You're at a really fancy restaurant with

your friends. The type of fancy where

the waiter comes around with that little

metal bar and scrapes breadcrumbs off

the tablecloth and you're stuffed. You

gorged yourself. You're kicked back in

your chair. You had some wine. You had

steak. And the bill comes out. You're in

such a good mood, you don't even think.

You just grab the bill, look at your

friends, you say, "This is on me." You

open the bill. You think that's a lot of

zeros. That is exactly what is happening

to the market after this study just

dropped yesterday.

The conclusion of the study was that

over $600 billion in spend on AI at the

corporate level doesn't really get you

much. The study in question was put

together by the National Bureau of

Economic Research. And what they did is

they took 6,000 CEOs from four

countries, US, UK, Germany, and

Australia. They sat them down and they

asked for their quantified experience in

their companies on what this AI rollout

is looking like for them. The results

would make Sam Alman's eye twitch. 90%

of them revealed that AI is having zero

impact on employment or productivity.

And that's not surprising considering

their employees aren't using it. AI use

for people that have their tools at

those companies, 6,000 companies,

average AI usage is about 1.5 hours per

week for the average employee. And 25%

of them, a quarter, aren't using it at

all. This is not a Vibe statistic. This

is a peer-reviewed study with a massive

sample size. This has actually happened

before, but last time it took 15 years

to show up in the data. Apollo economist

Torsten Sllock made this connection with

something called Solo's paradox, which

is named after Robert Solo. It was

coined in 1987. And what he noticed is

that, hey, we're getting computers

everywhere. They're in offices. We're

using them for everything, but are they

actually making us more productive? It

doesn't seem like it. and it's not

showing up in the productivity

statistics. If we have all this new

tech, you'd think that people would be

more efficient and be able to crank out

more work, right? He summarized this

nicely by saying, quote, you can see the

computer age everywhere, but in the

productivity statistics. So, tech

proliferated, but it did not make us

more productive. And he was able to

quantify this, but his name is making

the rounds again. If we have all these

AI tools, these AI tools are supposed to

be so good. We're told that they are

going to achieve AGI and they are going

to replace humans. We're going to be out

of work in mass. We're seeing some early

statistics. You remember I've been

citing the MIT study that says that most

AI pilots don't return any value. And

you could dismiss that is just one

study. Maybe they had an on sample

group. But now we have this NBER

peerreed article sourced from 6,000 CEOs

across four countries and it's reporting

the same thing, almost identically the

same thing in terms of the statistics.

Although the other one was on a AI

pilots, this is on continuous AI usage.

So what is going on here? Why are we not

more productive because we have access

to so many of these AI tools? And

really, I'm of two minds. I am an

optimist, but I'm not an optimist in the

same way that Sam Alman is an optimist.

I don't believe that AI is going to

replace human intelligence anytime soon.

That being said, I use AI every day. It

has made coding fun for me. Again, as a

software leader, I do not get time to

write code anymore. I manage engineers

and I manage departments. Previously, it

takes up your whole day. You do not get

to put hands on a keyboard, which is

unfortunate because I really liked

coding. Now, with a a wife and being the

sole breadwinner, and having a

three-year-old around the house, another

kiddo on the way, there's just not a lot

of time to write code for me anymore. If

I sit down on the weekend, I really

really don't want to instantiate a new

React app and get all the blah blah blah

spun up. I just have a good idea and I

want it to work on the computer. And AI

has made that possible. Again, it's

quite amazing. In fact, while I did the

research and wrote the copy for the H1B

exposed website that I mentioned a few

videos ago, I did not write any of the

code. I just chatted with claude code

and had it shape the website for me

until it was designed how I wanted and

acted and worked how I wanted. All that

is to say these tools are genuinely

helpful at some level. I don't think

many people dispute that at this point.

But the problem is scale and that's the

core argument that I hope you pick up on

from this channel that I have a nuanced

position with this where I think that

the tools are useful. Not I think they

are in fact useful. These tools are in

fact useful. They are not going to

produce the massive worldchanging

productivity gains that we were promised

though. Yes, maybe in 10 years they

will, but not over the next year or two.

And the financial bit, the bubble of

overvaluation of companies that don't

actually produce anything real and

obviously as a result of this study and

others aren't actually providing a lot

of value is way overblown. Way

overblown. One of you on a comment in my

last video mentioned that the AGI bubble

has popped, but the AI bubble hasn't

popped. And I think that's a fantastic

way to look at it. But as I've mentioned

before, I don't anticipate this wave to

crash and roll back publicly until one

of the major companies falls, until

somebody dramatically misses their

earnings. The writing is already on the

wall. Wall Street is concerned with

what's going on. You see this every time

a company posts a beat in their earnings

report like Amazon for instance and then

announces the insane amount of capex

they're planning on spending for AI

infrastructure over the next year.

Remember that's about 680 690 billion

total uh from all of the five

hyperscalers combined for 2026 their

capex capex guidance we have that from

their earnings calls. So, I'd expect

status quo at least for the next six

months in those earnings calls reports.

Actually, 374 of the S&P 500 companies

mentioned AI in their latest earnings

call. Outside of the Magnificent 7,

there's zero signs in profit margins or

earning expectations. And as we've been

reporting on since last fall, when

everyone called me a nut job and tinfoil

hat for saying that this was a funding

bubble, there is a very clear funding

bubble here. It is so clear that the

tech sponsored media is even reporting

on this. Now, your brief refresher on

that, it's like Nvidia is like a drug

pusher. He shows up to OpenAI. Jensen

Hong walks in there and he says, "Do you

want $100 billion, although it's now

just $20 billion because he walked back

that initial commitment." And Sam Alman

is like, "Sure, I could use $20

billion." And just as he's taken the

money from Jensen, Jensen's like, "One

condition. Got to use that 20 billion to

buy GPUs from us." That's the only

condition. You got to give it back and

we'll give you GPUs and we get the tax

write off because it's depreciation on a

hardware asset. That's how it works.

Money's going around in circles. And

where is it coming from? The VCs. And

eventually, it's going to come from

government bailouts. Once we see

somebody like Nvidia fail or one of

these large hardware manufacturers,

we're going to get some bailouts. You

and I are going to pay to continue to

subsidize it once the party stops and

everybody has cashed out except us.

Except us, not us. We won't hear about

it until after the fact. This is the

bubble I'm talking about is the finance

the AGI bubble. So, you are not crazy

for being skeptical. Your observations

on this technology are valid. If you've

experimented with it, if you've seen how

it's rolled out at enterprise scale, you

are just reading data that everybody

else is willfully ignoring. If your

company is spending millions on AI tools

with no clear ROI, this is why I want to

re-emphasize. Productivity gains for

individuals are real, but this macro

story of revolutionizing organizations

is not proven. If you want the facts and

figures on this, as well as exclusive

content after earnings calls, be sure to

sign up for the newsletter down there in

the description. If you're not

subscribed, let's fix that now. Hit the

subscribe button, hit the bell. Thank

you for watching, and we'll see you in

the next one.