enormous.
Speaker 1 (30:56):
Right now, UBI is fifthally impossible without massive disruptive new.
Speaker 2 (31:00):
Taxes, taxes that simply haven't been implemented and frankly face
huge political opposition. Plus there's another wrinkle. The trials we've seen,
like the big one in Finland from twenty seventeen twenty eighteen,
what did that show? It showed that while UBI did
reduce stress and improve well being for the participants, which
is good, it also seemed to increase unemployment slightly because
people receiving the basic income stopped actively searching for work.
Speaker 1 (31:23):
So it potentially solves the poverty problem, but might actually
reinforce the displacement problem.
Speaker 2 (31:28):
That's the concern raise in the analysis. Yeah, it's complex.
Speaker 1 (31:31):
Okay, If retraining and UBI are problematic, what about regulation
trying to control the pace or impact of AI deployment.
Speaker 2 (31:40):
The regulatory landscape is well, shocking is one word for it.
Shockingly porous. Europe's big EUAI Act, passed in twenty twenty four,
got a lot of headlines.
Speaker 1 (31:51):
Yeah, it was a lotted as a global first, banning
high risk systems like social scoring, some facial recognition.
Speaker 2 (31:58):
Right, but there's a massive loophole, critical one, which is
the act specifically exempts what it calls productivity.
Speaker 1 (32:03):
Tools preativity tools.
Speaker 2 (32:05):
You mean the very large language models, the clauds, the
GPTs that are responsible for the mass job destruction we've
been talking about. Regulators focused on the scary or Wellian
big brother stuff, but they completely missed the immediate economic
threat decimating the workforce.
Speaker 1 (32:18):
So the result is zero enforcement against AI driven labor displacement.
Speaker 2 (32:23):
Pretty much zero effect on that front. And in the
United States, the source material notes quite simply there is
no comprehensive federal AI law whatsoever. Nothing. The world's largest
economy is flying completely blind.
Speaker 1 (32:35):
Into the storm facing this complete policy vacuum. Has there
been any resistance from workers?
Speaker 2 (32:41):
There has You're seeing the emergence of what some are
calling let out.
Speaker 1 (32:45):
Two point zero neo Luddites.
Speaker 2 (32:46):
Yeah. An anonymous group called the AI Resistance Network claimed
responsibility for coordinated arson attacks at three WAYMO depots in
twenty twenty five, trying to physically destroy the autonomous vehicle.
Speaker 1 (32:57):
Infrastructure and organized labor.
Speaker 2 (32:59):
We also so saw the Teamsters union organizing blockades at
Aurora's big trucking hub in Dallas. They managed to delay
about forty thousand shipments of crucial goods trying to exert pressure.
Speaker 1 (33:10):
How did the authorities respond to this resistance.
Speaker 2 (33:12):
Swiftly and unforgivingly. The FBI is now officially classifying AI
sabotage like the Waimo arson, as domestic terrorism. The state
is clearly protecting the technology deployment, not the displaced workers.
Speaker 1 (33:25):
And even when conventional policy proposals are made, like taxing
the tech, they.
Speaker 2 (33:29):
Get killed instantly. The sources highlight a proposal by respected
economists Darren Asimoglu and Simon Johnson. They suggested a modest
two percent tax just on AI capital expenditures, a basic
robot tax. What was the goal of the tex specifically
to fund transition stipends and maybe new vocational training for
workers displaced by AI? Seems reasonable, right?
Speaker 1 (33:51):
Yeah? What happened?
Speaker 2 (33:52):
Tech lobbyists reportedly killed that proposal in Congress within forty
eight hours of it.
Speaker 1 (33:56):
Being introduced forty eight hours.
Speaker 2 (33:58):
The political will to regulate this industry for the benefit
of labor, the sources conclude, it simply does not exist
at the moment.
Speaker 1 (34:06):
Okay, So, given this trajectory, the blinding speed of adoption,
that ninety nine point nine to eight percent cost collapse
in cognition and this almost complete policy failure. The sources
conclude by looking forward, and it's a necessary look, even
if it's pretty dark. They set twenty thirty as a
kind of critical inflection point, yeah, a crunch point, and
they outline three potential futures for humanity, three stark scenarios
(34:28):
based on the current path.
Speaker 2 (34:30):
Right, The first one, and sadly, the one they give
the highest probability seventy percent likelihood, is called the Great Hollowing.
Speaker 1 (34:36):
The Great Hollowing? What does that look like?
Speaker 2 (34:38):
In this scenario, AI continues its relentless march. It captures
maybe forty percent of global GDP by twenty thirty five,
simply because the cost of machine cognition is so incredibly low.
Speaker 1 (34:48):
Forty percent of GDP. What does that mean for jobs?
Speaker 2 (34:52):
It means an estimated one point two billion jobs vanish globally.
Speaker 1 (34:56):
One point two billion.
Speaker 2 (34:57):
That's the projection mass unemployment, wage collapse. It triggers such
severe social unrest that governments are forced to impose immediate
capital controls, trying desperately to stop the hoarding of wealth
by the AI owners.
Speaker 1 (35:11):
These the capital controls work, not.
Speaker 2 (35:13):
Really, according to the scenario, It just causes the AI
firms which are highly mobile to instantly relocate to more
permissive regulation free zones. Thinks Singapore, maybe the UAE, which
further cripples Western economies that try to regulate.
Speaker 1 (35:27):
And the end result for society.
Speaker 2 (35:29):
Inequality reaches levels that the sources. Compared to feudalism, you
have a vast majority subsisting on meager government stipends or
charity provided by the tiny elite who own the AI.
The means of production.
Speaker 1 (35:40):
That's bleak okay, seventy percent chands with great hauling with
scenario two.
Speaker 2 (35:44):
Scenario two is called the neo ludite backlash. This one
is given a twenty percent likely NEOLAA.
Speaker 1 (35:48):
Backlash, so the resistance succeeds sort of.
Speaker 2 (35:51):
In this scenario, public fury, fueled by the human toll
the stories of Maria jesh Emily multiplied by millions, becomes overwhelming.
It forces governments around twenty thirty perhaps to enact a
global moratorium on further development and deployment of generative AI.
Speaker 1 (36:08):
They hit the brakes. Stop the job losses.
Speaker 2 (36:11):
Sounds better, It stops the immediate bleeding, yes all, but
the consequence, according to this scenario is severe global stagnation.
Innovation just stalls completely in the West. Why, Because AI
is becoming fundamental to scientific discovery. Are in d efficiency?
You turn it off, you slow everything down. Meanwhile, nations
like China maybe Russia, who likely ignore the ban and
(36:32):
continue aggressive AI.
Speaker 1 (36:34):
Development, they leap frog ahead.
Speaker 2 (36:35):
They leap frog ahead, They dominate global technology and military development,
economic growth. The West, having banned the key driver of progress,
enters a projected fifty year period of relative economic and
geopolitical decline. It trades the short term pain of displacement
for long term stagnation.
Speaker 1 (36:52):
So you either get feudalism or stagnation. Is there any
optimistic outcome?
Speaker 2 (36:55):
There is a third scenario, but the sources give it
only a ten percent likelihood. It's called the abundance pivot.
Speaker 1 (37:01):
The abundance pivot. Okay, what happens here?
Speaker 2 (37:03):
This is the future where we actually manage to harness
the incredible productivity gains from AI for broad societal benefit.
Productivity becomes so massive that it eventually can fund things
like a universal twenty hour work week, maybe by twenty forty.
Speaker 1 (37:18):
And UBI becomes viable.
Speaker 2 (37:19):
UBI finally becomes fiscally viable in this scenario. Why because
a robust global system of robot taxes or AI capital
taxes is successfully implemented, generating an estimated fifteen trillion dollars annually,
enough to cover the cost we talked about earlier.
Speaker 1 (37:35):
So humanity reaches a kind of post scarcity.
Speaker 2 (37:37):
Era potentially yes, a world of abundance driven by machine intelligence.
But and this is the absolutely critical caveat in the
source material, humanity only reaches the state after surviving the
twenty thirties bloodbath O, meaning we likely still have to
go through something resembling the Great Hollowing first, the mass displacement,
the unrest, the inequality spike, before society finds the political
(37:58):
will and the mechanisms to pivot towards abundance. The journey
to utopia is littered with massive societal risk and suffering
hashtag tag outro Wow.
Speaker 1 (38:09):
This has been a deeply necessary, I think, but also
a very challenging deep dive into this source material.
Speaker 2 (38:14):
It's heavy stuff.
Speaker 1 (38:15):
It really is. The central and I think undeniable argument
we've pulled from every paper, every forecast here is that
the mass destruction of millions of jobs by AI, it's
not some hypothetical risk for our kids to worry about.
It's a present catastrophe. It's defining the economic reality of
the twenty twenties right now.
Speaker 2 (38:32):
Yeah, we've mapped the sheer speed of it, that unbelievable
ninety nine point nine eight percent cost collapse in cognition.
We've mapped the winners, haven't we The soaring market caps
of Nvidia, the one point two billion dollar stock packages
for CEOs like Altman. That's seventy eight percent of wealth
flowing straight to capital.
Speaker 1 (38:46):
And we've tried to put human faces on the losers.
Maria the truck driver now on food stamps. Rajesh, the
call center lead in Manila who can't find work, Emily
the skilled copywriter, now ghostwriting for fifteen bucks an hour.
Speaker 2 (39:01):
It's brutal, and the clock is really ticking, isn't it.
The projections say by twenty thirty, just a few years away,
another three hundred million jobs globally are likely to vanish.
Speaker 1 (39:11):
Three hundred million more. It feels like a silent coup,
the greatest and certainly the swiftest transfer of wealth and
power in modern human history, happening right under our noses and.
Speaker 2 (39:22):
The sources they bring it back to that historical context
near the end the First Industrial Revolution, the original Luddites, right,
they protested the machines taking their weaving jobs, and yes,
they were crushed by the state.
Speaker 1 (39:34):
They lost.
Speaker 2 (39:34):
They lost, but their children or certainly their grandchildren, eventually
inherited the factories. They became the modern working class, and
ultimately they did benefit from the massive new productive capacity
the machines created through unions, regulations, the welfare state.
Speaker 1 (39:49):
Okay, so what's the final thought here? The takeaway?
Speaker 2 (39:52):
This leads us to the final lingering question the source
material poses for you, the listener, to really grapple with
the question is whether today's displaced workers, the Mariahs, the Rajeshas,
the Emily's, will inherit anything at all this time, or
whether the masters of AI, the owners of the algorithms,
will simply hoard the entire future, leaving the displaced with
(40:13):
neither jobs nor any share in the revolutionary abundance the
technology clearly promises.
Speaker 1 (40:18):
Will they inherit anything or will the future be hoarded?
That's a chilling question to end on. Thank you for
walking us through that.
Welcome back to the Deep Dive. We are here to
take the stack of source material you shared with us,
the research papers, the internal memos, the corporate projections, and
will really extract the critical knowledge you need. Today. We
aren't discussing potential future disruptions. We are diving headfirst into
what these sources characterize as an accelerating economic crisis.
Speaker 2 (00:22):
Yeah, and the premise here is pretty aggressive, almost terrifying,
you could.
Speaker 1 (00:26):
Say, it really is. It demands immediate attention. The core
argument artificial intelligence isn't just augmenting human labor. The sources say,
it's obliterating.
Speaker 2 (00:36):
It obliterating, that's the word they use, and on a scale,
and importantly, at a speed we just haven't seen since
what the Industrial.
Speaker 1 (00:43):
Revolution exactly two centuries ago.
Speaker 2 (00:45):
And that speed, you know, that's the critical difference. It's
what makes us Deep Dive so urgent for you, the listener.
I mean, think about it. The initial warnings just a
few years back, they projected maybe seventy five million jobs
displaced globally by AI.
Speaker 1 (00:58):
Which sounded like a lot at the time.
Speaker 2 (01:00):
It did, but the sources point to this grim acceleration.
The updated twenty twenty five World Economic Forum forecast, it's
drastically bleaker. We're now talking eighty five million jobs globally
displaced by twenty twenty.
Speaker 1 (01:15):
Seven, eighty five million by twenty twenty seven. That's a
massive jump in projections over a really condensed timeline.
Speaker 2 (01:22):
It is it suggests the catastrophe, or well the shift
is unfolding much faster than anyone predicted.
Speaker 1 (01:28):
And what's crucial here, and the sources really hammered this home,
is that this isn't just theory anymore. It's not some
future fear. They argue, quote, the apocalypse is not coming.
Speaker 2 (01:38):
It has arrived, right, So we need to ground this,
make it real. What are the specific current corporate actions
that show this is already happening? Where's the reality check?
Speaker 1 (01:46):
Okay, let's look at some big names.
Speaker 2 (01:48):
We can start with the giants. Yeah, look at IBM.
Back in twenty twenty three, they announced they were freezing
hiring for about seven eight hundred back office.
Speaker 1 (01:55):
Roles, right, the administrative stuff.
Speaker 2 (01:57):
Exactly, and they specifically cited AI's capacity to handle those
routine tasks. That's not waiting for displacement. That's a preventative
strike against labor.
Speaker 1 (02:04):
Okay, IBM freezing roles. What else.
Speaker 2 (02:06):
Even more dramatically, you have Klarna, the huge Swedish fintech company.
They used AI to actively shrink their existing workforce by
fifty percent. Fifty percent, how after successfully deploying AI for
customer service interactions. And this wasn't, you know, just letting
people leave naturally, this was systemic replacement.
Speaker 1 (02:28):
That distinction feels really important workforce shrinkage versus just not hiring.
Speaker 2 (02:33):
It's vital. It means the company is actively taking salaries
benefits off the balance sheet entirely. It's a direct substitution.
Speaker 1 (02:41):
So where's the most compelling evidence that AI is really
out of the lab and hitting the mainstream.
Speaker 2 (02:46):
Well, it's interesting. You see it in two sort of
very different sectors simultaneously, transportation and high end diagnostics.
Speaker 1 (02:53):
Okay, you claim.
Speaker 2 (02:53):
So transportation, we now have autonomous Waimo fleets. They're logging
millions of driverless miles in places like Phoenix, San Francisco,
dense urban area, deplacing.
Speaker 1 (03:02):
Ride share drivers, delivery drivers.
Speaker 2 (03:04):
Thousands of them daily. And at the same time, you
look at the medical field, a really high skill domain,
right radiology.
Speaker 1 (03:09):
Yeah, it requires years of training.
Speaker 2 (03:11):
Google's DeepMind AI now routinely surpasses the diagnostic accuracy of
human radiologists.
Speaker 1 (03:17):
Wait, surpasses, not just matches surpasses.
Speaker 2 (03:20):
So when the machine is cheaper, faster, and more accurate
than a highly paid specialist, will the economic floor for
that job just disappears instantly?
Speaker 1 (03:30):
Okay, So our mission today for you listening is to
give you that necessary shortcut through all this evidence. We're
going to dissect the mechanics. First, how AI actually destroys
this value, focusing on this concept of cheaper cognition. Then
we'll try to quantify the human toll. Look at specific
industries that are already, as the sources say, in freefall, trucking,
(03:52):
customer service, even white collar jobs.
Speaker 2 (03:54):
Yeah, the place is feeling the immediate impact.
Speaker 1 (03:56):
And finally we have to expose the policy vacuum, the
frame total failure of governance to keep up with this
accelerating shift.
Speaker 2 (04:03):
Which you know might be the most dangerous part of
all this, the lack of response.
Speaker 1 (04:06):
Okay, let's unpack this core mechanism. Then it's one thing
to say AI is smart, right, but we need to
understand how it delayers this well economic annihilation. The sources
keep coming back to this phrase cheaper cognition.
Speaker 2 (04:19):
Right, if you think about the last say fifty years
of the digital age, it was all defined by Moore's.
Speaker 1 (04:24):
Law, cheaper transistors, more computing power.
Speaker 2 (04:27):
Exactly, cheaper processing. But the next decade, maybe longer, it's
defined by AI providing cheaper, faster cognition.
Speaker 1 (04:34):
Cognition like thinking.
Speaker 2 (04:36):
Thinking, writing, analyzing, diagnosing, translating the fundamental output of millions
of white collar jobs, specialized jobs. So when that fundamental
output becomes nearly free, I mean, what happens, the economic
justification for paying millions of salaried workers just evaporates.
Speaker 1 (04:53):
Why pay a specialist analyst one hundred thousand dollars a year.
Speaker 2 (04:56):
When in LLM, a large language model can perform the
synth sist of maybe one thousand analysts simultaneously and it
works two hundred and forty seven, No benefits, no salary,
doesn't need a desk, and the source of stress.
Speaker 1 (05:08):
This isn't just a marginal improvement. It's not like getting
ten percent cheaper.
Speaker 2 (05:11):
No. No, it's a structural collapse in costs. That's why
the speed is so terrified.
Speaker 1 (05:14):
And this is where it gets really interesting. We have
a data point that just crystallizes the economic difference. I
want you, the listener, to really pay attention to this
cost collapse for routine language processing. Get rewind just a
bit twenty nineteen. If you wanted, say, one thousand words
of professionally drafted, human quality text generated via an API.
Speaker 2 (05:36):
Like for marketing copy or a basic.
Speaker 1 (05:38):
Report exactly, the cost was roughly sixty cents zero dollars.
Sixty cents.
Speaker 2 (05:43):
Sixty cents seems reasonable.
Speaker 1 (05:44):
Now, fast forward to twenty twenty five. Open AI's latest
model OH one preview. It delivers equivalent often actually superior
output for wait for it, zero point zero zero zero
one two dollars from sixty cents.
Speaker 2 (05:57):
That's a ninety nine point nine to eight percent price drop.
Speaker 1 (06:00):
Six years, six years. That's staggering. That's not evolution, that's
instantaneous collapse, right, I mean, think about previous disruptions. Getting
in ninety nine percent costroduction took decades, like manual typesetting
to digital printing back.
Speaker 2 (06:12):
This speed is the thing, the novel destructive element. Absolutely so,
if your whole livelihood depends on routine language generation, drafting
legal summaries, synthesizing earnings reports, writing ad copy, your economic
floor just vanished. Almost one hundred percent drop, and it
happened while you were like still negotiating your annual rais.
Speaker 1 (06:32):
Now what's truly frightening, as the sources make clear, is
that this pattern, this radical cost collapse, it's not just text.
It's universal.
Speaker 2 (06:40):
Yeah, it's happening across basically every cognitive domain. At the
same time, the economic incentive for capital earners, for businesses
to switch is just overwhelming.
Speaker 1 (06:49):
Now give us some examples beyond text.
Speaker 2 (06:51):
Okay, take image generation. A human illustrator might charge what
four hundred dollars or more for a commercial grade image
plus licensing fees revisions standard practice mid journey V six.
It produces equivalent, sometimes even more creative, high resolution output
for about zero four cents per image four cents four
cents versus four hundred dollars.
Speaker 1 (07:11):
So the decision to hire a human it's not just
about saving a bit of money anymore. It becomes frankly
an economic absurdity for many routine tasks.
Speaker 2 (07:19):
Okay, images, what else coding? Coding? Yeah, gethub copilot. It's
now reportedly responsible for writing forty six percent of developer's code,
almost half almost half. It cuts the intellectual heavy lifting
for basic programming dramatically, and this directly led firms like
Accenture to report stunningly a thirty percent cut in demand
(07:40):
for junior programmers.
Speaker 1 (07:41):
So the entry level cognitive jobs are getting hit first.
Speaker 2 (07:44):
The first casualties. Absolutely, And if your job involves a
sophisticated data analysis, spreadsheets, modeling, andthropics, claude three point five
sonnet model performs the same deep multi variable analysis four
hundred times faster than a high highly skilled human analyst
four hundred times.
Speaker 1 (08:03):
The human just can't compete, not on speed, not on cost,
not on scale. So this efficiency isn't just a threat,
it's becoming an economic mandate for any business that wants
to stay competitive very much, which leads us directly to
what the source is called the substitution flywheel. This explains
the mechanism for mass job loss, why it scales so
incredibly quickly. It's a critical three step process, right.
Speaker 2 (08:23):
The flywheel starts with the first step, initial substitution. This
is the bit we see most clearly.
Speaker 1 (08:29):
One AI system comes in and.
Speaker 2 (08:30):
Immediately replaces let's say ten human jobs, a team, a department.
Speaker 1 (08:35):
Maybe okay, step one, but the.
Speaker 2 (08:37):
Real danger, the reason it scales is the second phase,
scaling via cloud deployment. Because AI systems are fundamentally software.
Speaker 1 (08:45):
You can copy them infinitely deploy them anywhere exactly.
Speaker 2 (08:47):
That one AI system that replaced ten people, it can
be instantly distributed globally through the cloud, instantly replacing ten
thousand human jobs, often without moving any physical infrastructure or
needing any local train. It's centralized and incredibly.
Speaker 1 (09:02):
Swift ten thousand from one system. And then step three compounding, ah.
Speaker 2 (09:07):
The compounding phase. This is the part that means the
human worker, well, they can't realistically catch up even with retraining,
because the AI system improves itself through reinforcement learning feedback
loops gets better every single day, widening the performance gap
between the machine and the human.
Speaker 1 (09:24):
While the human needs years to learn a new skill.
Speaker 2 (09:27):
The AI needs milliseconds to update its model based on
new data. It's a constantly accelerating gap.
Speaker 1 (09:33):
Do we have a specific example of this flywheel in action?
Something concrete?
Speaker 2 (09:38):
We do, and it's a chilling one. The sources pulled
it from an internal document leak. Apparently Amazon they're fulfillment centers.
Speaker 1 (09:45):
Okay, Amazon warehouses.
Speaker 2 (09:46):
A single AI system was rolled out its job optimize
logistics and crucially labor scheduling, and the result the elimination
of a staggering one hundred and thirty thousand seasonal warehouse.
Speaker 1 (09:59):
Jobs one hundred and thirty thousand in the.
Speaker 2 (10:01):
Fourth quarter of twenty twenty four alone, just like that,
mass obsolescence in action, one centralized diploma decision, wiping out
thousands upon thousands of jobs instantly right before the peak
holiday season. Two.
Speaker 1 (10:13):
Wow, that illustrates the scale perfectly. Okay, so we've established
the economic mechanism, this relentless drive for cheaper cognition amplified
by this substitution flywheel. Now let's pivot. Let's move from
the abstract cost collapse to the specific tangible industries that
the sources describe as being already in well free fall.
(10:34):
Let's start with transportation. This one feels particularly significant, maybe
politically sensitive, just because of the sheer scale.
Speaker 2 (10:39):
Right absolutely, the job title heavy and tractor trailer truck drivers.
It's the single largest common job classification in the US
how people, about one point nine million workers. The scale
of the potential devastation here is just enormous. It hits
the American middle class right where it lives.
Speaker 1 (10:58):
So what are the projections, saying they're damning.
Speaker 2 (11:00):
McKenzie's analysis, which the sources reviewed, estimates that one point
four million of those truck drivers will be unemployed by
twenty thirty.
Speaker 1 (11:08):
One point four million. At one point nine million, Yeah.
Speaker 2 (11:11):
That's almost three quarters of the entire long haul workforce
wiped out in just seven years from now.
Speaker 1 (11:16):
Can we see this happening already? Is there real world evidence?
Speaker 2 (11:19):
Oh? Absolutely, it's not hypothetical. Companies like Aurora and Kodiak,
they've already logged over two point five million autonomous miles,
mostly on key logistics corridors, particularly in Texas, and.
Speaker 1 (11:30):
The big shipping companies UPS FedEx.
Speaker 2 (11:32):
They're rapidly adopting these systems. UPS specifically was cited in
the material for cutting ten thousand drivers in twenty twenty.
Speaker 1 (11:39):
Five, ten thousand drivers right.
Speaker 2 (11:40):
After deploying two simples driverless technology on the high volume
I ten corridor. That's the big one running across the
entire South Okay.
Speaker 1 (11:48):
So that's the long haul freight side. What about cities,
ride chairs deliveries, that's.
Speaker 2 (11:53):
The other shoe dropping. The sources point out the urban
dimension two every week, Waimo and CRU's Robotaxi fleets just
in San Francisco are now completing about one hundred and
fifty thousand fully driverless rides.
Speaker 1 (12:05):
One one hundred and fifty thousand rides a week live less.
Speaker 2 (12:10):
And every single one of those is a fair a
shift that a human uber or lift driver isn't getting.
The problem is, these autonomous systems aren't just niche fixes.
They're a universal threat across the whole transportation sector, delivery vans, buses, taxis,
tractor trailers, everything.
Speaker 1 (12:27):
But hang on, what about the classic argument that AI
creates new jobs. You know someone has to maintain these
complex autonomous trucks, right, AI repair technicians.
Speaker 2 (12:35):
The sources looked at that. They do acknowledge some new
rules are created, but the number is negligible compared to
the displacement. Think about it. An autonomous truck driven optimally
by AI probably needs less maintenance than a human driven one.
It drives more consistently, less wear and tear.
Speaker 1 (12:52):
Fewer sudden breaks, smoother acceleration exactly.
Speaker 2 (12:56):
So the sources estimate for every one hundred drivers displaced,
you might need only three new specialized sensor technicians or
maybe remote operators, three new jobs for every one hundred lost.
Speaker 1 (13:07):
The net result is still catastrophic.
Speaker 2 (13:10):
Job loss, overwhelmingly so.
Speaker 1 (13:11):
And this has ripple effects, doesn't it. It's not just
the driver who loses their job. There's an economic multiplier effect.
Speaker 2 (13:17):
Precisely. You have to think about the whole infrastructure that
supports one point nine million truck drivers stopping, resting, refueling every.
Speaker 1 (13:25):
Few hours truck stops.
Speaker 2 (13:27):
We're already seeing devastating consequences in those crucial truck stop towns.
The source material mentions Breezewood, Pennsylvania, classic example, right off
the turnpike. They're reporting a forty percent decline in business
revenue already forty percent?
Speaker 1 (13:39):
What does that mean locally?
Speaker 2 (13:40):
Think about the local economy just vanishing, the independent mechanics
who rely on human driven trucks, needing repairs, the massive
fuel sales, the taxes that support state roads, and the
local diners, the motels, the small retail shops build entirely
around serving those transient drivers.
Speaker 1 (13:59):
So it hits me canics, waitresses, gas station attendance.
Speaker 2 (14:02):
Entire communities face potential extinction when the trucks don't stop anymore.
Speaker 1 (14:06):
Okay, if the trucking industry is facing a kind of
physical collapse customer service seems like it's undergoing an instantaneous
auditory replacement.
Speaker 2 (14:15):
Yeah, that's a good way to put it. The scale
here is also huge. In twenty twenty three, about two
point eight million Americans worked in call.
Speaker 1 (14:21):
Centers two point eight million, and AI is.
Speaker 2 (14:23):
Impact by twenty twenty five. So basically now, AI voice
agents have already captured and estimated sixty percent of all
Tier one support functions.
Speaker 1 (14:30):
Percent of the basic inquiries, the simple.
Speaker 2 (14:32):
Stuff, exactly. And the corporate rush to automate is just overwhelming.
Why because the cost savings are immediate and they are massive.
Speaker 1 (14:40):
You mentioned Clarna earlier.
Speaker 2 (14:41):
Right, Their AI is handling two point three million chats
every single month. That's the equivalent output of about seven
hundred human.
Speaker 1 (14:49):
Agents, seven hundred salaries gone gone.
Speaker 2 (14:52):
And Airbnb they won't even further. They aggressively replaced twelve
hundred support staff how using eleven Labs hyperrealistic voice clones.
The tech is apparently so good, so realistic, that the
substitution is seamless. The average customer calling in probably can't
even tell they're talking to an AI.
Speaker 1 (15:11):
Wow, and it's not just simple chants or questions, is it?
Speaker 2 (15:14):
No? AT and T is using Google's Dialogue Flow that's
their AI platform, to resolve a staggering eighty percent of
complex billing disputes.
Speaker 1 (15:21):
Billing disputes those often need negotiation, maybe some.
Speaker 2 (15:24):
Empathy, apparently not anymore. Eighty percent resolved without any human
intervention whatsoever. It eliminates the need for human discretion and
tasks we used to think required it.
Speaker 1 (15:33):
But the sources they shift focus pretty dramatically when talking
about the global consequences here, right, it's not just us
call centers.
Speaker 2 (15:39):
No, and this is crucial. They highlight the global human cost,
specifically in the massive outsourcing hubs.
Speaker 1 (15:45):
Think about the Philippines, Yeah, a huge center for BPO
business process outsourcing.
Speaker 2 (15:50):
Exactly, home to one point two million BPO workers. These
are the outsourced global call center jobs. This industry is
a cornerstone of their national economy, supporting millions of Filipino families.
Speaker 1 (16:03):
So what happens when that economic base gets pulled out
from under them so quickly?
Speaker 2 (16:07):
The trauma is profound. The sources site WHO Data World
Health Organization reporting that suicide rates and former call center districts,
particularly in Manila, have spiked an astonishing three hundred percent
since twenty twenty three, three hundred percent since twenty twenty three.
This isn't just an economic data point anymore. It highlights
an immediate humanitarian crisis caused directly by the speed and
(16:29):
scale of this displacement.
Speaker 1 (16:30):
That's a visceral outcome. It really shows that yanking the
economic engine out of a region so suddenly it leads
to systemic societal failure, not just individual job losses.
Speaker 2 (16:41):
Yeah, and for a long time, you know, white collar
workers felt relatively safe. The assumption was AI would automate
the repetitive, maybe glue collar stuff, but not our jobs,
not the thinking jobs.
Speaker 1 (16:50):
They were profoundly wrong, weren't they.
Speaker 2 (16:52):
Completely The sources argue that AI's assault on creative and
professional industries is particularly well cruel because it targets what
they call the very essence of human identity, are unique
skills of creation, of abstract reasoning.
Speaker 1 (17:07):
So where's the evidence of this white collar devastation.
Speaker 2 (17:11):
It's becoming clear. Journalism, for instance, The New York Times
is already using AI to generate seventy percent of its
routine stock listings and company earnings recaps.
Speaker 1 (17:20):
The bread and butter financial reporting.
Speaker 2 (17:21):
Exactly functional, high volume writing that no longer requires a
human journalist. That single application eliminated one hundred and eighty jobs.
Speaker 1 (17:30):
There, okay, journalism, What about more purely creative fields animation?
Speaker 2 (17:34):
Animation took a huge hit. Disney reportedly laid off four
hundred specialized animators. Why because generative AI tools like Stable
Diffusion V three could cut the storyboard time for complex,
high budget projects from three weeks down to just three hours.
Speaker 1 (17:48):
Three weeks to three hours, the.
Speaker 2 (17:49):
Time to market advantage, the cost savings for shareholders. It's
almost impossible to ignore that kind of efficiency gain.
Speaker 1 (17:55):
And even for those creatives who managed to keep their jobs,
the economic ground is shifting under them, isn't it?
Speaker 2 (18:01):
It absolutely is. The economic leverage is just vanishing. The
sources note that forty two percent of professional copywriters lost
their jobs in twenty twenty four, forty two percent, and
the survivors, the few survivors, they're often reduced to becoming
AI prompt engineers basically people who supervise the models tweak
the inputs. But and this is key, they're earning on
(18:24):
average forty percent less. Why less because their core value
is no longer seen as their unique creative genius it's
shifted to curation workflow management. Their fundamental skill has been
devalued almost overnight.
Speaker 1 (18:37):
That erosion of the middle class firewall. That feels like
the most alarming long term trend here.
Speaker 2 (18:42):
It really is.
Speaker 1 (18:43):
Let's look at the data the sources provide comparing human
and AI accuracy in these highly skilled, traditionally safe fields.
The numbers are pretty staggering.
Speaker 2 (18:51):
Yeah, this data really destroys that old assumption that humans
are always necessary for quality control or nuanced judgment. Often
the machine is simply superior quantifiably.
Speaker 1 (19:00):
Okay, lay it out for us.
Speaker 2 (19:02):
All right, consider these stats pulled directly from the source
material radiology field requiring years of intense medical training. Right defo,
Google Health's AI achieves ninety four point three percent accuracy
and diagnosis the human baseline ninety one point two percent
higher accuracy from the AI significantly higher. That single performance
gap immediately puts an estimated one hundred and ten thousand
(19:25):
radiology jobs at risk globally.
Speaker 1 (19:27):
Wow. What about law legal research?
Speaker 2 (19:30):
Specialized legal lms like Harvey dot AI are boasting ninety
six percent accuracy and synthesizing case law and predicting outcomes
human paralegals and junior associates around eighty eight percent accuracy.
That potentially risks four hundred twenty thousand jobs in the
legal support sector.
Speaker 1 (19:46):
Four hundred and twenty thousand. And finance financial analysis.
Speaker 2 (19:50):
Bloomberg GPT train on decades of financial data achieves eighty
seven percent accuracy and forecasting market trends and generating complex reports.
The average human analyst around eighty two percent. That threatens
about two hundred and fifty thousand jobs.
Speaker 1 (20:03):
So the machine isn't just cheaper, it's often better at
the core tab That's the.
Speaker 2 (20:06):
Harsh reality, and that quantitative superiority leads directly to the
kind of corporate actions that you, the listener, really need
to understand are happening.
Speaker 1 (20:13):
Now, where do we see this most clearly?
Speaker 2 (20:15):
High finance is a great example. Investment banking. Goldensachs obviously
a powerhouse. They cut twelve hundred investment banking analysts in
twenty twenty.
Speaker 1 (20:24):
Five, twelve hundred analysts.
Speaker 2 (20:26):
Why after deploying their own proprietary large language model specifically
trained to generate pitch books.
Speaker 1 (20:33):
Pitch books that's like the core deliverable for junior analysts exactly.
Speaker 2 (20:38):
It used to be the cornerstone of starting a finance career.
Hours and hours of grinding out slides. Now the AI
does it.
Speaker 1 (20:46):
So what happened to the remaining analysts. Did they get promoted?
Speaker 2 (20:48):
No, they weren't promoted. They were effectively demoted. Their role
was reduced to being AI validators, meaning they just check
the model's output. They're policing the machine, not doing the
primary analysis, and their pay reportedly cut down to junior
associate levels. The entire value chain got flipped upside down.
The machine does the work, the human verifies. Okay, so
(21:09):
we've seen the mechanics the industries. We really need to
shift the focus now to the systemic failure that this
level of job destruction causes when millions of jobs just
vanish almost simultaneously. The economic fallout is sort of twofold,
isn't it. You get collapsing tax revenue on one side
and explosive inequality on the other.
Speaker 1 (21:28):
But first, the sources really emphasize we can't just talk statistics.
These numbers have to be made real exactly.
Speaker 2 (21:35):
This is where the abstract economics just crashes into lived reality.
These aren't just numbers on some nber spreadsheet. These are
people's livelihoods collapsing. Right now, the sources give us some
specific stories to focus on.
Speaker 1 (21:47):
Let's hear who are we talking about?
Speaker 2 (21:48):
Okay, First, there's Maria. She's a forty two year old
heavy truck driver down in Laredo, Texas. She was, you know,
the backbone of her family, paid off her house, driving
one hundred and ten thousand miles a year, solid middle
class job. Then Aurora's driverless trucks hit the I thirty
five corridor, her main route. Her dispacter basically told her, look,
(22:10):
we love you, Maria, but the AI doesn't take bathroom breaks,
it doesn't.
Speaker 1 (22:14):
Need sleep, just like that job gone.
Speaker 2 (22:16):
Job gone. Now she's reportedly on food stamps, stuck with
this expensive commercial driver's license that's becoming rapidly worthless.
Speaker 1 (22:23):
God, that's brutal. Who else?
Speaker 2 (22:25):
Then there's Rejesh. He's twenty nine, a call center team
lead over in Manila in the Philippines.
Speaker 1 (22:30):
Right the BPO sector we mentioned.
Speaker 2 (22:32):
Yeah, he was a dedicated middle manager doing well overseeing
a team of forty agents. His entire team, all forty
plus him, displaced basically overnight ow the US client they
served switched to an eleven labs AI one that spoke
perfect localized tobolog better than some of the agents. Probably
two thousand people in his immediate district lost their jobs
in that.
Speaker 1 (22:51):
Wave, two thousand. What's Rejesh doing now?
Speaker 2 (22:53):
Now? He applied to like three hundred jobs, but they
all demand AI literacy, and he asked, pretty poignantly in
the source material, how can I use AI literacy to
feed my kids tonight? The skills mismatch is immediate and devastating.
Speaker 1 (23:08):
Yeah, retraining doesn't help when you need income now. And
there was a white collar example too.
Speaker 2 (23:12):
Yes, Emily, she's thirty five, a copywriter living in Brooklyn.
Speaker 1 (23:15):
Good creative professional.
Speaker 2 (23:16):
Doing well, pulling in one hundred and twenty thousand dollars
a year at a pretty prestigious ad agency. Comfortable, upper
middle class life. Now her client base, they prefer the
twelve dollars output they can get from mid Journey for
images and Claude for text for their campaigns.
Speaker 1 (23:31):
Twelve dollars versus her.
Speaker 2 (23:32):
Salary exactly, So she's left scrambling. The sources say she's
ghostwriting LinkedIn posts for fifteen dollars an hour.
Speaker 1 (23:39):
From one hundred and twenty k to fifteen dollars.
Speaker 2 (23:41):
An hour, the value of her entire skill set built
over a decade destroyed.
Speaker 1 (23:45):
These stories, they really highlight the speed of the displacement,
don't they.
Speaker 2 (23:50):
That's the key, and it's why the government systems just
can't cope. The US Census Bureau reported, get this, two
point one million workers filed for unemployment, specifically citing AI
displacement as the reason. In twenty twenty five.
Speaker 1 (24:01):
Alone, two point one million sighting AI. How does that
compare to the year before.
Speaker 2 (24:05):
That figure is triple what it was in twenty twenty four. Triple.
The system is just being completely overwhelmed.
Speaker 1 (24:11):
Okay, so jobs disappear, this fast productivity technically soars, But
who captures the wealth? This is what the source is
called the one percent capture right, or the inequality bomb exactly.
Speaker 2 (24:23):
There was a critical paper cited from the NBER, the
National Bureau of Economic Research, published in twenty twenty five.
They meticulously tracked where the labor savings generated by AI
adoption actually.
Speaker 1 (24:35):
Went in the findings.
Speaker 2 (24:37):
The findings are stark. A staggering seventy eight percent of
all generated labor savings accrue directly to capital owners shareholders.
Speaker 1 (24:44):
Basically seventy eight percent, what about the workers?
Speaker 2 (24:46):
Another fifteen percent goes to the relatively small group of
high school workers needed to supervise and maintain these new
AI systems, the AI supervisors as they call them.
Speaker 1 (24:56):
Okay, seventy eight plus fifteen, that's ninety three percent. What's
left for the displaced.
Speaker 2 (24:59):
Worker a minuscule seven percent, just seven percent trickles down
to the millions losing their jobs, and that's usually just
through meager severance packages or frankly highly ineffective retraining programs.
Speaker 1 (25:10):
This concentration of the winnings, it feels unprecedented. We see
Nvidia's market cap soaring to three point four trillion dollars.
That's up twelve hundred percent since twenty twenty two.
Speaker 2 (25:20):
It's astronomical. And you also have figures like Sam Altman's
twenty twenty five compensation package cited in the Sources one
point two billion dollars in stock one point two billion. Well,
millions are losing jobs exactly.
Speaker 1 (25:32):
This explosive wealth creation for a tiny few is happening
at the exact same time as sharp economic decline for
the vast majority of blue collar and even mid skill
white collar workers.
Speaker 2 (25:42):
But isn't the counter argument always? While these founders, these investors,
they took massive risks to build this technology. Why are
the sources so focused on the wealth transfer aspect?
Speaker 1 (25:52):
Because the transfer isn't just about a few massive CEO packages,
it's systemic. The data clearly shows that median wages for
non college educated man who make up a huge chunk
of the displaced workforce in areas like trucking and manufacturing,
have fallen six point two percent in real term since
twenty twenty three. Adjusted for inflation. They are earning less.
Speaker 2 (26:11):
So the economic benefit is just incredibly polarized, which brings
us to the G and E coefficient, the measure of inequality. Right.
The Genie coefficient measures income inequality. Zero is perfect equality.
One is maximum inequality, where one person has everything. The
US figure hito point four to nine in twenty twenty five.
Speaker 1 (26:30):
Zero point four to nine. How does that compare historically?
Speaker 2 (26:33):
That's the highest it has been since nineteen twenty nine.
Speaker 1 (26:35):
Nineteen twenty nine, the year the crash.
Speaker 2 (26:37):
The sources draw a very deliberate parallel there. Nineteen twenty
nine led to the Great Depression, which eventually resulted in
the New Deal. Massive systemic changes to protect labor regulate capital.
Speaker 1 (26:48):
And the argument is AI is the modern accelerant driving
us back to that nineteen twenty nine level of inequality.
Speaker 2 (26:53):
Exactly, but crucially without the political mechanisms or maybe the
political will for a new Deal style response this time around.
Speaker 1 (27:01):
And we see companies making structural choices that reinforce this
hoarding of the gains, don't we. The Walmart example was striking.
Speaker 2 (27:07):
Yeah, Walmart's big AI deployment in their e commerce division
reportedly saved them two point one billion dollars annually in
labor costs. Just labor costs two.
Speaker 1 (27:15):
Point one billion. What could they have done with that?
Speaker 2 (27:17):
Well, the sources calculate that money could have paid forty
two thousand warehouse workers a decent fifty thousand dollars salary each,
forty two thousand jobs funded, But they did do that now. Instead,
the company used that labor saving to help fund a
massive twenty billion dollars stock buy back, returning the capital
directly to shareholders.
Speaker 1 (27:37):
The corporate incentives are just structurally geared toward capital now,
not labor, it.
Speaker 2 (27:42):
Seems that way, and this also creates another looming crisis,
a fiscal crisis for the government. Household displaced workers, well,
they aren't paying income tax, they aren't paying payroll taxes.
The irs, according to the sources, projects a potentially catastrophic
one point one trillion dollar federal tax revenue shortfall by
twenty thirty.
Speaker 1 (28:00):
One point one trillion dollars just from labor displacement and
wage stagnation.
Speaker 2 (28:04):
Purely from that effect. And think about states like Michigan
or Ohio where maybe twenty twenty five percent of the
workforce is concentrated in manufacturing and transportation. The two sector
is getting hit hardest and fastest for them. This tax
revenue collapse isn't just a problem, it's potentially existential.
Speaker 1 (28:19):
Okay, So if the problem is this systemic, this massive,
this immediate, what's the policy response been. This is where
the source material gets frankly, deeply disheartening. It paints a
picture of total failure. Governance just isn't keeping pace.
Speaker 2 (28:35):
Not even close.
Speaker 1 (28:36):
Let's look at the common solutions people propose and why,
the sources argue they've basically failed already.
Speaker 2 (28:41):
All Right, where do we start? Retraining? Is usually the
first thing people mention, right.
Speaker 1 (28:45):
Yeah, the classic answer to industrial change when one industry declines,
you retrain for the new one. For decades, the advice
was simple, learn to code.
Speaker 2 (28:54):
That mantra it's now completely obsolete. Why because, paradoxically, coding
is arguably one of the first high value cognitive jobs
that AI successfully ate.
Speaker 1 (29:04):
We saw the GitHub data earlier right.
Speaker 2 (29:06):
Githubs twenty twenty five report sixty one percent of all
new code is now AI generated. You can't effectively retrain
millions of people into a field that is itself rapidly
becoming automated. It makes no sense.
Speaker 1 (29:17):
Okay, so learn to code is out. What's the new advice?
Learn to manage AIS. Become a prompt engineer.
Speaker 2 (29:23):
That's the new buzzphrase. But the problem is scale. It
just doesn't scale to absorb eighty five million displaced workers globally,
or even the millions displaced in the US.
Speaker 1 (29:34):
How many prompt engineering jobs are there?
Speaker 2 (29:36):
Roughly the sources estimate only about one hundred and eighty
thousand specialized prompt engineering or AI oversight jobs worldwide. Compare
that to say, fourteen million displaced coders alone, let alone
truck drivers, call center agents, analysts.
Speaker 1 (29:51):
The math just doesn't work.
Speaker 2 (29:52):
It doesn't The source material really emphasizes the difference here.
Past disruptions, like say, the shift from farming to fact work,
created large new classes of jobs that could absorb millions.
This disruption, it seems to be eliminating entire classes of
jobs and replacing them with highly centralized, machine managed systems
that only need a tiny handful of human supervisors.
Speaker 1 (30:14):
Okay, so retraining looks like a dead end for the masses.
What about universal Basic income UBI? That's often floated is
the big solution for technological unemployment?
Speaker 2 (30:22):
It is, and it's politically attractive in some ways, but
the sources just point to the immediate financial reality. Yeah,
it doesn't work right now?
Speaker 1 (30:29):
Why not? The math?
Speaker 2 (30:30):
The math is brutal. Andrew Yang's proposal, which was kind
of the benchmark, one thousand dollars a month for every
American adult, would cost roughly three point two trillion dollars
a year just for the US.
Speaker 1 (30:41):
Three point two trillion. And how much tax revenue is
AI realistically generating that could fund this?
Speaker 2 (30:47):
Current projections cited put AI related tax revenue at only
about one hundred and eighty billion dollars annually. The gap
is just enormous.
Speaker 1 (30:56):
Right now, UBI is fifthally impossible without massive disruptive new.
Speaker 2 (31:00):
Taxes, taxes that simply haven't been implemented and frankly face
huge political opposition. Plus there's another wrinkle. The trials we've seen,
like the big one in Finland from twenty seventeen twenty eighteen,
what did that show? It showed that while UBI did
reduce stress and improve well being for the participants, which
is good, it also seemed to increase unemployment slightly because
people receiving the basic income stopped actively searching for work.
Speaker 1 (31:23):
So it potentially solves the poverty problem, but might actually
reinforce the displacement problem.
Speaker 2 (31:28):
That's the concern raise in the analysis. Yeah, it's complex.
Speaker 1 (31:31):
Okay, If retraining and UBI are problematic, what about regulation
trying to control the pace or impact of AI deployment.
Speaker 2 (31:40):
The regulatory landscape is well, shocking is one word for it.
Shockingly porous. Europe's big EUAI Act, passed in twenty twenty four,
got a lot of headlines.
Speaker 1 (31:51):
Yeah, it was a lotted as a global first, banning
high risk systems like social scoring, some facial recognition.
Speaker 2 (31:58):
Right, but there's a massive loophole, critical one, which is
the act specifically exempts what it calls productivity.
Speaker 1 (32:03):
Tools preativity tools.
Speaker 2 (32:05):
You mean the very large language models, the clauds, the
GPTs that are responsible for the mass job destruction we've
been talking about. Regulators focused on the scary or Wellian
big brother stuff, but they completely missed the immediate economic
threat decimating the workforce.
Speaker 1 (32:18):
So the result is zero enforcement against AI driven labor displacement.
Speaker 2 (32:23):
Pretty much zero effect on that front. And in the
United States, the source material notes quite simply there is
no comprehensive federal AI law whatsoever. Nothing. The world's largest
economy is flying completely blind.
Speaker 1 (32:35):
Into the storm facing this complete policy vacuum. Has there
been any resistance from workers?
Speaker 2 (32:41):
There has You're seeing the emergence of what some are
calling let out.
Speaker 1 (32:45):
Two point zero neo Luddites.
Speaker 2 (32:46):
Yeah. An anonymous group called the AI Resistance Network claimed
responsibility for coordinated arson attacks at three WAYMO depots in
twenty twenty five, trying to physically destroy the autonomous vehicle.
Speaker 1 (32:57):
Infrastructure and organized labor.
Speaker 2 (32:59):
We also so saw the Teamsters union organizing blockades at
Aurora's big trucking hub in Dallas. They managed to delay
about forty thousand shipments of crucial goods trying to exert pressure.
Speaker 1 (33:10):
How did the authorities respond to this resistance.
Speaker 2 (33:12):
Swiftly and unforgivingly. The FBI is now officially classifying AI
sabotage like the Waimo arson, as domestic terrorism. The state
is clearly protecting the technology deployment, not the displaced workers.
Speaker 1 (33:25):
And even when conventional policy proposals are made, like taxing
the tech, they.
Speaker 2 (33:29):
Get killed instantly. The sources highlight a proposal by respected
economists Darren Asimoglu and Simon Johnson. They suggested a modest
two percent tax just on AI capital expenditures, a basic
robot tax. What was the goal of the tex specifically
to fund transition stipends and maybe new vocational training for
workers displaced by AI? Seems reasonable, right?
Speaker 1 (33:51):
Yeah? What happened?
Speaker 2 (33:52):
Tech lobbyists reportedly killed that proposal in Congress within forty
eight hours of it.
Speaker 1 (33:56):
Being introduced forty eight hours.
Speaker 2 (33:58):
The political will to regulate this industry for the benefit
of labor, the sources conclude, it simply does not exist
at the moment.
Speaker 1 (34:06):
Okay, So, given this trajectory, the blinding speed of adoption,
that ninety nine point nine to eight percent cost collapse
in cognition and this almost complete policy failure. The sources
conclude by looking forward, and it's a necessary look, even
if it's pretty dark. They set twenty thirty as a
kind of critical inflection point, yeah, a crunch point, and
they outline three potential futures for humanity, three stark scenarios
(34:28):
based on the current path.
Speaker 2 (34:30):
Right, The first one, and sadly, the one they give
the highest probability seventy percent likelihood, is called the Great Hollowing.
Speaker 1 (34:36):
The Great Hollowing? What does that look like?
Speaker 2 (34:38):
In this scenario, AI continues its relentless march. It captures
maybe forty percent of global GDP by twenty thirty five,
simply because the cost of machine cognition is so incredibly low.
Speaker 1 (34:48):
Forty percent of GDP. What does that mean for jobs?
Speaker 2 (34:52):
It means an estimated one point two billion jobs vanish globally.
Speaker 1 (34:56):
One point two billion.
Speaker 2 (34:57):
That's the projection mass unemployment, wage collapse. It triggers such
severe social unrest that governments are forced to impose immediate
capital controls, trying desperately to stop the hoarding of wealth
by the AI owners.
Speaker 1 (35:11):
These the capital controls work, not.
Speaker 2 (35:13):
Really, according to the scenario, It just causes the AI
firms which are highly mobile to instantly relocate to more
permissive regulation free zones. Thinks Singapore, maybe the UAE, which
further cripples Western economies that try to regulate.
Speaker 1 (35:27):
And the end result for society.
Speaker 2 (35:29):
Inequality reaches levels that the sources. Compared to feudalism, you
have a vast majority subsisting on meager government stipends or
charity provided by the tiny elite who own the AI.
The means of production.
Speaker 1 (35:40):
That's bleak okay, seventy percent chands with great hauling with
scenario two.
Speaker 2 (35:44):
Scenario two is called the neo ludite backlash. This one
is given a twenty percent likely NEOLAA.
Speaker 1 (35:48):
Backlash, so the resistance succeeds sort of.
Speaker 2 (35:51):
In this scenario, public fury, fueled by the human toll
the stories of Maria jesh Emily multiplied by millions, becomes overwhelming.
It forces governments around twenty thirty perhaps to enact a
global moratorium on further development and deployment of generative AI.
Speaker 1 (36:08):
They hit the brakes. Stop the job losses.
Speaker 2 (36:11):
Sounds better, It stops the immediate bleeding, yes all, but
the consequence, according to this scenario is severe global stagnation.
Innovation just stalls completely in the West. Why, Because AI
is becoming fundamental to scientific discovery. Are in d efficiency?
You turn it off, you slow everything down. Meanwhile, nations
like China maybe Russia, who likely ignore the ban and
(36:32):
continue aggressive AI.
Speaker 1 (36:34):
Development, they leap frog ahead.
Speaker 2 (36:35):
They leap frog ahead, They dominate global technology and military development,
economic growth. The West, having banned the key driver of progress,
enters a projected fifty year period of relative economic and
geopolitical decline. It trades the short term pain of displacement
for long term stagnation.
Speaker 1 (36:52):
So you either get feudalism or stagnation. Is there any
optimistic outcome?
Speaker 2 (36:55):
There is a third scenario, but the sources give it
only a ten percent likelihood. It's called the abundance pivot.
Speaker 1 (37:01):
The abundance pivot. Okay, what happens here?
Speaker 2 (37:03):
This is the future where we actually manage to harness
the incredible productivity gains from AI for broad societal benefit.
Productivity becomes so massive that it eventually can fund things
like a universal twenty hour work week, maybe by twenty forty.
Speaker 1 (37:18):
And UBI becomes viable.
Speaker 2 (37:19):
UBI finally becomes fiscally viable in this scenario. Why because
a robust global system of robot taxes or AI capital
taxes is successfully implemented, generating an estimated fifteen trillion dollars annually,
enough to cover the cost we talked about earlier.
Speaker 1 (37:35):
So humanity reaches a kind of post scarcity.
Speaker 2 (37:37):
Era potentially yes, a world of abundance driven by machine intelligence.
But and this is the absolutely critical caveat in the
source material, humanity only reaches the state after surviving the
twenty thirties bloodbath O, meaning we likely still have to
go through something resembling the Great Hollowing first, the mass displacement,
the unrest, the inequality spike, before society finds the political
(37:58):
will and the mechanisms to pivot towards abundance. The journey
to utopia is littered with massive societal risk and suffering
hashtag tag outro Wow.
Speaker 1 (38:09):
This has been a deeply necessary, I think, but also
a very challenging deep dive into this source material.
Speaker 2 (38:14):
It's heavy stuff.
Speaker 1 (38:15):
It really is. The central and I think undeniable argument
we've pulled from every paper, every forecast here is that
the mass destruction of millions of jobs by AI, it's
not some hypothetical risk for our kids to worry about.
It's a present catastrophe. It's defining the economic reality of
the twenty twenties right now.
Speaker 2 (38:32):
Yeah, we've mapped the sheer speed of it, that unbelievable
ninety nine point nine eight percent cost collapse in cognition.
We've mapped the winners, haven't we The soaring market caps
of Nvidia, the one point two billion dollar stock packages
for CEOs like Altman. That's seventy eight percent of wealth
flowing straight to capital.
Speaker 1 (38:46):
And we've tried to put human faces on the losers.
Maria the truck driver now on food stamps. Rajesh, the
call center lead in Manila who can't find work, Emily
the skilled copywriter, now ghostwriting for fifteen bucks an hour.
Speaker 2 (39:01):
It's brutal, and the clock is really ticking, isn't it.
The projections say by twenty thirty, just a few years away,
another three hundred million jobs globally are likely to vanish.
Speaker 1 (39:11):
Three hundred million more. It feels like a silent coup,
the greatest and certainly the swiftest transfer of wealth and
power in modern human history, happening right under our noses and.
Speaker 2 (39:22):
The sources they bring it back to that historical context
near the end the First Industrial Revolution, the original Luddites, right,
they protested the machines taking their weaving jobs, and yes,
they were crushed by the state.
Speaker 1 (39:34):
They lost.
Speaker 2 (39:34):
They lost, but their children or certainly their grandchildren, eventually
inherited the factories. They became the modern working class, and
ultimately they did benefit from the massive new productive capacity
the machines created through unions, regulations, the welfare state.
Speaker 1 (39:49):
Okay, so what's the final thought here? The takeaway?
Speaker 2 (39:52):
This leads us to the final lingering question the source
material poses for you, the listener, to really grapple with
the question is whether today's displaced workers, the Mariahs, the Rajeshas,
the Emily's, will inherit anything at all this time, or
whether the masters of AI, the owners of the algorithms,
will simply hoard the entire future, leaving the displaced with
(40:13):
neither jobs nor any share in the revolutionary abundance the
technology clearly promises.
Speaker 1 (40:18):
Will they inherit anything or will the future be hoarded?
That's a chilling question to end on. Thank you for
walking us through that.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com