00:00:00
if I had to guess about what has changed
00:00:03
from 2021 to 2025
00:00:06
is that most companies have realized
00:00:09
that buying yet another
00:00:13
vertical software solution is not going
00:00:15
to help their business that it typically
00:00:18
adds bloat it adds cost and it adds
00:00:22
people and I think starting in 2023
00:00:26
what people started to guess is at some
00:00:29
point in the near future you're going to
00:00:32
have some AI way of rewriting all of
00:00:36
this vertical software and I think
00:00:38
that's why it stopped growing i don't
00:00:40
think this SAS market ever had the
00:00:44
return on equity that it was supposed to
00:00:47
and I think so many companies have woken
00:00:49
up from this hangover saying there's got
00:00:51
to be a better way it can't always be
00:00:54
yet another tool yet another program yet
00:00:57
another multi-year delay
00:01:00
yet another price escalator and I think
00:01:03
that that the jig is totally up for
00:01:06
software you're referring to the
00:01:08
Salesforce and the SAS category Chimoth
00:01:11
and what you're doing at 8090
00:01:13
specifically yeah well it's it's not
00:01:15
just us but like if you look at anybody
00:01:16
that's rebuilding software
00:01:19
it is so much easier to rebuild software
00:01:22
from scratch today like my team of 30
00:01:25
people can transact hundreds of millions
00:01:29
of dollars of work not because we are so
00:01:32
prolifically amazing but frankly because
00:01:34
well I think the team is good but
00:01:36
honestly because the underlying tool
00:01:37
chain gives you a level of leverage and
00:01:40
so if you rebuild the software
00:01:42
development life cycle using these tools
00:01:46
you can't help it but become much more
00:01:48
efficient and you can't help it but
00:01:50
deliver custom solutions that are
00:01:52
meaningfully cheaper and I think Jason
00:01:54
if you look at the entirety of the
00:01:56
software that runs the world we're going
00:01:59
to rebuild it soup to nuts all of that
00:02:02
and the tool you're referring to just
00:02:03
for the audience is the AI co-pilots
00:02:06
that are making that are contributing 30
00:02:08
40% to code bases at Microsoft and less
00:02:11
less specifically that because those are
00:02:13
those are good for individual people but
00:02:15
the software development life cycle is
00:02:17
more the horizontal end to end of making
00:02:19
things got it so what we do internally
00:02:22
at 8090 is we have an entire process
00:02:24
that starts from the PRD all the way out
00:02:25
to the functioning code and we use
00:02:28
different techniques at each step but
00:02:30
what you get is a 50 60 70% increase at
00:02:34
each step which then compounds and so
00:02:37
you have the ability of a team that
00:02:39
would otherwise be able to service tens
00:02:41
of millions of dollars be a team that
00:02:43
can service hundreds of millions and
00:02:45
then a team that would otherwise service
00:02:46
hundreds can service billions let me ask
00:02:48
you guys your response to this theory if
00:02:52
there is going to be this kind of
00:02:54
accelerated
00:02:55
call it custom software rebuild of
00:02:58
business models and you take the S&P 493
00:03:02
do you think that we enter an era where
00:03:04
there is a similar dispersion as we're
00:03:07
talking about seeing in the MAG 7 with
00:03:09
the S&P 493 where there are going to be
00:03:12
probably the biggest money-making
00:03:14
opportunities for investors that we've
00:03:16
seen in decades
00:03:19
between those that do adopt and do
00:03:21
rebuild using AI and those that don't
00:03:25
for 100 100%
00:03:27
i had a call yesterday with one of the
00:03:29
largest private equity funds in the
00:03:31
world hundreds of billions of dollars
00:03:33
under management and we're doing
00:03:34
something with them at 8090 with one of
00:03:37
their most important assets
00:03:40
and when you're an owner of a business
00:03:43
and you can direct
00:03:46
specific change
00:03:48
and you can rip out
00:03:51
hundreds of millions of dollars of
00:03:54
software licenses and replace it with
00:03:57
tens of millions of dollars of highly
00:04:00
customized software
00:04:03
it's an enormous lift to OPEX and
00:04:05
business model quality so why doesn't it
00:04:08
happen more the reason it doesn't happen
00:04:10
right now for this S&P 493 is that the
00:04:13
IT organizations inside all companies
00:04:17
essentially speak a different language
00:04:19
than the CEO the CFO and the board so if
00:04:21
the CEO CFO and the board of directors
00:04:24
of the S&P 493 speak English the IT
00:04:27
organization speaks Mandarin Chinese and
00:04:30
you get away with saying all kinds of
00:04:31
i'll give you an example i went
00:04:33
to a CIO conference one person that I
00:04:37
met an $18 billion a year IT budget
00:04:43
what the does that actually even
00:04:45
mean to spend $18 billion a year on it
00:04:50
i'm not saying that this is a mag seven
00:04:52
company guys and when you take that
00:04:54
example and you multiply it by 50 and
00:04:56
100 and 493 examples of people spending
00:05:00
money there's an entire cartel of
00:05:03
influence that's been built in software
00:05:05
that's going to get undone because
00:05:07
you're not going to be able to justify
00:05:08
it free absolutely correct and the
00:05:10
response from the SAS industry is
00:05:13
changing from the per seat model as the
00:05:16
number of employees at these companies
00:05:17
continues to get lowered obviously
00:05:19
Microsoft a lot of layoffs andy Jasse
00:05:22
talking about layoffs they're moving
00:05:24
from the per seat model they're not
00:05:25
taking this uh laying down uh they know
00:05:28
that people are going to make custom
00:05:29
software so what they're doing is
00:05:30
they're moving to a consumption model so
00:05:32
you're seeing people charge per call per
00:05:35
customer support call etc
00:05:38
and well it's I'll tell you why it
00:05:39
doesn't work it's working in combination
00:05:41
hold on hold on let me finish the other
00:05:42
thing they're doing is they're
00:05:43
dramatically lowering the number of
00:05:45
people and the developers they have on
00:05:46
their team and then a lot of what's
00:05:48
happening in the background is the third
00:05:49
piece they're doing is they're starting
00:05:51
to uh do rollups and people are starting
00:05:54
to talk about how can we take you know
00:05:56
20 of these SAS companies lower them
00:05:58
just like you're doing to compete your
00:06:01
thought
00:06:02
playbook well I just wanted to comment
00:06:04
on this like consumption based pricing
00:06:06
it doesn't work and what I mean is you
00:06:09
can have some adoption in the short term
00:06:11
the best example is Snowflake but in the
00:06:13
long term it destroys your business and
00:06:15
the reason is because you don't know
00:06:17
which data is valuable and you're not
00:06:19
going to put up with a variable business
00:06:21
model that increases more and more cost
00:06:23
because you need to trap everything and
00:06:25
so what happens is all of these other
00:06:27
companies develop around you people go
00:06:30
back to Postpress people go to Superbase
00:06:32
they find all of these ways of saying
00:06:34
snowflake makes no sense and the reason
00:06:37
is because in this world nobody's going
00:06:39
to pay consumption because you're like
00:06:41
how do you expect me to you know hold
00:06:43
and store and pay for terabytes and
00:06:46
terabytes potentially a day of data it's
00:06:49
not sustainable we'll see if intercom
00:06:51
Salesforce HubSpot and we see if all of
00:06:54
those people start Slack start losing
00:06:56
their customer base or if they lower
00:06:57
their pricing to make it just too easy
00:06:59
to keep those systems in thomas your
00:07:01
thoughts yeah so two quick thoughts uh
00:07:03
number one Chimath to put a kind of a
00:07:05
mathematical frame on this right we know
00:07:08
that Anthropic is kind of the level zero
00:07:11
of code generation they're they're doing
00:07:12
incredibly well powering companies like
00:07:14
Cursor right i think and this is order
00:07:17
of magnitude correct that Enthropic in
00:07:19
Q1 added 70% of the net new ARR in the
00:07:23
SAS industry right defined by public SAS
00:07:26
companies right so let's just think that
00:07:28
the company in AI that is most powering
00:07:31
the disruption of SAS added 3/4 of the
00:07:34
net new of the entire industry right so
00:07:36
that's kind of point number one I think
00:07:38
Freedberg point number two I think what
00:07:40
we're seeing in the Max 7 right where
00:07:42
we're starting to have debates about
00:07:43
who's well positioned and who isn't
00:07:44
who's going to win and who isn't right
00:07:47
is actually as it was in the past 5
00:07:49
years going to be a broader lens into
00:07:51
the S&P 493 I think inside of boardrooms
00:07:56
inside of every investment committee
00:07:58
you're going to see the exact same
00:07:59
conversations that we've been having
00:08:01
about the MAX 7 right who who's well
00:08:03
positioned who can win what are the
00:08:05
management teams maybe like Zuck that
00:08:07
are being aggressive and bold and
00:08:08
capturing the opportunity and which are
00:08:10
the ones that are not so for me as a
00:08:12
stock picker right I think over the next
00:08:15
five years
