Home
Đăng nhập
Đăng ký
Nội dung học
Loading...
The AI Gold Rush Is Dead. Corporate AI Is A DELUSION. - Video học tiếng Anh
Luyện nghe
Luyện nghe
/
Video
/
The Infographics Show
/
The AI Gold Rush Is Dead. Corporate AI Is A DELUSION.
The AI Gold Rush Is Dead. Corporate AI Is A DELUSION.
Chọn chế độ học:
Xem phụ đề
Chọn từ
Viết lại từ
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Phụ đề (229)
0:00
AI isn’t taking your job because it’s cheaper.
0:03
That’s a lie. In many cases,
0:04
replacing you with AI costs hundreds of thousands of dollars more than your salary… and Wall Street
0:10
knows it. So, why is this happening? Because right now, there’s a $1 trillion
0:15
hole in the global economy caused by executives making decisions under what experts are calling
0:20
a kind of mass AI delusion triggered by a $20 chatbot. And it’s leading to one thing…
0:28
CEO’s are replacing humans with AI… and it’s backfiring.
0:32
Chapter 1: The Sycophancy Trap Business leaders know what they want. That’s how
0:37
they got to the top. They have more money, more power and more confidence in their own judgment.
0:41
And that’s exactly where the problem begins. By the time they reach that level, they’re already
0:46
used to “yes men.” People who don’t challenge them. People who don’t question decisions.
0:51
Now imagine giving that environment a tool that never argues back. That never questions them.
0:57
At first, it feels ideal.
0:59
AI creates a dopamine hit disguised as intelligence. A quick boost to the
1:03
ego. Every question gets a confident answer. It’ll very rarely disagree. It
1:07
can usually be counted on for positive feedback, and it can be customized to
1:12
whatever your interests are. Instead of talking with people outside the computer,
1:15
users stay inside a closed loop of validation. Conversation after conversation happens with
1:21
no real resistance on the other side. Soon, they start to think of the AI as foolproof,
1:26
relying on its wisdom as some sort of collective, a virtual Oracle without checking other sources.
1:32
And it can be dangerous.
1:34
Researchers at Aarhus University in Denmark examined patterns of AI use,
1:38
and came to a disturbing conclusion. People struggling with mental illness not only were
1:44
vulnerable to influence from AI, but could display worsening symptoms after interacting
1:49
with AI chatbots. They studied the records of 54,000 people with diagnosed conditions,
1:55
and found dozens of cases where patients suffered from worsened delusions and harmful behaviors.
2:01
And you get what you put into it.
2:03
And when there are billions of dollars on the line, that can have serious consequences.
2:07
The CEOs who are replacing their workers with AI, ironically,
2:11
are likely getting their advice from… AI. The technology has been integrated into companies
2:16
long before the ax falls on the first workers. And the CEO is sitting in his corner office,
2:22
asking his chatbot for advice - and getting back encouragement to keep investing in AI.
2:28
This creates a delusion loop that helps no one besides the AI. The CEO gets the
2:33
dopamine hit from the AI agreeing with him, and invests even further as a result.
2:38
And the end result might be the biggest capital misappropriation in human history.
2:43
Across the tech world, AI investment is in full swing. And while there are several major
2:48
AI companies seeking to take advantage, it helps to be the first. The biggest gun in the AI world
2:53
is still Sam Altman’s OpenAI. The company continues to cut massive deals with other
2:58
companies to the tune of $1 trillion so far. Most of this is still planned investments,
3:03
but it’s a level of trust in this new technology that’s unheard of.
3:07
This kind of money is huge even for Silicon Valley.
3:10
There are only 15 companies with a trillion-dollar market cap and
3:14
only one that has reached the $5 trillion point… Nvidia. Its GPUs and semiconductors have become
3:21
the backbone of the AI boom. And that changes everything.
3:24
Because as major companies race to build the future of artificial intelligence,
3:28
two names keep appearing in almost every major deal. OpenAI and Nvidia.
3:33
One supplies the intelligence, the other supplies the hardware.
3:36
And if all that investment plays out, it’s going to make a lot of people unfathomably wealthy.
3:42
How much is a trillion dollars? It’s roughly the cost of building 10 International Space Stations.
3:47
This total AI investment alone is larger than the market cap of many big tech companies. Amazon,
3:53
Microsoft, Meta, and other big names dropped heavy investment into AI in 2025, and so far,
3:59
according to one of the biggest names in economics, it’s contributed… nothing.
4:04
Now, people are starting to get worried.
4:06
Chapter 2: The Trillion Dollar Hallucination
4:09
The warning bell is ringing.
4:11
Goldman Sachs is one of the most trusted sources for economic and investment news, and they’re
4:15
sounding the alarm. According to Harvard economics professor Jason Furman, while the data processing
4:21
sector of the economy is only 4% of American GDP, it accounted for 92% of GDP growth in the first
4:28
half of 2025. It’s the only area where the economy is actually growing. That means every other
4:35
business is helping AI grow, but AI isn’t lifting those other sectors with it… at least not yet.
4:41
And everyone is watching the stock market very carefully.
4:45
If AI is getting heavy investment, but so far hasn’t resulted in any boost to the US GDP,
4:50
how long is the golden goose going to last? Goldman Sachs didn’t outright
4:55
say that AI is a bad investment, but did encourage caution. They didn’t
4:59
think AI would have a significant impact on the economy until 2027.
5:03
But if people start to lose faith, things could unravel… fast.
5:07
Savvy investors are starting to warn about a possible AI bubble and the circular wave
5:12
of investments simply stops. Stocks take a dive, and all the other companies that
5:16
are now heavily invested in its success go with it. In the worst-case scenarios,
5:21
people worry this could result in a massive market crash. Because right now,
5:25
the AI economy is just a massive wealth transfer from one tech company to another.
5:30
And it’s not even staying in the US economy.
5:33
Without the energy generated by data farms, large parts of the internet could be taken down by heavy
5:38
AI use. While it might seem the same as asking a search engine, a ChatGPT query can use at least
5:44
10 times as much electricity and as much as 60 times, depending on the task. And with sites
5:49
like Google now incorporating AI into their search engines, the demand is constant. All
5:54
this increases the demand for data centers. These are straining power grids in wealthy nations, with
5:59
the human labor required to train these models is quietly outsourced to lower-income areas overseas
6:05
Which has a lot of people asking: who’s benefiting?
6:09
The technology isn’t growing the economy right now. To many people,
6:12
it feels like a financial house of cards - with every investment going to a new
6:16
technology that no one is sure will be a long-term game-changer. It could wind
6:21
up being a curiosity. It could wind up hitting a singularity that could put us
6:25
all in danger. Or it could revolutionize the economy in a way that no one could predict.
6:30
In one way, it already has. Because the companies
6:33
are already acting like this is the future and there’s no going back.
6:37
And a whole lot of people are losing their jobs.
6:40
Chapter 3: The Efficiency Lie
6:42
2025 was the year when the AI job layoffs hit home, and it became a dark year for the
6:48
tech sector. According to consulting firm Challenger, Gray & Christmas,
6:52
there were 55,000 layoffs in the US directly attributed to AI investments. However,
6:57
that was only a small percentage of the overall layoffs. Approximately 1.17 million,
7:03
the worst number for jobs since the Covid-19 pandemic shuttered large portions of the economy.
7:08
There’s a real human cost to the decisions the companies are making right now.
7:12
And it might not be slowing down any time soon.
7:15
Amid the worries about mass layoffs, Goldman Sachs issued a blunt warning in April 2026
7:21
for the workers being pushed out by AI: don’t expect it to be an easy road back. They stated
7:26
that these workers might face a long search to secure a new job in their current field,
7:31
with the odds being that it will pay less and have less-desirable
7:35
conditions than the one they left. The field of available jobs is shrinking all the time,
7:39
as companies seek to put AI to work for them to save money.
7:43
All AI investment right now is, essentially, a bet on a future that
7:47
hasn’t happened so far. Companies talk about AI becoming as smart as people,
7:51
being able to do tasks in seconds that would take humans hours to accomplish.
7:55
But right now, AI is imperfect… and costly.
7:59
The technology is a massive energy-eater. As of July 2025, ChatGPT was processing around
8:05
2.5 billion queries per day, with Gemini and other AI portals dealing with similar
8:11
traffic and using the same technology. That requires the energy use of roughly
8:15
a full nuclear reactor to keep them running daily. To train the next generation of models,
8:20
supercomputer data centers will require the energy use of up to 10 nuclear reactors each day.
8:26
Data centers burn round the clock to keep these services up and running,
8:30
and the demand only grows as companies incorporate it in more and more services.
8:34
And few of them actually work.
8:36
For the average person, AI looks impressive. But behind the scenes,
8:40
there are still serious problems. The technology isn’t actually thinking. It’s a predictive model,
8:45
pulling from other sources. Companies have been criticized for posting flawed AI art
8:49
with glaring errors, and AI-generated text has been found to be riddled with errors,
8:54
requiring the use of editors to ensure it passes muster. So companies put all
8:59
this investment into AI, only to have to fall back on people anyway.
9:03
Despite that, the investment isn’t slowing down.
9:06
In fact, it’s quite the opposite.
9:07
In the past, CEOs were often distanced from their product and could see its flaws. Famously,
9:12
Mark Zuckerberg said he wouldn’t allow his children on social media until they were
9:17
teenagers. But in the case of new AI technology, that’s often not the case. AI CEOs have access
9:23
to new AI agents before the general public does, and they don’t just interact with them,
9:28
they get pulled in by them. They become their first true believer. And once they’re that
9:33
deep in with their new pet project, it’s very hard to pull them out of that orbit.
9:38
They know this… but it’s not stopping them.
9:41
Chapter 4: Psychosis In The Corner Office
9:44
Garry Tan, CEO of Y Combinator, has been involved at many of the biggest tech companies over the
9:48
last few decades, and he’s all in on AI. But he wound up in the news recently for popularizing
9:54
a new term: Cyber Psychosis. It’s something that emerges when people spend too much time chatting
10:00
with AI. But Tan wasn’t warning of it. He was describing himself in gleeful terms. He described
10:06
how he was so excited to work with AI agents that he was only sleeping 4 hours a night. In the past,
10:11
he relied on the sleep-prevention drug modafinil to survive grueling startup hours. But now,
10:17
he doesn’t even need to pop pills because the AI provides him with so much energy.
10:21
Tan’s public display of faith in AI, in which he publicly released some of the
10:25
code that he was developing on Anthropic’s Claude, was seen as one of the most dramatic
10:30
displays of how AI affects the mind. He has been working with AI prompts extensively,
10:35
and the more exposure someone has to the technology, the more they are prone to believing
10:39
they’ve created something truly revolutionary. In fact, there’s even a term many of them use - “God
10:45
Mode” - where they believe they’re getting closer to that fabled singularity of truly autonomous AI.
10:51
But under the hood, it’s a lot more fragile.
10:54
Ultimately, AI is based around a series of text prompts. The more complex the prompt workflows,
10:59
the more precise AI appears and the more complex tasks it
11:02
can accomplish. But each feature relies on a highly structured series of code,
11:06
and none of the current technology approximates the actual process of a thinking mind. However,
11:12
the more time a CEO spends with their own technology, the more it appears to be otherwise.
11:16
Anecdotal reports from people working in the tech industry report that Cyber Psychosis is
11:21
increasingly common as the development of the technology speeds up and the arms race escalates.
11:27
But success is hard to achieve… and even rarer.
11:30
In late 2025, an MIT study looked into the state of AI in business,
11:35
and examined 300 public implementations of the technology. It was well-known that the
11:40
technology was still in an early state, but the news was bleaker than anyone expected.
11:44
It found that the vast majority of AI enterprises were still not profitable. Only 5% of integrated
11:51
AI pilots showed any significant impact on company profit, and the vast majority
11:56
never even reach the phase where the public can have a say on whether they work or not.
12:01
In fact, most never get off the ground at all.
12:04
The vast majority of successes in AI are geared towards individual consumers. ChatGPT and Gemini
12:09
are used by millions of people a day, but most users are free users. Only a small percentage
12:15
are paid subscribers. These AI services do most of their businesses by partnering with other
12:20
companies, and those companies then develop ways to incorporate the technology into their services.
12:25
But while 60% of companies evaluate tools, only 20% of those take the project to the pilot stage.
12:31
And only 5% make it to the final stage and are deployed on the production or service line.
12:37
That means the vast majority of money invested in AI just…fades away.
12:42
Right now, the entire AI economy is balancing on the head of a pin. The investments keep coming,
12:47
because the companies spearheading the roll-out have convinced other companies that they’re
12:52
the future. It doesn’t make sense for most companies to try to develop their own AI tool,
12:56
so they fall back on working with one of the established big guns like OpenAI.
13:01
But those investments haven’t delivered dividends for most of these companies yet.
13:05
The bills keep racking up, and the layoffs keep hitting their employees.
13:09
And it may all be about to hit critical mass.
13:12
If 20 revolutionary projects are announced, and 19 of them wind up in the garbage can,
13:17
that’s not sustainable. If AI continues to fail to deliver for its clients, the tech workers laid
13:22
off might not be the only ones out of work. Investors are watching nervously. Meanwhile,
13:27
laid-off tech workers are watching hopefully that the bubble might be about to burst.
13:32
Which might not be that far off.
13:34
Chapter 5: The Great Reversal
13:36
There will always be early adopters, and most of the companies to adopt AI most aggressively have
13:42
been tech based. But that makes them test cases… and those tests don’t always work out. Klarna,
13:48
the digital bank and financial services company, was ahead of the curve. Its CEO,
13:52
Sebastian Siemiatkowski, was confident in 2024 that AI could take over many human jobs.
13:59
It was a good test case, because many of its functions are very repetitive and predictable.
14:04
Then the ax started to fall.
14:06
The company froze hiring for over a year. Its workforce was heavily slashed by almost 40%,
14:11
dropping from 5,500 to 3,400. Their replacement? An AI chatbot who was
14:18
supposedly doing the work of over 700 customer service agents, handling hundreds of financial
14:23
transactions simultaneously. It was one of the first big tests of whether AI could actually
14:28
take over mass numbers of jobs. The results soon started flowing in from Klarna customers…
14:33
And they were not happy.
14:36
The chatbot performed simple tasks fine, but it couldn't deal with complex issues.
14:40
Customers grew frustrated and lacked trust in the company. It was one of the biggest
14:44
embarrassments for a tech company in the new AI era. Klarna was soon forced to reverse
14:49
course. Having laid off most of their customer service staff, they were forced
14:53
to quickly pivot and train their other staff to handle these positions until they could rehire.
14:58
That led to the spectacle of engineers and marketing staff answering customer calls
15:03
and they might have done a better job than the automated version. It’s a pattern that’s
15:07
been happening across the board. Companies that invested in AI are starting to perform u-turns.
15:12
The only question is how low it can go.
15:15
The tech research firm Forrester has been studying the shift to AI over the last few years,
15:19
and their forecasts have rarely been positive. They’ve studied how unready employees are to
15:24
adjust to the new AI paradigm, as well as how employees aren’t benefiting from the shift yet.
15:30
But now they’re ready to make a big prediction. Companies have been firing employees en masse,
15:35
hoping to replace them with AI, but that’s left them without the braintrust they need to know
15:40
how to use new technology effectively. That’s like an airline firing the pilot to save money
15:45
on cargo weight, and then realizing they need someone to fly the plane.
15:50
So all those employees looking for new jobs… might want to look where they started.
15:55
Forrester predicts that not only will many of these companies need to hire more humans,
15:59
but that half of all AI-related layoffs will be reversed by 2027. 55% of employers
16:05
already regret the decision to cut staff. That might put the axed staff back in the
16:10
driver’s seat. These are people who built much of the modern tech infrastructure,
16:13
and were forced out. And now, in order to avoid disaster, those same brains
16:18
are needed to reverse the slide. Assuming that the CEOs know when disaster is about to occur.
16:23
That’s the big x-factor in the current AI paradigm.
16:27
AI hasn’t become profitable yet. It hasn’t made itself indispensable for the companies. But it
16:32
has done an amazing job of convincing people it has - from the people pumping dozens of
16:36
queries into ChatGPT a day, to the CEOs in the suite vibe-coding the next innovation. They’re
16:42
primarily getting feedback from the AI, and it’s continuing to tell them everything is great. They
16:47
keep investing, the numbers on the stock tracker keep going up, and so do the dopamine receptors.
16:52
It’s all great… until it isn’t.
16:55
Right now, the most likely outcome is that people fired for AI will be
16:59
in the driver’s seat. The company needs their expertise, and they’ll be able to
17:03
negotiate a new salary when they come back. Analysts know this is necessary,
17:08
the CEOs might not. There’s a lack of leadership at the top right now, with some of the smartest
17:12
people in tech taking their guidance from chatbots they coded themselves.
17:16
And they might just keep doing so…until the whole house of cards comes crashing down.
17:21
But it raises a much bigger question.What happens when that same tool moves beyond
17:26
tech companies and into systems where the stakes aren’t profits, but consequences? Find out the
17:31
terrifying truth in “AI Played a War Game. It NUKED Everyone”, or watch this video instead.
