Anthropic's AI Exposure Index & Job Market Impacts

Guys, there is a lot happening with the job market and AI right now, and Anthropic has

just released what they're calling an AI Exposure Index.

Basically, they're showing what jobs are the most exposed to AI.

Actually, got my news on this originally from Connor on LinkedIn.

If you're not following him on LinkedIn already, I don't know what you're doing.

This is where yet the best AI news.

But there's a lot of really incredible things where they're basically showing you what areas

are most exposed to AI, and more importantly, what they're calling the observed AI coverage.

Meaning, look, AI can do these tasks, but how much of those tasks is being done by AI.

And there's some massive discrepancies in different areas.

So we're going to be covering that all in the show today, kicking this over to you, Connor,

because you're the first person I saw to report on this, and your post has gone absolutely

mega viral.

I mean, I see over 1200 likes on this thing.

Tell me about what struck you the most with this kind of labor market impact of AI.

Okay, so first of all, can I just, Jane, can I tell you something I'm so embarrassed about?

So I kind of saw this, and I was in the middle of something else.

So the way that I do LinkedIn posts a lot is like, I don't know, I'll just try to make

it as readable as possible, right?

So I grabbed the chart from the AI, from what was it called the exposure index, grabbed

the most interesting chart, which is sort of like a radial chart, thrown into Canva,

kind of called anthropic AI exposure index, labor market impacts of AI kind of thing.

And I just sort of like, I just kind of create a quick thing in Canva about it, but

Jane, I have a typo in there.

If anybody wants to go like, look, and I misspelled the word, and people start like crushing

me.

I'm like, oh, no, I tried to edit, I couldn't edit it.

And then it goes viral.

And right now, Jane, that's like 130,000 impressions of like, really this one, this is what

you want to be like, yeah, no, don't have to spell it.

I'm like, all right, so anyway, so sorry, and I kept saying like the mistake is mine,

not anthropic.

Okay.

Back to the topic.

But Jane, this is the world that you and I live in, friend.

I just thought this was super interesting, because you know why I liked it so much, Jane,

it wasn't hysterical.

I think anthropic does some of the best reporting out there.

I, by the way, you know, McKinsey, PCG, there's a lot of good reporting out there, but a lot

of times also, they're telling a story, but they're also, look, everybody has their interest.

I'm not saying that.

I'm just saying that I think anthropics work is just very good.

And they just throw things out pretty regularly.

So for this, what they found is something that's probably not that, you know, surprising

where it's like, you know, management, business, software, things like that are really exposed

to this stuff.

But the thing, Jane, you sort of like touched on it, is that it's not that companies

are firing at very high rates, they're just like hiring less.

And this was sort of like, I think this is the big story.

It's like this entry level knowledge work that's sort of super quietly shrinking.

And this is like that bottom of the career ladder that people talk about.

And it's just dropping and dropping and dropping.

So clearly, like, again, no surprise computer programmers hugely exposed, like, whereas

Cooke's mechanics, you know, they're not.

But the, but the thing that sort of like, I thought was like super interesting was like,

and I sort of like wrote this.

I'm just going to like read, sorry, to like read a sentence for my own LinkedIn post

here.

But it says, like, the people most at risk now aren't getting fired there.

Sorry, the people most at risk right now are not the ones getting fired.

They're the ones who can't get in the door.

The early career workers trying to break into knowledge work.

And that's what we're seeing, right, Jane?

Because like, even right now, like there's a job's report right now is sort of like a

little glum, you hear Jack Dorsey saying, oh, we fired 4,000 people because of AI,

which is of course, like, it's just a complete fabrication.

I mean, they did fire people, but it wasn't because of AI.

I think the thing, Jaden, that I would sort of like love to get your take on here is

also just just this in general, but also like this overhang, guys, the gap between what

AI can do and how it's showing up and working and people individually.

It's colossal.

I cannot express how colossal it is.

When you hear people saying it's going to destroy all white color work, it kind of can

right now.

The problem is like, not the problem.

The good thing, actually, is that like these things are just very slow to dissipate

through work through behaviors.

People say like, oh, well, we just hired the different people.

I'm like, but people are already in those jobs.

This stuff is slow to take place, but I cannot emphasize this enough.

I wish I had a microphone to the planet.

We have to upskill now.

Guys, you have to get good on this stuff now.

The best protection against losing your job is really understanding how AI can amplify

what you do.

It's just that and starting to sort of soapbox this, but it's just so critical.

Anyway, Jaden, you and I talked about that a lot.

I just wanted to bring it up.

Okay.

Well, based off of, for the chart, for those that aren't looking at the pictures of

the chart, I'm going to explain a couple areas that I think are most striking to me.

And that is that AI, what it is best at, what it can really do quite well right now, is

management, business and finance, computer and math, architecture and engineering.

And then also we have like office and admin.

And what's interesting is the areas that it's currently doing the most, the current, the

area that it's currently being applied to the most, is business and finance, computer

and math and office and admin, also sales.

And we see that a lot.

There's a little bit, a little bit in legal, a little bit in education, a little bit in

arts and media.

Those are kind of all the top places that it's currently being used at, but, and all of

those like legal and arts and media also are some of the top places that it is, you know,

has capabilities.

Kate, the thing that was the most interesting to me though is where is it the most capable

and yet the least used one of those, one of those that might not surprise you is healthcare

practitioners.

It's, you know, it's pretty good for this, but it's very, you know, kind of small

the amount that you're using it.

Yeah.

And you can, you can look at like, well, there's like HIPAA compliance.

There's a lot of regulation.

It's like tricky.

And I think that does, that is a, a true point.

The other one that really blew me away and I feel curious to take on this Connor architecture

and engineering.

It is great at this, you know, theoretically, no architecture, no engineer is touching

AI.

Nobody.

It's, you know, that, that was an interesting one too.

I sort of picked up on that.

I thought that, and so this is just, guys, if you're listening to the sound of our voice,

this is just Jaden and Connor talking like we don't have insight information on this.

Something on that that I'm just guessing is there's certain things obviously with like

coding and stuff like that.

And I think architecture falls into an interesting category where the AI is exceptional at this

because essentially architecture is a kind of a combination of like design and civil engineering

in some way or another, right?

And on the civil engineering front, it's outrageously good, right, but on the, but on the, on

the sort of like the stylistically, it's sort of like anything else, right?

It can kind of come up with a lot of ideas, but people still want a human architect.

They don't necessarily want, they want to kind of like here, like, hey, I live in this

town.

What are people doing?

They don't want just like AI being like, this is the cool new thing, right?

In the same way like my daughter Lucy who's about to turn 15 probably doesn't want AI like

dressing her because she knows what's cool.

Like she doesn't want to sort of like hear in theory what's cool, right?

The stakes are too high.

So I think that in architecture, I can't believe more architects aren't using it.

That seems very strange.

It is that sort of like massive overhang, but also like, I don't know.

So, you know, legal is a good example, right?

There's a ton of like legal tech.

So Jayden, let me just sort of like share my theory on all this because I was talking to

a guy yesterday about this, this CEO of a company about this yesterday, kind of a smaller

company, but like, but we were just sort of like brainstorming about this.

My theory and like why it's like crushing and coding, why it's crushing and legal is,

and this is what we do with AI mindset with like big, you know, companies, we do this big

behavioral shift with some of the biggest companies in the world.

And the idea here is that if you are sort of like trying to get people to work in a different

way completely, right?

Like they do like name the job and now you want them to do it with AI.

Now they have to like think with AI, it's a behavioral change.

It's not like, it's not AI doesn't really work like a digital transformation.

We say this all the time.

Digital transformation is like, hey, we're switching out, you know, fax machines for email

or switching out this, you know, pencil and paper for Excel.

You're doing the same task, but now you're just doing it like the, the perfect example

that's like a hedge fund, a slower algorithm, now a faster algorithm, right?

But you're doing the same thing.

Why AI adoption so slow is here asking people to think differently, to take things that

they, you know, would never use sort of like a computer for something and now all of

this and think with it.

So it's much more like a person.

So what we say is that the reason that adoption so slow is that your brain just has a hard

time looking at software and thinking about as a person.

So why is like legal doing well?

It's because legal tech has been around before AI.

And so now instead of like that legal tech, they're just switching it out for AI legal

tech, right?

Same with software.

Software always had these, you know, I had had a lot of different things.

But now you just have it with a faster sort of thing.

Whereas in like office administration, it's not like you had a computer doing something

and now you have a better computer doing it.

It's like you had a human doing it and now you have to kind of like think how you change.

So one of the things on this radio here in terms of this overhang is that like it's easier

when you've already been doing something with software and now you sort of like just replace

with AI.

It's harder in architecture where you've never used, you know, like tools and things like

this and this.

I mean, obviously that you some, but now you're asking people to think in a whole different

way.

I think that's just going to be slow.

That's that massive overhang.

And it's that massive variance between what AI can do and what people are used to doing,

you know?

Yep.

100%.

I actually want to show you guys something interesting for anyone watching over on YouTube,

but otherwise we'll explain.

If you go and search for AI for architecture on Google, you're going to see the top result

is actually just, it's called architecture AI because I was like, oh, like what tools

exist out there?

And I'm kind of scrolling through.

Literally the top two results are like AI tools you use as an architect on Reddit.

So Reddit is like the second result.

The first result is called architecture AI.

It's literally just a custom GPT that somebody built.

So the chat GPT, yes, but it's like the entire category of architecture and AI is getting

just like is ranking on Google that should be incredibly a competitive space by just

a custom GPT, which by the way, you just click a button and it's like, help me design

my dream home.

And so like I just clicked it while you were talking.

And it just makes some pictures of like a nice dream home.

But like, guys, this custom GPT that makes like four pictures of what your dream home

could look like is the number one ranking thing?

Okay, contrast that.

If you go and search for something like code with AI, okay, we're seeing like people

have their sponsored results.

There's clawed chat GPT, there's deep AI, there's Devon, there's cursor, there's Gemini,

like there's all the top companies are trying to crush and rank for this, right?

It's a super, anyways, it's so fascinating to me that it feels like some industries are

completely forgotten.

And the other thing that I want to bring up on this chart is that, well, these are kind

of a whole bunch of areas that anthropic is focused on.

There are so many industries that I think get very little attention and people think very

little of that this just hasn't touched at all.

And in it will, I think AI will eventually go everywhere, especially when you start thinking

about humanoid robots.

But things right now like installation and repair, transportation, production, construction,

like there is just just miniscule specs of AI doing anything in these industries right

now.

And so I think those are areas we should all expect to see grow a lot.

If you're interested in building new products, if you're interested in launching new products,

I mean, I would look at these charts to kind of decide what areas are good to go after.

I mean, if you're trying to start a new AI company architecture and engineering where

the models are very capable and the usage is very low and you literally would have to

compete with a custom GPT, like this seems kind of like a no brainer.

So anyways, as far as servicing new customers, I think these charts are also pretty interesting

for looking at that.

Yeah.

Well, that's an entrepreneur talking out of Jaden, right?

So Jaden just says, you know, next week Jaden's going to have like, you know, Jaden's

architecture AI or something like that, even for like just everybody out there, even

if you're not like an entrepreneur, even if you're not like the, you know, mad scientist,

like Jaden, you know, the idea is there's just so much room to grow here.

I, you know, I was just talking to the college kids about this, they're like, what should

we go into?

I was like, listen, it doesn't matter what you go into.

What matters is when you get into a company that you are able to sort of like figure out

how AI can, you know, can really sort of like change processes and don't worry about just

like if you're using AI, nobody cares if you're using AI.

People care if you can help them change a process or workflow and if you can elevate your

team.

Don't get like, don't focus on you getting great at AI.

The most important thing is can you sort of like elevate everybody else and just this chart

to me shows the huge opportunities out there guys, and this is the thing, right?

Like you have so many opportunities.

We're so grateful that you guys like chime in here and get in here and leave us comments,

ratings reviews.

We love ratings and reviews.

We're so grateful for them.

Also, please don't forget to check out AIbox.ai Jaden's company, which allows you to check

out all these models and compare and contrast them all together, which is what I do all the

time.

Anyway, build your own workflows, build your own apps, all this kind of like really cool stuff,

put stuff on the marketplace.

Jaden was just sort of showing how architecture AI blew up.

Guys, you can build this stuff on AIbox.ai.

If it feels intimidating, this is where you go AI box AI.

Next time you see us, here are a voice, you are going to be a multi millionaire because

of this absolute guaranteed results with an asterisk guaranteed result or not guaranteed.

But guys, do check it out.

Once I get super fun way to kind of get started on this, so check it out and we will see

you in the next episode.