Latest / 1 to 100 / E57: The $100M SaaS Company Nobody Wants to Buy

Transcript

E57: The $100M SaaS Company Nobody Wants to Buy · 1 to 100

0:00
I don't know what to do with traditional software companies
0:02
anymore. I think that this is the way we
0:05
can talk about it. Yeah.
0:07
I mean, I think the big, I think the big topic right now is, and
0:12
I literally just was grabbing coffee with a friend before we
0:16
started recording. But I remember 3 years ago,
0:18
like, you know, I'm as delusional as the next founder,
0:22
OK? I am full of fucking founder.
0:25
OK, I'm full of it, but I'm not stupid.
0:29
Yeah, and the problem with being not stupid is that the only
0:34
companies going public right now are going public that are
0:39
absolute monsters that are not figuratively defying gravity,
0:46
literally defying gravity. They are going to space.
0:49
OK. Like that's who's going public.
0:53
So, so like the, the question is like, I don't what happens to
1:00
traditional software companies. I don't, I don't know what
1:04
happens anymore. I don't know what happens like
1:07
if you're, if you're running a $200 million software, by the
1:11
way, I have like tons of friends in this category.
1:14
They're running between a 200 and $800 million software
1:17
company at 2010 to 20% EBITDA, growing 30% year on year, 120%
1:25
net retention. So like adding a few, like
1:27
adding a little bit of new customer, where do you sell that
1:30
business? Yeah.
1:32
Who's the buyer? Dude, it's so interesting too
1:34
because like so in the past you would look at, I don't know,
1:38
let's say you're being really conservative and you want to
1:39
look at a, an EBITDA multiple. So you take a business like that
1:45
that's really well established. Sure.
1:46
Like a 25 X EBITDA multiple, yeah.
1:48
Yeah, exactly. And so you look at that and
1:50
you're like, OK, there's this business that does, I'll just
1:52
use 100 million. So it's easy math, 100 million a
1:55
year in revenue, 10% margin or as 10% EBITDA on that.
2:00
And so you're talking about 10 million a year in profit and
2:04
you're going to put a 2025 X multiple on that.
2:07
So you're going to sell it for 2:50.
2:08
If you don't believe that that is going to continue to grow
2:12
over time, which I think there's a lot of questions about that,
2:15
maybe you can just gut that business in terms of like from a
2:19
cost profile standpoint, maintain a lot of that revenue,
2:22
maybe even grow it a little bit. And you'd bump that to say like
2:25
a 50% profit margin. In that context, it still takes
2:30
you five years to get your money back.
2:33
And then you've got to believe that it's going to be at least
2:38
stable beyond that time frame. Otherwise it's not worth buying.
2:42
So that's, that's a, it's a really interesting problem.
2:45
I think that's why we see like such a massive shift in the
2:48
valuations of software businesses.
2:51
But even even with that shift, like if the worst case scenario
2:55
were to happen and, and people just like stop start cancelling
2:57
and mass and everything goes to AI, there's no value in those
3:01
businesses. So.
3:03
I'm just even wondering at any price who buys them?
3:06
Let's just say Jeremiah, you could buy this company at 100
3:09
million, OK, 100 million in revenue, 10% EBITDA.
3:13
So you could just kind of like look at it and be like, I'm
3:18
going to snipe this thing for 100 million bucks, right?
3:21
Because you're like a friend down the road has like the
3:24
ability to give you that $100 million to go buy that business.
3:28
What do you go tell that friend? First of all, do you want to buy
3:30
it? Yeah, I mean, and then and.
3:33
Then what the hell do you tell your friend who's going to give
3:34
you the money? The price has to be low enough,
3:36
right? Like if it if it were, let's
3:41
well, OK, so like let's let's take that example.
3:43
Let's say a business has been is $100 million a year business,
3:46
it's been operating at 10% EBITDA margin.
3:48
If I was looking to buy that business, I would say I don't
3:51
believe that business is going to grow over time.
3:54
Is it possible? Sure, but I wouldn't, I wouldn't
3:56
believe in growth. There's some sort of trajectory,
3:59
right. So like where's it at in this
4:00
trajectory hasn't been going up because if it's been going up,
4:05
does it's going to turn the other direction?
4:07
So there's some, some amount of time that it takes for that to
4:10
happen, right? And I don't know what that time
4:12
frame is. There's a momentum piece.
4:13
So you've got, if you have positive momentum, it has to go
4:17
negative before you start to get into the revenue decline.
4:20
And I, my belief would be that there's a, there's probably two
4:23
to four years for that to happen.
4:25
Likely in this current context, maybe that changes.
4:27
But I think that's like that's probably the window I would look
4:30
at if it's positive currently, if it's already negative, then
4:33
who knows. And then at that point it's
4:35
like, OK, well, why is it only a 10% margin?
4:38
Could you boost that to 50607080IN software?
4:43
Those kinds of numbers are possible if you have the right
4:45
cost structure. But if you, if I was looking at
4:48
that business, I would say, OK, could you actually get the cost
4:50
structure into such a place where you think you could pay
4:53
that off in two years and then everything after that is
4:56
positive? I'm not an investor in this
4:58
context, so I, I can't say like these are the right numbers.
5:00
But if it were me putting my money into something that's kind
5:03
of the way I would look at it. It's like, do I think I'd get my
5:05
money back out of this in a couple of years?
5:07
And if not, then I don't think I'd be interested.
5:10
Yeah. I just think the problem is that
5:12
people I can't think of who wants to buy these businesses.
5:15
That's the actual problem. I mean I.
5:17
Would it used to be the public markets?
5:19
What? I would, I mean, it's probably
5:21
PE, right? Like somebody who thinks they
5:23
can change the cost profile business.
5:25
What do you mean dude? Have you seen what's happening
5:27
in PE for software? Did you see what happened in
5:30
Medallia? I did not see what happened in
5:31
Medallia, but I, I don't know I it ultimately comes. $5 billion
5:36
write off dude. PE is not supposed to put up
5:39
zeros. Yeah.
5:40
OK. But that's they bought it under
5:42
different assumptions. I think if you buy it now,
5:44
you're buying it with a different expectation of what
5:46
the future looks like. OK, but you still have OK, but
5:50
your cut, but your PE fund exists right as it exists today.
5:54
So you can say anything you want to about the new assumptions
5:56
about the new companies that you buy.
5:58
You can't like magically change the whole world.
6:00
This is this is like the whole problem.
6:02
The whole problem is like the PE fund has, this has the
6:05
traditional PE fund who buy software companies has their
6:08
existing portfolio. They've taken on a bunch of
6:10
debt. That debt was what's the
6:13
interest rate? The interest rate doesn't seem
6:14
to be going down a whole lot. Leverage is expensive.
6:17
So you're going to now lever against a thing that you're not
6:20
sure about how long it's going to grow?
6:22
That's true. I think your only real shot is
6:24
like if you're raising money now with a different thesis at a
6:28
different expectation and even that you have to be very
6:31
aggressive on what you think the type, the payback window is on
6:33
that money. The interesting though is like,
6:35
I actually still think that a well run business, so like a
6:39
monday.com, like they're, they're down what 80% or
6:42
something like that from last fall.
6:44
I, I shouldn't, I don't even know if they're well run.
6:46
All I know is the headline stock price, but like I'm just going
6:49
to use them as an example. They're deeply embedded into
6:52
businesses already. They have good revenues.
6:56
My understanding was that that isn't actually revenue based at
6:59
all. It's just based off of like
7:00
market expectations and all that, that decline in a context
7:05
like that, That business can probably figure out, if it's
7:09
well run, can figure out a way to change the business enough
7:13
that it can continue to generate revenue into the future.
7:16
It probably has to look like a different product, different,
7:19
like lots of things have to change within that business.
7:22
But it doesn't mean that business is going to go away
7:24
necessarily. And so I still think there's a
7:26
lot of these businesses that will survive.
7:27
The problem is which ones, and nobody really knows the answer
7:30
to that. This is a good point though,
7:32
like because one of the things I guess that's still an advantage
7:36
for, you know, traditional software businesses is the
7:38
revenue quality could be good still, right?
7:41
We're churning, right? We're churning, but we're
7:43
scrappy and we're small and we're trying to do stuff like
7:46
more efficiently. But like, you know, if you're
7:48
deeply integrated into the businesses like Monday, Monday,
7:51
so deeply integrated, like click up with Slack, you know, Google
7:55
Chat exists. I think it's better than Slack.
7:58
No one's churning from Slack to Google Chat as free, Google
8:01
Chats free. And it's like so well integrated
8:03
with Google Suite. Everyone's using Google Suite
8:04
for spreadsheets and Google Drive and this and that is the
8:06
one charm from Slack because they're just so built into Slack
8:09
and Slack doesn't even Slack just chat.
8:11
It's just instant message. It's not like you know, Monday
8:14
or click up where everything is in Monday or click up.
8:16
I don't know. I think that what do you guys
8:17
think? Do you think the traditional
8:19
software can be a little less scared just because the revenue
8:23
quality of software businesses is so good?
8:26
No, no, I think this should be scared shitless.
8:31
I I kind of a joke, dude. What kind of an I mean, it's
8:35
like you don't know me. We come on this podcast every
8:38
week. Right.
8:41
No, I, I also agree, like I, I mean, for what it's worth, like
8:46
I, I see, I see people who run software businesses that seem to
8:49
have zero concern about what's going on right now.
8:53
You need to think about. Everyone.
8:54
Dude. Dude.
8:55
Jack. OK, I'm look.
8:57
Sorry to cut you off, but like, if you think about everyone
8:59
who's come on this podcast who runs a traditional software
9:02
business, right, all of them are kind of still OK and doing very
9:06
well. Like I would say actually not
9:08
just still OK, they're doing very well.
9:09
And maybe this is just before the apocalypse, but like almost
9:12
everyone we've spoken to is doing fine, right?
9:16
And trying to shape their businesses, still trying to
9:18
shape it, still trying to think, you know, how do I make my
9:20
product have AI features, have native AI products, books or
9:23
actually function within codecs or applaud.
9:26
But they they haven't done it yet and they still are growing,
9:29
right? But this one, right, we're
9:30
talking about small people, right?
9:32
I imagine they're experiencing higher and higher rates of
9:34
churn. I don't know what, what the,
9:38
what the runway looks like before it starts to get painful.
9:42
But like, my understanding is that our friends are OK, right?
9:45
I don't think our friends are OK.
9:50
Let me tell you that what they say on a podcast is not the same
9:54
as what they say on text. Let me, I'll just say that
9:57
because I'm not trying to like say anything about any
9:59
particular person. But I first of all, I promise
10:02
you that the second thing I'll say is, look, I'm not.
10:06
This is like not a dig at Replo in particular, given that like
10:10
some people who like in the SMB world who use us or who
10:15
historically used us will compare us to them.
10:18
Did you see Cody's tweet that said like just to use Claude?
10:21
Yeah, yeah, yeah. I was just pulling it up.
10:23
That's the correct answer. Just use Claude, right?
10:27
That's the correct answer. OK, now again, Cody's ahead of
10:33
the curve, but that's the right answer.
10:35
So it'll take some period of time for everyone else to catch
10:37
up. What does a company let's just
10:41
pretend to someone else because I don't want anyone to think
10:43
like I'm purposely pitting it like I'll by the way, anyone who
10:47
uses us as just a page builder, which at this point should be 0
10:50
people, but anyone who uses us in that context should also not
10:53
do that. So I just want to be clear that
10:58
I'm not like trying to make like a like a dig point.
11:00
I'm actually trying to make like a strategic point or or take a
11:05
let's actually use someone else because that way it disavows the
11:09
particularities Wix, what is Wix trading at right now on a
11:13
revenue multiple basis? I don't.
11:15
Know it isn't like less than one like .7 or something.
11:18
I saw something and they also have record revenues, right?
11:22
That's on increasing revenues. And Jack, they have the AI
11:27
version of Wix, they have a base 44 and it's at $140 million run
11:31
rate and went from zero to 140 in two years.
11:34
So they already have the AI version of it.
11:35
Internally, they're executing their executional animals
11:41
essentially, right? Trading at less than one XY
11:46
because of this thing like by the way.
11:49
Even base 44 isn't future proof exactly.
11:54
Dude, we just rebuilt our website on Claude.
11:57
Like Fermat. Fermat's website is being
11:59
rebuilt on Claude by someone who has never coded in their life.
12:04
I'm embarrassed about how long this took me to pull up.
12:08
We got there eventually. Yeah.
12:09
No, it's crazy, guys. I mean, the amount that you can
12:12
do with Claude and so quickly as well.
12:14
I mean, I remember I showed you guys the landing page that I
12:19
built with Claude. That's like old news already.
12:21
Like the landing page did well. And you know, our brand is
12:23
growing and we're tiny, but that's like whatever.
12:26
But yeah, I've been playing around with how you can use
12:31
cloud to do landing pages even even better and than what we
12:35
saw, and it's just incredible what you can do.
12:38
Let me see if I can quickly show you this.
12:42
I used to be a Rep low shell, right?
12:44
I used to be like, Replo's amazing.
12:46
You know, I really liked the appeal to authority, right?
12:48
Oh, you can replicate these high converting landing pages from
12:51
other brands and then obviously, you know, adapt them to suit
12:54
your brand because it's assuming that those teams that like, I
12:57
don't know, grooms and Lollipop or whatever it happens to be
13:01
have the best agencies and the best marketers and they've
13:04
tested, you know, ruthlessly. They have a pretty good accurate
13:07
picture of what the market is buying right now.
13:10
And then you can go and replicate that.
13:11
That's not really what Replo does, but they have like a
13:13
template library and whatnot. I'm not sure what the product
13:15
looks like today. This was like 2 years ago, but
13:18
now you can just do that include, right?
13:20
You can. And it's you don't have to be
13:21
smart. You can just like ask for it.
13:23
You can say, hey, go take this Grunes landing page from that
13:27
like highest impression ad in Facebook ad library and remake
13:30
it for me. Let me see if I can show you
13:32
this quickly because I did exactly that and I asked it to
13:36
make it very customizable so that I could change it and make
13:41
it suit like whatever I wanted to use it for.
13:45
And look at this. So this is in Shopify and all of
13:48
these are Shopify sections so I can edit them.
13:52
I can be like, OK, this, I don't know, the CTA can change the
13:57
banner here to say, I don't know, never sold one of our
14:02
gummies and I don't know, make it red.
14:06
So it's not just like dumping raw code onto the Shopify store.
14:10
It's fully customizable to match the brand that's replicating the
14:13
page for their purposes. So I, I don't, I just, I don't
14:16
know why anyone would use red upload now you know.
14:20
It's, it's interesting because when I started getting into
14:23
coding, it's probably like 2010 ish, 9:10 somewhere around
14:29
there. And like the way that you did
14:34
things if you didn't know what you're doing or actually, even
14:37
if you didn't know what you're doing is you'd go to, to a forum
14:41
and you would copy a code snippet and you would go in at
14:44
it. It was a template, right?
14:45
That's, that's ultimately what it was.
14:47
It's like you would get this template and you would modify
14:48
the template to do what you need it to do.
14:50
My first business in selling software was WordPress
14:57
extensions or Woo Commerce extensions, which is basically,
15:00
it was basically a template. It was code, a file of code that
15:05
you would go and you would install on your website and then
15:06
you can make some modifications to it.
15:09
And that was the that was the way that we shipped things for a
15:13
long time. And basically you needed
15:15
somebody who understood the code well enough and understood the
15:19
logic and stuff behind it to be able to like take a template and
15:22
make it better. But Gen.
15:24
AI is so much better at that because ultimately, like it has
15:28
all of this context, all of this knowledge, and it can just give
15:30
you the perfect, maybe not perfect, but close, as close to
15:34
perfect of a template to start from as you could get.
15:37
It's just, it's a way, way better version of like in the
15:41
past, I would go and manually pull these different things and
15:44
take me a day to figure out how to do this thing.
15:47
Like building a landing page like that literally would have
15:49
taken me like pre Shopify, pre all that kind of stuff.
15:52
It's like a day of work if you know what you're doing to go do
15:55
that. And I'm copying code snippets I
15:57
have saved from this landing page I did in the past.
16:00
Like when I worked in an agency for a while, that was what we
16:02
did too, right? It was basically just like you
16:04
build up this library of sections and you go and you add
16:07
those code sections and you modify them to meet the needs.
16:10
And Jenna is just so much better at that than any other thing has
16:14
ever been. I mean, if you put the appeal to
16:18
authority side to the side, just the time to product is so much
16:24
faster, right? You don't need people.
16:26
But then if you reintroduce the appeal to authority, it's
16:30
something like grooms. Grooms is they are not like a
16:32
legacy player that has like came on to the space and you know,
16:37
done incredibly well. They came on so fast.
16:39
They came on so aggressively. They're expert operators, right
16:42
and they disrupted right. I think within months they had
16:45
like AG1 posting ads that said groans alternative, right.
16:49
So they just completely disrupted so fast and they've
16:52
done so incredibly well, but even then, they're spending so
16:55
much money on landing page experts on web designers or
16:58
whatever happens to be. And some of those landing pages
17:01
don't become their top converting landing pages.
17:03
They don't carry the brand through revenue ranges, right?
17:06
But if you're able to identify exactly which landing pages are
17:08
doing really, really well and then replicate them in minutes,
17:11
right? It's just it, it does what
17:14
Shopify does before it lets people compete on almost a level
17:19
playing ground. You now don't need the army of
17:21
experts testing different pages, right?
17:23
The experts agreements, for example, don't launch hundreds
17:25
of pages and most of them won't get very much performance, but
17:28
they're still the best in class, right.
17:30
But those, those top pages, now anyone can have access to them,
17:34
right? So, you know, kind of the the
17:35
promise for Shopify making website design possible in the
17:40
1st place. I think it's incredible.
17:41
I'm so excited. I think, you know, but I also, I
17:44
do worry about software businesses like Replo.
17:47
Do they survive just because they're so integrated or do
17:49
people genuinely just go, I don't need this anymore, I got
17:52
clawed right? I mean, it's no, it's no
17:55
different than it's always been in software.
17:57
It's just that this is accelerating, which is like you,
18:00
you have the ability to maybe no difference in the right phrase,
18:05
but there's always been disruptors and startups that
18:09
come in and they do something different, better, faster,
18:12
cheaper, whatever. And that's your, that's your
18:14
little piece of competitive advantage.
18:16
And you go after the incumbent and eventually the incumbent
18:19
falls off. And then that whatever that
18:21
startup is becomes the new incumbent.
18:23
And then somebody comes and takes their spot eventually.
18:25
And AI is just doing that. I, I just wonder there's still
18:30
always a, there's a path to incumbent existing.
18:33
And we should have been curious what you think about this.
18:35
Like, like a sales force exam, for example, is not a piece of
18:39
software I like to use, but it still exists and it has a very
18:43
big place in the, the ecosystem, even alongside all these
18:47
incumbents that are the just represent that have come up over
18:51
the years. And I'm curious if like we'll,
18:54
if you think we'll still see some of that in the future where
18:56
there are players that exist that have been around for 10:20,
18:59
30-40 years because there's a segment of the market whose
19:05
needs they meet. And then everything else is kind
19:07
of like these AI, either AI powered startups or just holy AI
19:12
or whatever that is. OK, I mean, I think like 2.
19:16
So just really quickly we can like tie like tie a bow on this.
19:21
Like will people continue to use these other things?
19:24
I, I just struggle to see how that happens over some
19:30
reasonable period of time. I think like the existence of an
19:33
incumbent has less to do with like people will continue to use
19:38
the thing the way that it has previously existed and more to
19:41
do with that incumbent actually changing how they operate.
19:45
And like Salesforce was one of the first to like, of the larger
19:49
companies to like become headless and become, I mean,
19:54
dude, they became MCP first before, before House, right?
19:58
Like Olivia is going to come on. Like the House is a startup and
20:02
House announced their MCP like 2 days ago and Salesforce
20:06
announced that they're going to be headless like 2 months ago.
20:10
So you know who's moving faster, right?
20:14
Like I I think that we shouldn't just assume that because you're
20:18
the incumbent, you're slow. There are incumbents who are
20:22
really fast and who are going to be winners because they're
20:26
shifting to what the new reality is.
20:28
So that's like one comment. The other one is, I think that
20:32
there are certain workflows that are much more likely to like
20:38
fade away slowly, but they're still going to fade away.
20:42
And so then it comes back to this, like, strategic question,
20:47
which is like, what do you do with the business?
20:50
Like, do you just run it while it slowly fades away?
20:53
Do you try to sell it? Who buys it ultimately?
20:58
Like, we as humans have to like, operate this business or we have
21:03
to not, and we have to find something else that we decide is
21:06
better for us to do. Yeah.
21:08
So I don't know. I think that that's a much
21:09
harder question. It's just like, will humans
21:12
bother with businesses that are not eventually going to make it,
21:16
basically. I get, I guess or one thing that
21:21
seeing all these posts from these incumbents that are moving
21:25
very quickly, I'm realizing is that the things that you guys
21:28
have been sharing are literally the future of these fast moving
21:32
incumbents, right? You were saying this two weeks
21:34
ago, right? Flattened organizational
21:36
structure, fewer layers, faster decisions, right?
21:40
And now it seems like pretty much all of the incredible, like
21:45
you were saying before, the businesses that are going public
21:47
that are, you know, going to space, they're all doing it.
21:50
They're all doing exactly all of these things, right?
21:54
I think it's super interesting to see it play out, but it's
21:56
also interesting to see the things that, you know, we
21:58
haven't talked about in some of some of these announcements,
22:02
right? The one in here that I think is
22:06
really interesting from Coinbase, from Brian Armstrong
22:08
at Coinbase is this idea and I saw Jensen from talking about
22:15
this as well, is the idea of AI teams, right?
22:19
And in just one individual that runs an AI team.
22:22
I can't. I'm not even sure if it's in
22:26
this post actually. Yeah, there we go.
22:28
We'll be concentrating around AI native talent who can manage
22:30
fleets of agents to drive outsized impact.
22:33
We'll also be experimenting with reduced pod sizes, including one
22:36
person teams with engineers, designers and product managers
22:40
all in one role. I mean, part of this seems, you
22:45
know, almost optimistic, very, very optimistic.
22:46
But also, I, I mean, it's happening, it's already
22:50
happening every, you know, a lot of startups have these people
22:52
in, in their teams. And I imagine that this is, you
22:56
know, a very real role in Coinbase today.
22:58
I mean, it's interesting, there are people out there with these
23:02
skill sets, right? Like my, a couple of my startups
23:05
that I've had in the past were me just like writing the code,
23:09
talking to customers, being the PM, like all of that stuff.
23:13
I literally did every single function in the business.
23:15
I was a solo operator who knew how to do all these things.
23:18
If I had AII would have moved so fast because I knew how to do
23:25
all of those things. And now it's like I just, I
23:29
managed basically. So it's a little bit different.
23:31
But in that context, I think if you could take somebody who has
23:36
a skill sets and just to say go for it, like here's your job, go
23:38
figure out how to do this thing. They can move so fast now.
23:42
Yeah, Jack, just so you know, we're doing this right.
23:46
By the way, the crazy thing organizationally that happens
23:49
when you do this, that is a little bit missing, but it
23:52
actually was, it started to get talked about in the click up
23:57
note that got sent around yesterday.
24:00
It's going to do it's going to do a few other things.
24:03
So the first thing it's going to do is you have to start to teach
24:07
your team how to think about career advancement in the
24:09
context of impact, where that impact is driven by their
24:12
ability to manage like a larger and larger fleet of agents with
24:17
higher complexity and higher context.
24:19
So it's like you're just going to grow in context and
24:21
complexity and you're going to be able to have agents that do
24:23
the work for you. OK, so fine.
24:26
So that's that's like one, one thing as opposed to managing
24:28
people, which is like traditionally how we told people
24:32
that they would advance in their career.
24:34
The second thing is that you have to now compensate on that.
24:38
And so we are going to like, I mean, it doesn't matter if my
24:41
team hears this because we're going to do it.
24:43
So the we're going to like change our comp structure such
24:47
that it's going to over reward people who are able to do that
24:50
most effectively, both in cash and in equity.
24:55
And so the outcomes that people like the way people will think
24:59
about their job, their impact, the their growth, their
25:04
compensation growth will be directly related to their
25:07
ability to drive impact where they are.
25:09
Because yeah, it kind of turns out you can do it.
25:12
I'll, I'll give like a micro example.
25:14
Oh, go ahead, go ahead. I.
25:15
Was just going to ask, how do you have a, have you already
25:20
thought through like the the levers and mechanisms of of that
25:23
comp increase? Like is it?
25:26
What are the the inputs that determine whether comp is goes
25:30
up in that context? I think that you in this
25:33
situation can literally pick business objectives.
25:38
You can just say people who move the needle on our business
25:42
objectives directly on their own will get overcompensated.
25:51
And I think you can literally say that because unlike in old
25:55
times when you would say like, oh, you're responsible for this,
25:58
like sub KPI that ladders up to this overall company goal, you
26:01
could literally just say like, no, no, no, I'm trying to get
26:06
I'm going to make up something, right?
26:07
I'm trying to get $10 million customers.
26:10
And someone can just say, OK, great.
26:12
I'm going to go find a way to get $10 million customers,
26:18
right? Or it's like, hey, we need like
26:22
a product that accomplishes X and someone could just say,
26:25
like, I'm just going to do the whole thing.
26:27
Nice. Yeah, Yeah, Yeah.
26:27
I mean this is already the the way that sales roles work or
26:30
Commission based sales roles, right.
26:32
So why shouldn't it be proportionate to impact in the
26:35
business, right? No, it's not the way Commission
26:38
based sales rules work. This is the issue.
26:39
See Commission based salespeople.
26:41
What they do is they say like oh I need marketing to do this.
26:43
I need someone to build this and I said I need someone to no, no.
26:48
But they all still say like, oh, but product me, I need this from
26:50
product, I need this from marketing, I need this like they
26:53
all still say I need someone else to do something.
26:56
And what I'm saying is, if you don't say that and instead you
27:00
just do it, you will be overcompensated.
27:03
Right. Because now you could just do
27:04
it. Do you see what I'm saying?
27:06
Like the new, the nuance matters, Yeah.
27:10
Or like an engineer could just say like, hey, I'm just going to
27:13
build a product and like, send it to a friend of mine and like,
27:17
see if he'll pay me $1,000,000 for it.
27:19
It's like I'm going to over reward that, you know?
27:23
Incredible. Yeah.
27:25
I want to see what you guys think about this.
27:28
I mean, this kind of ties into what Brian Armstrong was saying.
27:32
I actually don't know if there's any difference.
27:34
So we'll watch the video again and see if if he says anything
27:37
different in this video. Potential to ultimate many
27:41
aspects of human labor very soon.
27:43
I'm curious. That's not mine.
27:44
The degree to which you feel a responsibility for making that
27:48
transition. Go Well, What AI will do is to
27:49
make tasks that we do in our job more efficient.
27:52
Our job is not to wrangle a spreadsheet.
27:54
Our job is not to type into a keyboard.
27:55
Our job is generally more meaningful than that.
27:56
I'm fairly confident that AI will drive productivity, revenue
27:58
growth, and therefore more hiring.
27:59
There's a belief that the world's GDP is somehow limited
28:01
at $100 trillion. Well, what's likely to happen is
28:03
AI is going to cause $100 trillion to come. 200 trillion
28:05
dollars, $300 trillion. There's no fundamental limit to
28:08
the size of GDP. And so, and the number of people
28:10
in the world that is participating in this GDP, 9
28:12
trillion GP, GP is actually quite small.
28:13
And So what if we brought the rest of the world, the rest of
28:15
society, the rest of the world's population, empower them with
28:17
AI, give them the opportunity to participate in GDP.
28:19
It causes GDP to control. I think that's very likely.
28:21
Yeah. And it's very likely that
28:23
companies will become more prosperous.
28:24
But it's for certain. This is for certain.
28:26
Everybody's jobs will change as a result of AI.
28:27
Some jobs will disappear. New jobs are created.
28:29
Everybody will have to use AI because if you don't use AI you
28:30
lose a new jobs somebody. I just remembered that this is
28:32
extremely uncontroversial. There's nothing to talk about
28:35
here. That's the wrong clip.
28:38
Let's see here. Maybe it's this one and.
28:42
They are smart and they're saying, Oh my God, you know,
28:45
they're going to be no more radiologists because you know, I
28:48
can read X-rays better than radiologists can.
28:50
They're going to be no more software engineers because I can
28:51
program better than software engineer can.
28:53
These people are wrong. So what's really going to happen
28:54
is it's going to elevate all these people and it's like, it's
28:56
like you've been digging. Let's say you're a software
28:58
engineer, right? They're not going to give you is
28:59
you've been digging out a basement of your house with a
29:02
shovel and somebody's about to handle a bulldozer.
29:05
You should be so happy if you're digging the base with your house
29:06
and somebody says, hey, how about this?
29:07
I have a tool here that's gonna and what's really gonna happen
29:10
is we're gonna have so much productivity in our economy
29:12
that, for example, this is this is 1 effect.
29:14
A lot of people who have two earner income households.
29:16
One of the people is gonna drop out of the workforce.
29:18
That's why we're gonna have a labor shortage people because of
29:19
the productivity gains. You're gonna be able to afford
29:21
things. We're gonna have I critical have
29:22
to have deflation of certain core.
29:24
Assuming we let this technology play out and don't you know,
29:26
hamstring it with regulations too early, we will actually
29:28
have, you know, everything. New food will get cheaper.
29:30
OK. So I don't know if this one's
29:31
very controversial either. To me, these are both like
29:35
obvious. I'm not sure if there's much
29:37
that we can expand on these. Yeah, it's pretty high level.
29:42
Yeah, I mean, I think the the only comment I would make is
29:46
that what Jensen's point is, is like you will lose your job to
29:49
someone who uses AI is not fully appreciated by people.
29:56
I don't see the usage of AI with the level of urgency that one
30:01
would expect if you were actually worried that your job
30:04
was going to be taken, right. Not even close.
30:09
Right. I mean like I think.
30:11
Why, Yeah. Which is like a wild thing,
30:13
right? From an attitude point of view,
30:15
if like somebody on your team heard, wait, I can be
30:19
compensated more for having more of an impact on the business
30:22
goals of format, right? And they felt like they were
30:26
getting left behind. They felt fear or you know,
30:29
whatever emotions they feel, but they don't actually act towards
30:32
it. And I'd say that that's the
30:33
wrong inclination. It's an opportunity or an
30:36
opportunity to have more impact to dive in.
30:38
You were talking about your executive assistant trying to
30:42
find ways to make all of her job done by core.
30:45
That's incredible. That's the sort of approach that
30:47
you want. And I don't know if you can
30:48
change, you know, hopefully you can, right?
30:51
Hopefully you can take very talented people and give them
30:53
the right information that puts them on the right track, right?
30:56
So they they aren't wasted by some sort of like weird
31:00
political persuasion that like, oh, we should burn data centers
31:04
down. Or, you know, AI is this massive
31:06
red light and this massive problem.
31:07
And they actually become more powerful and become more
31:11
effective and achieve the business goals that their roles
31:14
are built for, right? Just, you know, changing over
31:17
time. But yeah, I'd be, I'd be
31:20
interested to hear from you guys what, what's work at your
31:22
businesses that getting those people, sometimes the, I don't
31:26
want to call them Luddites, but you're kind of getting those
31:29
people who are talented but are apprehensive or they're slow.
31:33
They're not early adopters, right?
31:35
And they, they need that encouragement.
31:36
What's what's worked for you guys and bringing those talented
31:39
people into being AI first? You got to show by doing because
31:44
it's not like any of us knew how to do this a week ago, right?
31:48
Did I tell you about this like bot that I built to give
31:51
feedback to my team? Yeah.
31:52
Tell you guys about this. OK.
31:55
Yeah. The the the only comment I would
31:57
make is like, so now other people are building similar
32:00
things. It's only a handful of other
32:02
people. But if you don't do it yourself
32:05
and show people how to do it, then like, why would anybody
32:08
else think that that's the ex. Like, you can't say I expect you
32:11
to do it and then not uphold yourself to that same behavior
32:16
because the issue is that it's not like you did it three years
32:20
ago and now you're expecting someone else to do it because it
32:22
didn't exist three years ago. So you have to do it at the same
32:25
time as everyone else is doing it, which is the new part of
32:29
this. But you do have to do it and
32:31
show people like you have to model, basically, yeah.
32:35
Yeah, I agree. It's it's all about modeling.
32:37
I think, like the and the interesting thing, that example
32:40
of your bot, right? You're modeling one thing there
32:45
and that one thing is probably not a good engineer on your team
32:49
probably doesn't see a whole lot of utility of doing that for
32:51
themselves. They've got other things they're
32:53
working out right. So for them to see the impact
32:56
that they might like, that's kind of cool, but like not not
32:59
really useful to me. They need to see something
33:01
that's useful to them as well. So there's all that.
33:02
It's not just like modeling one thing.
33:04
It's modeling lots of different things, lots of different people
33:07
modeling things. And I think that's how you get
33:10
that get people to understand and to be able to engage better.
33:14
So, and that's true for us too, right?
33:15
Like the, the things that a lot of the things that I do with AI
33:20
are things I've seen someone else do first.
33:22
They're not things that I did first.
33:24
I'm just seeing an example on saying, oh, that's cool.
33:27
I think I can go apply that over here.
33:29
And then I go use that thing. Yeah, the only thing that we're
33:33
trying to potentially teach people is the behavior, right?
33:37
Which is like, hey, don't be afraid to just do it and don't
33:41
assume that it won't work right. Like this is where the like same
33:46
founder delusion is useful because it's like we're kind of
33:50
not of most founders don't start a thing that is completely
33:54
unknown without high comfort with like if I I'm just going to
33:58
do it and see what happens, right.
34:00
And so that's like lucky that like, yeah, we don't know what's
34:05
going to happen because we haven't done this before by
34:07
definition. So we're just but, but, but
34:09
that's fine, right? It's like, not a big deal.
34:12
Real quick on the macro stuff, I think one thing that's really
34:14
going to be really interesting is how the, do people need jobs
34:18
in the future? I, I, I like to work.
34:21
I think I, I want a job. But it is interesting, like over
34:25
time, I mean, we, we all used to be farmers basically, right?
34:28
200 years ago, you spent the numbers I was just looking at
34:33
like call it 90% of your work went to shelter and food.
34:39
And now for the typical person, it's more like 50%, which is
34:42
still pretty high actually, but it's half of what it was 200
34:46
years ago. And AI has the potential to
34:48
maybe cut that in half again. And so there's like this.
34:52
And not only that, but with the technology advancements, that
34:55
money that's going towards non food and shelter is so much more
35:00
impactful of things you're getting with that are so much
35:04
greater than they were 200 years ago.
35:05
I mean, what were you getting 200 years ago?
35:07
Like maybe you're buying like pen and paper and you're right,
35:10
buying some books and things like that.
35:11
Now we've got these machines, they do incredible things.
35:15
So I think they'll just be like a like, I could see overall
35:18
wealth increasing and maybe like the, the deflation thing is an
35:22
interesting concept of what, what does that look like over
35:26
time? I don't know if that'll actually
35:27
happen. I think, I think both of them
35:29
are pretty optimistic on the future.
35:32
I think that's probably a more pessimistic views that maybe
35:35
warrant some consideration at least.
35:38
But but I do think that there is like a very optimistic path
35:42
potentially where if we just look at the history of how
35:45
technology has impacted just our our lives in a positive way,
35:50
it's been largely positive over time.
35:54
The the, the last thing I'll say is like, I was in Cincinnati
35:57
last week. Did I tell you guys about this
35:58
being in Cincinnati? Yeah.
36:01
And it's like very different. Yeah.
36:04
I'm like, OK, so that happened. And then this week my someone
36:10
who I work with here was visiting from New York for the
36:14
week, right? So I've been like asking folks
36:17
from our New York office to come here.
36:20
And I think like I've been trying to communicate to them
36:22
that like even the difference between New York and SF is
36:25
pretty big in terms of like AI adoption.
36:28
And by the way, like our New York team is like pretty
36:30
adoptive on AI and like under like it's pretty up to date on
36:33
how things are going. And we, we host this thing
36:36
called like Claude and cocktails and Claude where we help other
36:39
companies. Sorry, other people in New York
36:42
like get up to speed. I'm like how to be effective
36:44
with AI and things like this. But she spent a week here with
36:50
people here and her ex colleagues here.
36:53
And she was like, OK, I get it now.
36:56
Now I understand why you're being such, I guess, a pain in
37:00
the ass, such a pain in the ass about like coming to SF because
37:05
it is even more different than I expected.
37:08
And I think that that part of it is, yeah, not fully appreciated
37:18
yet. I think.
37:20
I think like the pace with which SF is moving is really hard to
37:26
understand, really hard to understand unless you're here
37:30
and, and like you see how like a whole city almost is moving at
37:37
the pace of AI. But then you go anywhere else in
37:40
the country or in the world and it is not the case that the
37:43
whole city is moving at the pace of AI.
37:47
And that is a very big difference in in like how things
37:52
will change. Yeah.
37:56
And I think it occurs an onus on like the people who are in it
38:01
too, as quickly as possible get other people in it too.
38:05
Like I think that that job is significantly underappreciated.
38:09
So there's these like announcements.
38:10
You just see the like deployment company by Open AI and like all
38:15
like all these forward deployed engineers getting hired all over
38:18
the place and the like a deal that Anthropic did to create
38:21
like a consulting arm. Are you guys seeing all of these
38:23
things by any chance? There's all these services
38:27
businesses that are being stood up by Open AI Anthropic like in
38:30
order to like help enterprises adopt AI faster.
38:35
The shortage of service offerings for people to adopt AI
38:38
is so massive. I think it is extraordinarily
38:43
under discussed. And I actually think that one of
38:46
the most important jobs for people who are in AI and
38:48
understand AI is to make sure as many other people get on the
38:56
boat as possible as quickly as possible.
38:59
It's interesting though. It was like my San Francisco, I
39:04
just think back to when I started working in software,
39:09
SAS, whatever you want to call it 15 years ago, basically like
39:14
San Francisco was the center of that.
39:16
Then the pace there was different than the concentration
39:20
of people who I feel like that shifted a little bit with COVID
39:22
and remote work. But at least at that time, San
39:25
Francisco felt different than the rest of the world in that
39:28
regard. Is this not just the same thing
39:31
now in a in a new type of a new wave of innovation?
39:37
Or do you think it's different than it was with the way the
39:41
world looked 15 years ago and San Francisco's place and
39:44
technology at that point in time?
39:46
In that way of framing it, it's the same, but not unlike the
39:52
other analogies that have been drawn.
39:54
It is simply the case that the velocity and magnitude is like 4
40:00
to. 20X higher depending on which number you look at because
40:05
like the fastest growing software companies at the time
40:07
did not add $14 billion of revenue in a month.
40:10
That's fair. So like again, depending on the
40:13
number you look at, you know what I mean?
40:17
Like I don't know how to answer that.
40:19
It's an interesting time and super, but the way that I see it
40:23
is a super abundance of supply to all sorts of different
40:27
demands, right. That like typically, I guess,
40:30
wouldn't be, you know, your go to for to be software software.
40:36
I remember you, you spoke about this before, Jeremiah.
40:39
You know, you're now doing things that it didn't make sense
40:41
to hire someone for leverage to go and do.
40:44
But now it makes sense because it's like what it costs a few
40:47
seconds of my time to get it done right.
40:49
And I just think that, you know, from from big things to small
40:53
things across so many different, you know, existing demand, We
40:57
just have a super abundance of supply.
40:59
So like, you know, air or information, right, We don't
41:03
have a Library of Alexandria and we don't have to buy books,
41:06
right? There is no scarcity of
41:08
information. We have, you know, extreme
41:11
competition because of so much as such super abundant supply of
41:15
information to provide us with that information.
41:18
Right before it was, you know, Google and Google search was
41:20
like how people, you know, found some information.
41:23
You, you know, go and checked a few pages to find that
41:25
information. I mean, before it was books and
41:27
then before it like knowledge was extremely scarce, right?
41:30
But now, you know, all these companies are just competing to
41:33
get you that information as quickly as possible, not just
41:37
and get the information, but execute as well, right?
41:39
It's not, hey, find me a hotel in this city and a restaurant
41:43
and, you know, and then I'll make these reservations with my
41:45
assistant or my wife will do it for me.
41:47
It's hey, go do it all for me, right?
41:49
I'm going here, go do it all for me.
41:51
And that's like a, you know, I guess more recreational than
41:54
like a, you know, a, you know, a businesses demand, right?
41:58
For, for a solution. But yeah, that's the way I, I of
42:02
course it's deflationary. It couldn't be more
42:04
deflationary, right? Yeah, I don't know.
42:07
I think then it will actually be deflationary because we have
42:11
currency debasement as well. If it's measured in any,
42:14
anything else, anything hard like gold, right, then yeah,
42:17
we've experienced a huge amount of deflation, all right.
42:19
Things are cheaper in gold times and hard asset times than they
42:23
ever have been, right? Yeah, yeah, that's for sure
42:26
true. I'll leave you with one idea and
42:29
then I really do got to go. But the one idea I'll leave you
42:31
with is think about visiting San Francisco right now, like going
42:36
to a different country to experience a different culture.
42:39
And that's sort of like what I'm encouraging people to do, which
42:41
sounds like a crazy thing to say, but it is worth spending,
42:46
like, one or two weeks here and being like, what the hell is in
42:49
the air here? Because I'm telling you, it is
42:52
different. Yeah.
42:54
Yeah. And I didn't fully appreciate
42:55
that until, like, yeah, people from new.
42:59
You wouldn't think New York is like doesn't get AI, but yeah,
43:06
it's just, it's just different, right?
43:07
And So what you're talking about, Jack, with like, all of
43:09
these impacts and like the abundance of it, I like you
43:13
should come here to feel it because it's like, OK, yeah, you
43:15
can eat pizza, but then you can go to Italy and eat pizza.
43:20
Yeah, good point. Yeah.
43:22
Yeah, alrighty guys, I really do.
43:26
Got to go to see you soon. Bye.