Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Before we get into the show, I have
0:02
to tell you about the lineup coming to HubSpot's
0:04
Inbound Conference this year. You all
0:06
know it's one of my favorite events, and this
0:08
year you're going to hear from some amazing guests
0:10
like Reese Witherspoon, yes,
0:13
the Reese Witherspoon, Derek Jeter, and
0:15
Andrew Huberman. I love the Huberman Lab
0:18
and listen to Andrew all the time. Make
0:20
sure to mark your calendars for September
0:22
5th through the 8th and join us
0:24
right here in Boston. This year, there'll
0:26
be multiple stages, industry experts,
0:29
talk tracks on everything from AI to sales
0:31
strategy.
0:32
It's the perfect combination of inspiration
0:35
and practical tips to help you grow your business. Tickets
0:37
are selling out fast, so head over to inbound.com
0:41
to get yours today.
0:44
Today, we're going to talk about the seven principles
0:46
to make you a winner in the AI era. I am your host, Kit Bodner,
0:48
CMO at HubSpot, joined by my co-host, Kieran
0:51
Flanagan, the CMO at Zapier, and
0:52
this is Marketing Against the Grain. Let's
0:55
get into today's show. Kieran,
0:58
we spend a lot of time talking about the future.
1:01
I know you're a big fan of the future,
1:03
but what about the past? Let's
1:06
get into today's show. Kieran,
1:09
we spend a lot of time talking about the
1:11
future.
1:12
I often think that you and I are probably not content
1:14
in our real lives because we spend
1:16
all of our time talking about the future. And
1:19
then when we get to the future, we're like, ah, not that cool. What about the future
1:21
now? Not that cool. What about the future and
1:23
I? Kieran, you and I were talking on WhatsApp. There's
1:26
a lot of uncertainty in artificial intelligence,
1:28
and most of the companies, most of the executives
1:30
we talk to are like, hey, what do
1:32
you think is going to happen?
1:34
And so what do you think is going to happen to fill
1:36
in the blank with some specific thing? Right.
1:39
Call it search engines, call it advertising,
1:41
call it sales reps, BDRs, what
1:43
have you. Like, there's tons of questions there, right? And
1:46
I don't think you and I know the answers 100%,
1:49
but what I do think is going to be really
1:51
helpful to everybody is that we took the time to
1:53
think about what do we know to be
1:55
true about
1:57
artificial intelligence? And can we
1:59
distill that data? down into principles and
2:02
use those principles as a potential
2:04
predictor of the future. You
2:06
and I love the idea of first principles.
2:08
We're first principle obsessed. And so we
2:10
wanted to do some very AI specific
2:13
first principles on today's show.
2:15
You ready to go? You ready? Yeah. All
2:17
right. I got our list. I'm gonna start with one of my favorite
2:19
principles and I want you to talk about what you think this one
2:22
means. The first one I'm gonna start with is in
2:24
an AI world, anything that depends
2:26
on just
2:27
brute force, human work and
2:29
time is going to be disrupted. So if you
2:31
use humans to do manual work,
2:34
all that manual work will be disrupted by AI. AI
2:36
will do that work
2:38
and somebody will come along and do it
2:40
faster and cheaper than you're currently doing
2:42
it. Yeah. Or it's
2:44
specifically in knowledge work. Which is like typically
2:47
where you would think about this, you would think of manual labor. And
2:50
now I have seen examples where
2:52
there are new stories of potential AI
2:54
robots gonna be able to do manual labor things
2:57
like build houses and all these things in the future.
2:59
But this is like very much specific
3:02
to knowledge workers and anywhere where you've had
3:04
to
3:05
fill the gap, where I just like grinding
3:07
something out with humans, there's just gonna be disruption
3:09
there. Again, I think of everything in terms of,
3:12
it's not the job that's being disrupted, it's
3:14
the tasks. The tasks, that is the way to
3:16
think about it. I love that. And human jobs are a series
3:18
of tasks and in the future automation
3:21
is gonna be able to automate much more of those tasks. Perfect.
3:23
And if you were thinking about this first principle, what you wanna say
3:26
is,
3:26
what are the parts of my business that I irrationally
3:29
rely on human time and effort and
3:31
manual work to do? And how do I automate
3:34
those before somebody else automates them
3:36
and puts me out of business?
3:37
Yeah, I think the other thing that's unwritten
3:40
in here that we didn't call out that I just thought of as,
3:42
like within a technology company, if
3:45
you judge your success by the size
3:47
of your team, that is gonna be heavily disrupted.
3:49
Yes, you are screwed. Because I think you're gonna be
3:51
much more judged on the quantity
3:54
of your output. Right?
3:55
Like today, there was a really great article
3:57
from the Slack CEO. Now there was nothing in it that was...
4:00
That was like jaw dropping, but everyone
4:02
seemed to think it is jaw dropping, which is you hire people
4:04
and they want to hire other people. And we pay
4:06
people based upon scope of work and scope of
4:08
work today. There's nothing wrong with this, right? Like
4:10
I know that it's really everyone wants to paint.
4:13
There is actually something wrong with it. There's
4:15
nothing wrong with paying people by
4:17
scope of work, right? The amount
4:19
of surface error that they control or the importance
4:22
of that work. And historically,
4:24
the amount of people you had to do that work,
4:27
the larger the tasks you were given or the larger
4:29
the goal you were given, you just need
4:31
more people to do that task or to achieve that
4:33
goal. Yeah. Yeah. There was over hiring. The problem
4:36
with technology companies is you have empire builders
4:38
where people hire just because they actually think that
4:40
makes them look better if they have bigger teams. And even if they
4:42
don't need bigger teams, they're not incentivized
4:44
not to have bigger teams, but
4:46
I think in the future, within this first
4:48
principle, you are going to be judged in your
4:50
output and your output could be the same with
4:53
two people plus AI. As it historically has
4:55
been with like 20
4:56
people, right? And I think that's the magnitude
4:58
of difference there. Look, the reason that
5:01
the scope of work compensation model isn't
5:03
good is because especially
5:05
now with AI and the rise of
5:07
automation, growth has to be nonlinear.
5:10
And nonlinear means that you get more
5:12
out than you put in. It's not just like, Oh, I had
5:14
a person and I get like same amount of
5:16
return of spin for every person
5:18
I add, it's like, right? No, that's automation and my
5:21
revenue grows exponentially, not
5:23
linearly. And when you
5:25
compensate people
5:26
on scope of work, you are incentivizing
5:28
linear growth. That's the problem. That's
5:30
the problem to kind of put it plainly. All right.
5:33
I got favorite ones. I've kind of expanded
5:35
into being a three-parter. We've talked a little bit about
5:37
this on the show before, but I want to underscore for everybody,
5:40
there are three generations of the internet,
5:42
web one, oh, the birth of the internet.
5:44
It democratize access to information. The
5:47
second generation of the internet web 2.0
5:49
democratize access to each other. Connections.
5:52
It was the social web. Web 3.0
5:54
this generation, the AI generation
5:57
democratizes understanding
5:59
of.
5:59
So the first principle
6:02
here is the understanding of complex topics
6:04
is now commoditized, which means
6:07
you could have an idea for a song or have an idea
6:09
for a piece of art. And
6:11
before, if you didn't have the skills to write music
6:13
or read music or do all those things, you couldn't do anything with it.
6:16
And
6:17
now, AI can help you basically
6:19
take that idea and turn it into an
6:21
end product, and it is democratizing
6:23
the understanding, at least at
6:25
a basic level, of all of those skills. I
6:28
actually would argue that that is slightly
6:30
different than I were talking about it, which is... Ooh, I love this. Go.
6:34
Yeah, it's not the understanding of complex topics.
6:36
What you've actually described is you don't need to understand them at
6:38
all.
6:39
And it's actually expertise is commoditized
6:42
because in the past, you
6:44
would have to understand how to do something to get that
6:46
idea out into the wild. Today,
6:49
or in the future, you will not have to understand
6:51
how to do that thing because you can just tell someone to do
6:53
it for you or tell an AI assistant to do it for you and get that
6:55
idea out of it. So
6:58
if I have to understand a complex topic, I still
7:00
have to get the information and kind of understand that. Okay,
7:02
well, how do I create that thing? Or how do I do that thing?
7:04
Or how do I get from A to B? But
7:06
actually, what AI does is I don't
7:08
have to learn any of that stuff, right? I just have to tell
7:10
it to do it and it will do it for me. So I wonder if it's
7:13
like
7:13
expertise in a lot of areas is now
7:15
commoditized. Well, expertise is definitely commoditized.
7:18
I think those two things go part and parcel. I don't think they're
7:20
actually mutually exclusive expertise and understanding. I
7:22
think they're very related. But if you're thinking
7:24
about the future of your business, what you want
7:26
to say here is the uniqueness
7:28
of understanding and the barrier to
7:31
entries of skills is not a moat
7:33
that is going to protect my business from being disrupted
7:36
as much in the future as it was in the past. Right.
7:38
Right. And AI is going to
7:40
have a higher potential to disrupt on
7:42
that expertise and understanding.
7:44
Right. In the past three to five years,
7:47
if you were a coder, you could have spent your
7:49
weekends learning how to be better than the other coders who
7:51
are going to be better. You could have
7:53
had to move all the same experience with you, but you put in those extra
7:55
hours. You could outwork them. And today
7:57
that coder has closed the gap. because
8:00
they've used in these kind of co-pilots and tools.
8:02
And that's the commoditization of expertise. Are you
8:05
saying we're moving from an era about work to
8:07
an era about think?
8:08
That would actually be, if we are, that
8:10
would be great for the world. I think it would be amazing
8:12
for the world and that's what you're tipping at. I don't know
8:15
if it's true, but you're kind of saying, hey, maybe we're
8:17
heading there, right? Yeah,
8:18
yeah, maybe we can all have like work-life balance,
8:20
but people who have just the better ideas
8:23
and able to, are able to get those ideas and
8:25
into the wild now will actually be much more successful
8:28
versus the person who is just willing to work 24-7 for
8:31
seven days a week. Well, one of your principles that
8:33
you shared was right online with that, which is creative
8:36
ideas are priceless. Execution
8:38
of those ideas commoditized. Explain what that
8:40
means. Yeah, when I think of AI, there's
8:43
three different categories of creativity.
8:45
There's being able to mix ideas together. There's
8:48
being able to create ideas within a
8:50
framework that you give it. And there's being able to come up
8:52
with net you ideas.
8:53
And if you look at AI across, we've
8:55
covered this across like text, image, video. If
8:58
for the most part is limited in those two buckets, right?
9:00
It can actually mix ideas together. Like
9:03
in terms of text, you can give it some content
9:05
from a book or something and ask it to
9:07
like take your idea and take the content from the book
9:09
and splice them together and give you some output.
9:11
And that output is like fine. Like if you're an okay
9:14
writer, it's fine. It will actually get you
9:16
to good. If you're a great writer, it doesn't actually improve
9:19
anything for you.
9:20
Image is very similar. I can take things,
9:22
work within this existing style guide and give you something
9:24
back and video very similar,
9:26
but it cannot create brand new ideas,
9:28
right? If I'm a marketer and I'm sitting there with these text images
9:31
and video tools, I'm gonna use elements
9:33
from each of those different things in different ways. I
9:35
think images is by far the most ship
9:37
level ready. Like I'm actually gonna take things and put
9:39
them straight into my production line. Video,
9:42
if I'm really good at video, I can take things there and
9:44
actually use them within my marketing. And in
9:46
text, it actually is good for like a research
9:48
assistant and an assistant.
9:50
But none of those things are gonna create net
9:53
new ideas. That's gonna really blow up my
9:55
marketing. And so creative ideas
9:57
are still priceless. And I would say even more so.
9:59
Because people who have great ideas,
10:02
I think get held back or all their
10:05
time gets sucked away by the execution of
10:07
those ideas. Just hard to use all these tools
10:09
and do things to bring them into market. And so I think
10:11
ideas are going to go up in value because
10:14
the execution of those ideas, and we
10:16
pay people a lot of money to bring those ideas to
10:18
fruition, is going to get commoditized. I
10:20
love that one. Another one that you had that
10:23
might be my favorite, and I might be
10:25
very, very jealous. Oh, yeah. I
10:27
nailed it. Which one? I
10:30
nailed it. You just told me which one I did. I
10:32
think it's the last one. You had the cost to
10:34
take risks are near zero. Right.
10:36
And that is pretty brilliant. You're basically
10:39
saying that the old adage of like
10:41
fortune favors, the bold, is actually going
10:43
to get far more true because
10:45
the
10:46
cost to actually try something
10:48
and see if it works is dropping to,
10:51
if not zero, very, very close to zero,
10:53
right? Right. We covered it on an episode
10:56
recently. In a not too distant future,
10:58
you're going to be able to build a chatbot equivalent
11:00
power of chat TBT that
11:02
can run 8.5 billion queries
11:04
a day
11:06
comparable to what Google does today for $650.
11:09
Can I not take that $650 and do some
11:12
incredible things and try things out like
11:14
the minimal viable version? So
11:16
when you want to test something and you want to launch
11:18
something new, you do a minimal viable version.
11:20
So what is a minimal viable version? Well, I want to use the least
11:23
amount of resources and take on the least
11:25
amount of expense to prove an idea is valid.
11:28
Now, minimal viable versions for a lot
11:30
of things still cost resources,
11:32
still cost some amount of money, and they are
11:34
still a detractor from trying a lot of
11:36
different things and taking risks. The cost
11:38
of doing those things is going to depreciate
11:41
so, so fast that I'm going
11:43
to be able to try so many different things.
11:45
You're right that it's related to the first thing, which again,
11:48
people with great ideas can get more
11:50
of those ideas out into the world to see which ones
11:53
actually resonate. And I think they're the people
11:55
that are going to be much more successful and they are going to benefit
11:57
a lot from the abilities of AI.
13:55
And
14:00
some amazing, amazing recent
14:02
episodes, Jina, like you need to trust yourself.
14:04
And here's how to start. She has an amazing
14:07
guest. The Savi Kumar shares
14:09
her insights on how self-talk confidence
14:11
and self-trust are all interconnected. If
14:14
you're struggling with self-sabotage or
14:16
negative self-talk, check out this episode.
14:18
This is just one example of all the amazing
14:21
episodes that Jina is doing right now. I
14:23
highly recommend checking it out. It's a great
14:25
podcast. You can listen to Gold Digger
14:28
wherever you get your podcasts.
14:56
It's a great episode. I would recommend taking leadership to
14:58
technology, entrepreneurship, relationships,
15:01
and much, much more. If I could
15:03
recommend
15:04
one episode, I would really recommend the recent
15:06
episode on scams in the pharmaceutical industry. Or
15:09
there's really a great episode on how you play
15:11
the status game. Both are incredibly addictive, shocking,
15:14
and captivating stories. Listening
15:18
to the Smart People podcast feels like befriending the
15:20
most interesting person at a dinner
15:22
party. There's never a dull episode. I've
15:25
got a few more principles left. This is one
15:27
of
15:27
mine that I think is really important is that information
15:29
asymmetry no longer exists. So what
15:32
does that mean? It means in the olden days,
15:34
if you were a buyer, you would have to go to that
15:36
company. Normally we'd have to talk
15:38
to a salesperson or somebody at that company to figure
15:40
out pricing, use cases of that product for your business, specifically
15:44
all of those things, and with AI. That
15:47
is quickly going away. And that
15:50
all the interesting things that I mean, if I've
15:51
ever shown to you a company has is available
15:54
to you if you want it without having to talk to anybody
15:56
else.
15:59
And that leveling of the information
16:02
playing field gives the buyers
16:05
way more power than they have ever had. And
16:07
will force companies to think about enabling
16:10
their buyers much more than they think about
16:12
enabling their sellers. Like it's inbound.
16:15
It's just an evolution of the inbound. Like inbound was created
16:18
and like the original that
16:20
sprung from consumers having much more
16:22
power as brands were created. And more things
16:24
on the internet and the internet rise in
16:26
popularity and people being able to like get
16:28
all the information they needed themselves.
16:29
But that wasn't true
16:32
of all the information. Like there's just a lot
16:34
of information that companies have that you have to
16:36
talk to the salesperson, talk to the support
16:38
person, like get connected to a human. And
16:40
I think in this world you're right. Like
16:42
you have an ever present agent
16:44
that can give you the information around the company because
16:47
a company's best interest is to give that agent
16:49
all the information they can so people can query at 24 seven.
16:52
And so I think all of that stuff does not get
16:54
locked away. And you actually
16:56
consumers have more control than they've ever had before. AI
16:59
is much more of a tailwind
17:01
for buyers than it is for sellers. AI is
17:03
way, way, way better for buyers than
17:05
sellers. Right.
17:06
OK, I think we have time for like two more, Karen.
17:09
So one that we've got here is
17:12
consumer impatience and entitlement
17:14
will be at an all time high.
17:17
I love you put entitlement in there. Like
17:19
those entitled consumers. No,
17:21
but entitlement, I think, is a very
17:24
important word where it's like the best way
17:26
to describe expectations because
17:29
when you have endless
17:31
choice, you get
17:32
to be entitled. Right. Like
17:34
if they're saying that there is going to be more choice than
17:36
ever before, then a consumer gets
17:38
to be entitled. That is a privilege that that
17:40
is just born out of those market circumstances.
17:43
Right. Well, the impatience is because you never expect
17:46
to have to wait in line for an answer again. I
17:48
think that's going to look archaic. The times
17:50
where I have to sit on the phone with my broadband company
17:52
for an entire day to try to get an
17:55
answer to something.
17:56
None of that stuff exists in the future. Consumers
17:58
are going to expect to get answers. whenever they need
18:00
answers in any kind of form, whether that's chat,
18:03
email, WhatsApp, it doesn't really matter.
18:05
They expect to get 24-7 answers because an
18:07
agent can give them those answers. And the entitlement
18:10
to me is like, they expect the concierge
18:12
experience because they expect agents to be able to
18:14
do most of the work. So none of this stuff is an added
18:16
cost to the company. The company has
18:18
agents set up, the agents do all that work for them.
18:21
And everyone gets the concierge experience.
18:23
I just think that the thing I don't know about
18:25
is like,
18:26
where do you find leverage? That's what I
18:28
always continue to think about is like, where's the leverage? Where's the
18:31
leverage? Where's the leverage? If everyone has a concierge experience,
18:33
what is it? Although they drive it around to your office
18:36
and bringing you tea and cakes, considering
18:38
you dine and telling you you're a great person
18:41
and picking you up and telling you lots of great things about
18:43
yourself. I don't know what the next part of
18:45
one-to-many
18:46
is. If the one-to-many experience becomes like
18:48
a very like one-to-one type experience. I think
18:50
that's actually a very good point. And it happens to
18:52
be exceptionally true. I
18:54
don't know what the differentiation is, but I know
18:56
in the short term, there's going to be a lot
18:58
of differentiation made by companies
19:01
who facilitate, you know, that speed
19:03
and around the clock access
19:06
to information and service versus
19:08
the old school way of waiting. Right. And that's,
19:11
if you're predicting the future around
19:13
AI, that's one of the things that I think is going to happen.
19:15
All right. Our last principle
19:18
of the day is that chat user interfaces
19:20
are just as popular and important
19:23
as graphical user interfaces. People
19:26
are going to want to use natural language
19:28
to do things online in
19:31
the same way that they've historically used graphical
19:33
point and click interfaces to do
19:35
things online.
19:36
Is that true? I do agree with that. Every
19:38
brand is going to have a natural language layer and chat
19:40
is going to be part of every single product
19:43
experience. I don't think you're going to use a product in the future
19:45
that doesn't have some way to chat with that product. Like
19:47
me being able to tell the product through natural language, please
19:49
do this for me. Please do this for me.
19:51
I think all platforms build that into
19:53
their app. Again, when we think about the Sam Altman
19:56
roadmap episode that we just did, we covered
19:58
what was open AI, secret.
19:59
roadmap, product roadmap, and sound map
20:02
and talked about in there. Hey, the thing we think
20:04
we got wrong about the plugins, and
20:06
these plugins don't have product market fit, is people didn't
20:08
want to create their app in chat
20:11
GPT. People wanted to bring chat GPT
20:13
into their app. And that's an important distinction, because it's what
20:15
you're saying here is, every
20:18
product has a chat GPT natural
20:21
language interface component to
20:23
it. Correct. And graphical interfaces,
20:25
graphical UIs, I don't think they go away.
20:27
We've just seen the biggest technology brand
20:30
in the world launch vision pro
20:32
glasses, which bring graphical UIs
20:34
that you can control with your hands into your
20:36
like everyday presence,
20:38
right? You actually wear it on your face.
20:41
And so I think those two things actually are just
20:43
maneuvering in like very different
20:45
ways, but actually are going to be very important to how
20:47
we like consume content, use software
20:49
in the future. Yeah. That's why I said they're going
20:51
to be equally as important, right? I think
20:53
we're going to live in a world where graphical user interfaces
20:56
and text or chat based user interfaces
20:58
coincide with each other. Right. And they're
21:00
going to kind of relate and rub
21:02
off on each other, right? Like one of the things about a
21:05
chat user interface is you kind of go back and forth
21:07
and iterate on something. That same iteration,
21:09
I think will move over to a graphical
21:11
user interface in different ways, right?
21:13
Like they'll start to blend and start to merge.
21:16
And what this means for everybody in the future
21:18
of AI, what does it mean for our business? It means
21:20
that customers are going to want again,
21:23
a very adaptable way to interact
21:25
with your business. And that interaction
21:27
is going to look like a mix of like a visual
21:29
interaction and a text based interaction, right?
21:31
Right. And balancing that and figure out where one is
21:33
better than the other is going to present a bunch
21:36
of opportunity to you as a marketer, to you as a founder.
21:38
And so that is what we are
21:40
talking about
21:41
here today. Those are our principles
21:43
for predicting the future of AI because
21:46
AI is uncertain.
21:48
But if you have those principles, you can
21:50
understand directionally the impact AI
21:52
is going to have and then use them to
21:54
figure out what strategic choices
21:57
and investment choices you need to make in your own
21:59
business.
21:59
be prepared for whatever change
22:02
might work out in terms of like what specifically
22:04
might happen. The case of the future,
22:07
as long as you have a rough understanding of the direction,
22:09
I think, you're normally way ahead of everyone
22:11
else. You don't need to know the answer ahead of time. You
22:13
need to know kind of the direction the answer is going.
22:15
Do you agree with that, Kieran? I agree. I
22:18
think this maybe relates to my very
22:20
last first principle, which I
22:22
do think is one we should like just cite as
22:24
fast movers get unfairly rewarded,
22:27
which I think there's always been a reward
22:29
for adopting
22:31
things before the masses.
22:33
I think in terms of AI, it has
22:35
never been more prevalent. It has never been more
22:38
important. You will never be more rewarded
22:40
than with this technology to be in that
22:43
fast mover bucket. Because again, I think commoditization
22:45
happens much faster. And so when everyone catches
22:48
up, you lose some of that advantage. But that advantage
22:50
for the first movers in integrating AI
22:52
into your work, integrating AI into your go-to-market, integrating
22:55
AI into your marketing is so much bigger
22:58
than those who are not doing that. And there's
23:00
still like a big discrepancy between what
23:02
we talk about every
23:03
day and what the common most companies
23:05
are actually doing. Correct. And are still
23:08
like way, way behind in trying to figure out
23:10
how to integrate this into their business. If you were watching
23:12
the show, you were already ahead. Exactly. And
23:14
if you want to keep being ahead, hit that subscribe button. Because
23:16
we are obsessed with staying
23:18
ahead on this show. And we are obsessed
23:21
with giving you the tools and the insights
23:23
to actually be ahead. And
23:25
the principles today, I think are a great
23:27
example of us try to do that. Leave a
23:29
comment in the YouTube if
23:31
you've got a favorite principle or if you
23:33
have one of your own that you think we left
23:35
off that we should have included in today's show. That
23:38
being said, that's first principles
23:40
for AI. It was awesome. Thank
23:42
you so much for watching today. And we'll be back real soon
23:44
on Marketing InstaDrain.
24:03
Howdy folks, Jacob Cohen here,
24:05
one of the hosts of the Hustle Daily Show, another
24:07
awesome podcast from the HubSpot Podcast
24:09
Network. Our show's got fresh daily takes
24:11
for anyone interested in learning about everything
24:14
from the billion dollar business of nacho cheese
24:16
to what the heck is going on with the housing market
24:18
to how AI is changing the way we live
24:21
and work. It's short, it's sweet, it's
24:23
got good intro music so I'm not really sure what more
24:25
you need. We'd love to have you come check it out
24:27
and you can tune in throughout the week wherever. You
24:30
get your podcasts.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More