Latest / Identity at the Center / #421 - The AI Identity Control Plane with Henrique Teixeira

Transcript

#421 - The AI Identity Control Plane with Henrique Teixeira · Identity at the Center

0:04
This is identity at the center. Welcome to the Identity at the
0:11
Center podcast. I'm Jeff, and that's Jim.
0:12
Hey, Jim. Hey, Jeff, how are you?
0:15
Oh, not so bad yourself. Good.
0:17
But you confused me. We're doing the Identity at the
0:19
Center podcast. I thought we were doing AI at
0:22
the Center. I want to talk about AI today.
0:26
Oh, what a what a what a change in topic.
0:30
I feel like we have to talk about AI every time.
0:32
It's just it's in the consciousness.
0:33
I know I've seen the comments sometime about talking too much.
0:35
AII get it. But it is top of mind for a lot
0:40
of people. I think we are still figuring it
0:42
out. And as people look back on these
0:45
episodes as like a time capsule, they look back and see these
0:49
Neanderthals trying to figure out how to manage AI in identity
0:54
security. Yeah, it'll be like Beavis and
0:57
Butt Head, like that's AI. OK, so I got two things AI
1:04
specific. First is like a pet peeve of
1:07
mine, which is that I still feel like we're trying to solve
1:12
everything using old IT Like, OK, so I'm working on this
1:17
project right now. It's called Project X 41123.
1:23
In other words, I'm not going to get into the specifics can.
1:26
I can I just say you should be using AI to come up with a
1:28
better project name. OK, please continue.
1:30
Definitely, definitely. And so it's based on like, hey,
1:35
one of the things is this can search for data and I'm like,
1:42
OK, so we go in there and you Google search.
1:44
Why don't we just put AI that way?
1:47
You can tell it like exactly what you want, give it all kinds
1:50
of specifics rather than just using search terms, and it
1:54
scours the data looking for the search terms and just returns
1:58
things in that way. And here's the real kicker,
2:01
right, is this particular application is searching for a
2:07
bunch of information, training courses about AI.
2:12
So I'm like, OK, but we're using like old IT to find this stuff
2:16
about AI. Doesn't that doesn't that seem
2:19
weird? So that was number 1, I think #2
2:24
was that I realized that. OK, so I had this idea like,
2:30
let's forget doing this Google search thing.
2:33
Let's just create an AI agent that goes out and scours all
2:38
these websites and comes back with, all right, well, here's
2:43
all the information that you need, right?
2:46
And then it like analyzes it and curates it and spits it back
2:50
just the data that you need. Well, the problem get this was
2:54
an identity management issue. It was like, well, the AI
2:57
doesn't have access to all those applications.
3:00
A lot of that data is gated, so you can't just go and scrape all
3:04
this stuff. You have to get access, you have
3:06
to get an account. And it's not like there's no way
3:09
to use your account in real time because the activity is going
3:13
out and scraping a bunch of applications, putting them into
3:16
the database, and then them becoming part of the overall
3:19
data set. So essentially it's like I'm
3:22
starting into this identity management to show us like, OK,
3:25
this is where the road brings the road.
3:27
Well the answer is just give that agent you know God access.
3:31
Let it do everything right. Forget privileged access, just
3:33
make it an admin and then all your problems are solved.
3:36
They can see and hear and do everything.
3:38
Well, that's what I was figuring too, but then I thought, OK,
3:41
when I go in and put the request in for that God access account,
3:46
I can just imagine the people laughing.
3:49
Coming from an identity security consultant.
3:52
Yeah, exactly. Like you're not getting that.
3:54
So try again. But that's the problem is like,
3:57
OK, without all these systems already connected and these were
4:01
websites that people built and they were gated by, you know, a
4:05
single sign on system, but they weren't going to use my single
4:08
sign on account. It's got to have the AIS got to
4:11
have the agents got to have some sort of credential that it can
4:16
go and even scrape this website. Not talking about like going and
4:20
hitting his back end. And what I learned is that a lot
4:24
of websites like create a web page that the agent can go and
4:31
scan this web page. So it's not going through and
4:34
like browsing the site. It's more or less hitting like
4:37
an API that the data is kind of flattened.
4:40
And I'm like, dude, I'm learning all this stuff because I'm
4:43
actually doing it in real life. And it feels like I'm in the
4:47
beginning of my identity management journey again.
4:50
That's the best part about it. I, I am is Just when you think
4:53
you got it all figured out, here comes something new to totally
4:56
rock your boat and put you back into the newbie corner.
4:59
Oh yeah, newbie corner. 6 1/2 years running an identity
5:02
management podcast. I'm learning some of this stuff.
5:05
But that you're right, spot on. Learning is the best part about
5:10
the job. It's actually fun, but it's it's
5:13
frustrating sometimes while you're doing it, but then you
5:15
come out the other end and you're like, oh, I'm an expert
5:17
in this now. Well.
5:19
That's how it works, right? Is you gain experience, you
5:21
learn about it, you teach yourself, teach others.
5:24
Once you can teach others, I think that's when you've got a
5:26
level of mastery that I think is, is something to hang a hat
5:30
on because teaching others is hard.
5:32
It's it's one thing to understand it, it's another to
5:34
be able to explain it to someone else and have them understand
5:37
it. Yeah, and I feel like a lot of
5:40
these concepts I knew in the abstract was like, now like I'm
5:45
going through this project and really dealing with it in real
5:50
time and having to solve these issues.
5:52
And so it's real learning, right?
5:55
Like you can, you can learn through books, definitely learn
5:59
the conferences and but the best teacher is real life.
6:04
True. And I love that Segway because
6:06
we can talk about conferences, the ones that we're going to be
6:07
at here pretty soon. We've got EIC coming up in May
6:11
and looking forward to seeing you there as well as our guests.
6:14
That will introduce in a second. That is in Berlin, May 19th.
6:18
The 22nd discount code gets you 25% off.
6:21
So people are taking advantage of of that.
6:24
It's ID AC-25 MKO. You can also find it on our
6:27
website, idacpodcast.com. Just Scroll down and got that
6:30
there. And then we'll be at ideniverse
6:32
in June. So that's June 15th to the 18th.
6:35
We actually have a 30% discount code, which I believe is the
6:38
best discount code you will find.
6:39
It's like 5% better than others. So if you use the code IDV 26-ID
6:45
A C30 percent, you get 30% off. Again, in our show, in our show
6:50
notes as well as on IDAC podcast.com.
6:53
That's there. I know some people have taken
6:55
advantage of that. Definitely appreciate it.
6:56
That kind of when you use those codes that helps show that, you
6:59
know, Jim and I can bring people to a conference and that then,
7:03
you know, allow us to conference say, oh, you know what, hey, we
7:05
should have that podcast go there and do some live episodes
7:07
and things like that. So it's greatly appreciated when
7:10
we see that kind of support out there.
7:11
So definitely appreciate the folks who take advantage of
7:14
those codes and hopefully they save some money and then they
7:16
can use that on identity beers that will be taking place at
7:19
both. You want to talk real quickly
7:21
about identity beers and then we'll.
7:22
Sure, I'm not the official promoter of Identity Beer, but
7:27
they are community events. Kind of the way it works is
7:31
these are not vendor sponsored events.
7:33
So vendors will be there, practitioners will be there, but
7:38
it's everybody just getting into it and talking and having a good
7:41
time. You don't have to feel like
7:44
you're beholden to the fact that, oh, XYZ company is paying
7:48
for my beer and my snacks. Pay for those things yourself.
7:53
But the payoff is that you know, you're there as like an
7:56
independent person. So I'm a big fan of identity or
8:00
I think it brings out people that I might not otherwise meet.
8:04
And I love the conversations. So it's like it's all about
8:08
having real world identity conversations.
8:11
Yeah, it's community and sometimes the conversations
8:13
aren't identity at all. It's just whatever.
8:15
And we're just meeting people and making new friends.
8:17
And I'm looking forward to it too.
8:19
So we'll have that in Berlin and also Las Vegas.
8:22
So I guess keep an eye out on Identa dot beer at Espenbaga's
8:26
website. So he kind of keeps track of
8:28
that kind of stuff. But you know, LinkedIn is also
8:30
good spot where you can kind of find some of that information.
8:32
Let's get to our guests, one of my favorite people.
8:35
He's also senior vice president of strategy with Sabian.
8:39
Definitely a member of the Identirati.
8:41
Henrique Teixeira welcome back for the 4th time on the podcast.
8:46
Fourth times already. Hey, Jeff.
8:49
Hey, Jim, what a pleasure. Thank you for having me again.
8:53
Yeah, like I said, you're one of my favorite people to talk to.
8:55
I think we just have so many good off the record
8:58
conversations. When I say off the record, I
9:00
mean we're not recording this. This episode might be a little
9:04
bit of insight to some of those conversations because we've been
9:06
kind of talking AI and not even maybe like the identity
9:10
management components of it, but sort of like societal things
9:13
that have been happening with AI.
9:14
So we might get into that towards the end, but it's
9:17
identity center. So let's try to maybe start with
9:19
some IAM things. But it's been about a year since
9:24
you were last on. I'm going to see you in Berlin,
9:27
I guess, very quickly. You're kind of in and out while
9:29
we're there. But give me an update.
9:30
What's been going on for the last year in your world?
9:33
Very interesting and especially hearing you guys talking about
9:36
identity here. I don't know when this episode's
9:39
going to go up, but next week actually I'm I'm, I'm going to
9:42
be flying to my hometown in Brazil and we are inaugurating
9:48
the chapter of it. Any beer in Latin America?
9:50
The first one there in Sao Paulo's.
9:53
Super cool. Also, the IAM tech day is a is
9:57
an interesting conference. It's organized by Alfredo Santos
10:02
and they have a few other locations, but the one in Brazil
10:05
was the one that initiated everything.
10:07
I'm, I'm speaking there as well. So super excited about that.
10:10
And, and like you said, the, the community aspect of identity
10:14
beer and in specific regime is what I like about it.
10:18
This is the connections we make and the opportunities that we
10:22
create for ourselves and other people that are interested to
10:25
talk about identity or to know more about this journey of this
10:29
profession or yes, no new people.
10:33
And I think I think this is on itself is a very, very
10:36
interesting benefit just for having those those chapters
10:39
around the globe and and pretty cool about that.
10:42
And Jim, very interesting choice of a topic for the opening.
10:46
AII don't think we'd talk about AI enough.
10:48
So thank you for that. Yeah, I think probably if you're
10:53
doing a shot for each time we say AI in this episode, you are
10:57
three sheets to the wind by now. Yeah, you're not going to make
11:02
it. Pace yourself.
11:04
Right, it's it's an interesting one, but thank you, thank you 4
11:08
times. Wow, I can't believe it, but
11:10
I'll be back anytime. Well, I appreciate that.
11:13
So by the time this episode goes live, I think it'll probably be
11:15
a week before EIC happens. And the one thing I kind of want
11:19
to note there is we call this thing identivier, but it's
11:21
really open for everybody. I think I'm pretty much on
11:24
record at this point. I'm not much of A drinker, but I
11:27
go, maybe I have a, you know, half of 1, maybe get a Coke or
11:31
something, but really. Alcoholic beverages, yeah, I, I,
11:33
I, I like to emphasize that every new chapter said, oh, non
11:36
alcoholic beverages are also welcome.
11:39
So we, we accommodate everybody. I think that has to be that way.
11:45
So I don't know how much you like with the intro, if you're
11:48
kind of thinking, man, I can't believe Jim runs an identity
11:52
podcast or what, But everybody's talking about this digitic AI.
11:57
And I know you've been talking about it a lot for the last
12:00
couple years. And you know, recently I've seen
12:03
a lot of things where you're talking about this AI control
12:07
plane for identity or AI identity control plane.
12:12
Can you talk a little bit about what that is?
12:15
Yeah, and I think Jim, everybody's trying to put AI in
12:18
a box with things like the project Glasswing that was
12:24
launched not too long ago. I think it was April 7th when
12:28
Anthropic announced this thing. With the hype and the amount of
12:34
marketing around this thing, Yeah, I would be scared as as
12:38
well, right about this thing. So when we think of the the
12:42
control plane, I think it's an intent of the industry to OK, we
12:46
got to, we got to control these things.
12:49
Like you, you said you you're learning everyday.
12:52
I'm learning everyday. And, and when I'm, I'm talking
12:55
with Gardner, my, my, my, my former colleagues, now I, I pay
13:00
them right? I to get advice, but also I, I
13:03
talk with, with Coppidreco, I talk with even the smaller firms
13:06
in this industry. There are so many different
13:10
terminologies. I, I know people used to
13:13
complain about me about the acronyms.
13:15
So there are new acronyms to me app and for example, if you, if
13:20
you go at I think the way Gardner's calling, those are
13:24
agent management platforms. But then if you talk with other
13:28
firms like the one by Francis Odom, right, Francis is running
13:32
the security on the software analyst Saab research.
13:37
I'm going to say that name, right?
13:39
So Francis is calling that UAD PS unify the Gentech defense
13:44
platform. So it's more on a defense and,
13:47
and, and detection type of thing.
13:50
So there's a lot of different names for this type of things.
13:53
I, I think people and our industry, they're still trying
13:57
to figure it out. What are the different things?
14:00
And, and I think who said something, I think was Simon,
14:06
Simon Moffat also, of course, writing amazing research on, on,
14:10
on the topic and saying he said something about identity
14:14
security, which is, and I think it's related to this, the way he
14:19
described it is because he, he maps a lot of vendors in the
14:23
identity security space. Well, I, I, I've mapped more
14:25
than 70 vendors out there calling out identity security
14:30
and they're not competing with themselves.
14:32
They're doing different things. So the same way that we hear
14:38
those different definitions and different products called, oh,
14:43
this is the control plane, this is a platform, whatever.
14:47
They may not be competing with themselves either.
14:50
I have that hypothesis because if you look at the way Gardner's
14:53
looking at it's more on, on the, they have the security element
14:57
of it. But then they talk about
14:59
libraries, tooling, dashboards, the observability like deploying
15:04
agents and desktops and this kind of stuff, right.
15:07
The operating and orchestration of that, the one that is called
15:14
UADP is about it's in the way they define it always.
15:19
DSPM plus DLP plus ITDR plus NHI as well.
15:23
Enough. Acronyms, but it's a little bit
15:27
of the DLP area, right the data security with identity security
15:30
together. So it's completely different
15:33
from the definition that Garner was also working with.
15:36
So I don't think they're competing and I wonder what
15:41
buyers would be thinking on that.
15:43
So saving themselves, right? So we have launched the the
15:48
control plane for identity as well, which is very identity
15:53
centric because that's that's the focus of the company.
15:56
And I think we're going to see more of that people perhaps over
16:00
using this. Same with identity, sorry, not
16:02
identity, but agentic or AI governance.
16:07
Governance is a word that has been around since the days of
16:10
forever for GRC people and governance, risk and compliance.
16:15
And what is governance of agents?
16:18
Is it about the ownership is about the other things?
16:21
So, so Jim, I think what struck me as interesting is that
16:28
there's a lot of different definitions.
16:30
They're not necessarily competing terminologies, but
16:35
they're definitely confusing. I think, you know, I don't want
16:40
to sound negative here, right? But I'm trying to figure out
16:44
what are the capabilities of an AI control plane, because you
16:49
mentioned identity security and it's like, OK, 70 companies say
16:54
they're doing identity security. If you ask them what are the
16:57
capabilities of an identity security tool, it would be like,
17:00
well, it's all the things that we do.
17:03
I did that research exactly for some of the identity Security is
17:07
a super powerful privilege access management tool plus
17:09
something else. Or it's going to be the next
17:12
generation IGA plus something else.
17:13
You're right. Yeah, it's, it's, it's
17:17
authorization, it's authentication, it's privilege,
17:20
access management, identity governments.
17:23
And I just write it off a few and I think those are things
17:26
that probably need to be there. Do you have a framework in your
17:31
mind of what the AI control plane needs to include?
17:37
Yeah, if you're looking for an identity problem statement,
17:40
which is well, this broadcast about identity, it has to be
17:45
always I, I, I like the idea of, of the, the three, right, the,
17:49
the rule of three. So if we, if we think of that as
17:52
3 pillars to that platform, there'll be the one piece of it
17:56
is discovery, the other pieces about the, the governance and
18:00
the life cycle of that. And the other one is the
18:02
containment control enforcement. You could also say, wow, this is
18:07
discovery, identity management, life cycle, access management.
18:12
I think those are universal things that we've been doing for
18:17
a lot of other things in identity.
18:19
And I think they apply very well for the challenge we see today
18:24
in energetic risks, right? When we think about the identity
18:27
risks or identity security, I think those 3 pillars, those 3
18:32
components, they apply very well for that.
18:35
I think that it is the answer that like I was kind of looking
18:39
for. So let me go over that again.
18:40
Discovery identity life cycle access management.
18:45
Access, yeah. Why those three items?
18:49
Well, if I think the biggest fear when people think about AI
18:55
and NHI today, similar but not exactly the same is I don't know
18:59
what's out there. So the visibility, the
19:02
observability, you got to find out the sanction and the shadow
19:06
stuff, right? I think that's the, the one
19:09
piece that perhaps differentiates from human
19:12
identity is that you kind of know the humans because you can
19:16
see them. Oh, I have 10,000 employees.
19:19
You have employee contracts, even vendors, which is harder to
19:23
see. You have vendor management
19:25
systems, You have other authoritative sources.
19:27
Whereas they fear with NHI that, wow, those things just pop up
19:31
like gremlins, right? And then it's hard to just to
19:35
keep up with that. So I think visibility becomes
19:37
like a bigger pillar for this type of problem than it was for
19:41
humans. Visibility of humans.
19:43
It was not a big deal in workforce, it was not a big deal
19:47
in CIAM, but it is for NHI, it is for AI.
19:51
So that that is 100% necessary. The second one and and was Simon
19:56
who wrote this, A lot of players in that space, they, they treat
20:00
the ownership. As a yeah, it's a secondary
20:04
thing. I think it's such like a massive
20:06
important thing. I think I've been talking about
20:08
this for more than a year, perhaps close to two years.
20:11
The ownership is the key pillar of everything that we do with
20:15
machines in general. AI or not AI machines in
20:19
general, they need to be owned by humans.
20:22
Same joke I tell every time. I hope time travel doesn't work
20:27
the way we think it works. And machines traveling time to
20:30
kill me because I think it may become offensive one day.
20:33
I say that we need to to own machines.
20:38
They, they can't be there by themselves doing anything what
20:41
they want. Now we humans, we are
20:43
accountable. We are the ones responsible for
20:45
that. So the ownership.
20:46
I've heard a theory. I've heard a theory that maybe
20:49
humans will work for the machine someday.
20:52
Oh man, I I OK. I think I think machines are
20:57
great and I just wonder if any machines are listening to this.
20:59
My name is Jeff and I'm I'm, I'm one of the good ones.
21:03
Yeah, see, maybe they will. They'll watch.
21:05
They didn't say. They'll they'll come back and
21:06
they'll save Jeff and they'll keep Enrique and, and but to to,
21:10
to Jim's point, we hear a lot about human in the Loop.
21:14
How crazy is that? Think about human in the loop.
21:18
It's almost like the loop is this huge thing and the human
21:21
has this tiny piece here. So we are part of a, a loop of
21:25
machines and we are their slave. Sorry, but you know what I mean.
21:31
And I think it's a it's a very interesting point to to think
21:33
about. And and Jim, I think where are
21:35
you that I think you brought up with me in a in a separate
21:38
conversation about this. That will be a point where the
21:41
machines are doing their jobs and they hire humans to complete
21:45
like, hey, I need to do Electro electrical work to finalize this
21:48
house. So I'll find a human like Uber.
21:51
I'll the human do the work for me, the connections and then
21:55
human. OK, I'll pay you some tokens and
21:57
you go happy in your life and put your VR goggles in your
22:00
home, right. And then that could be the
22:04
future, which I, I don't like that a bit, but when I hear
22:08
human in the loop, that's what I think about.
22:12
Yeah, it doesn't sound like in the Loop is the primary actor,
22:15
right? It's like you unimportant plebe
22:18
will let you know, but we're going to keep going.
22:21
It's kind of OK, so you cover. Yeah, let's, let's think about,
22:25
yeah. The human, the human in the
22:26
control plane, the human in the control plane.
22:30
Yes, I I I can support that idea.
22:33
You've got discovery, you've got identity life cycle, which I
22:38
think includes a lot of different things, includes, I
22:42
mean. Governance ownership the the
22:45
succession management so if if your owner.
22:48
Request approve, You've got the provisioning, you've got the
22:52
access reviews, you've got also role management, then you.
23:00
Go over his permissions, yeah. The third one was access
23:05
management. What does that include?
23:07
I think that the runtime stuff, right?
23:09
So think about what the challenges we see today with the
23:14
authentication authorization of AI is that now it's not only
23:17
policy based and and more, right.
23:20
So policies are not enough. So if you say, hey, you, you are
23:24
authorized to read and write my emails because I want you to
23:30
organize my, my inbox. So clean up my inbox.
23:35
The the bot may think, OK, let me erase everything then it's
23:39
clean. So no, that was not the intent.
23:42
So I think the access management, the runtime stuff,
23:44
Jim, we're going to see this more of intense based
23:47
authentication, intense based authorization, more on the
23:49
authorization side. But say, OK, you are authorized
23:52
to do a lot of things from a policy standpoint, but from an
23:56
intent you are more restricted. And that is a whole new concept
24:01
that didn't exist for humans, right?
24:03
How far along are we as an industry?
24:06
I mean are the tools that we make available for, I'm not
24:12
talking about your specific company as an industry, do we
24:16
have everything that people need to manage the problem and then
24:21
for organizations, are they going to get caught flat footed
24:25
with this? No, I think we're, we're, we're
24:28
not there even for humans. I think there's still a lot of
24:32
pieces using there, right. But I can, I can give examples
24:37
without this is not a shade on anybody, but I think a lot of
24:40
start-ups, they start with the visibility, which is a very big
24:43
problem. But a lot of them are doing what
24:46
we call IVAP, right. The identity, visibility and
24:48
intelligence platforms that the acronym Gardner uses or the ISPM
24:52
focus. They they do that perhaps into
24:55
some lightweight all the things. There are very few that do the
25:01
runtime authorization authentication piece like a
25:04
gateways and stuff. And and so no, I don't think
25:09
there's a lot of companies doing the three.
25:10
It's not just because my employer.
25:15
But so with the that framework that you gave, if the company is
25:23
or an organization is kind of in the beginning of dealing with
25:27
this, they realize they haven't done enough and they needed to
25:31
catch up. Where do they put the focus in
25:34
terms of getting ready to manage the identity of AI agents
25:40
properly? Yeah, think about this.
25:44
It really depends on the stage that company is, if you are a
25:48
startup or a company starting today, right?
25:51
So you don't have a lot of legacy stuff, you could start
25:53
with the outside of governance piece.
25:55
And, and I, I said this on a, on another interview the other day,
25:59
say the first agent that you launched, it should be already
26:02
with the owner. It's, it's nice.
26:05
And, and you're going to recertify access of that agent.
26:08
If you're not ready to govern that agent day one, you're not
26:12
ready to launch AI at all. Now if you are another type of
26:17
organization that have a lot of legacy, you have this huge
26:24
infrastructure that you built over years in cloud, you have
26:27
API keys everywhere. You may be benefiting.
26:30
Hey, I don't I don't know where things are starting with the
26:33
visibility. The hygiene piece may be my
26:36
recommendation. The access management in the in
26:41
the, the runtime in line enforcement I think will be a
26:45
third thing that we we can do, which is very interesting.
26:49
That is a different order. When we think of workforce,
26:53
right, that most companies that will start with an IDP, might
26:56
they have there an SSO, an MFA tool they start with that.
27:01
It's a tool to install. People authenticate through
27:03
that. I think for agents I see this
27:08
different order of things, at least for now.
27:11
It may be different in the future.
27:13
Yeah, I feel like a lot of insight there.
27:18
I think that last point also was a nice bow that you put on it,
27:21
which is different. Like humans usually organization
27:25
start with their IDP and then go into maybe like an IGA mode.
27:31
IGA and having yeah. I also kind of feel like with
27:36
this framework, all the components add value, but I
27:40
don't think you have to build any of them out completely.
27:43
Like you don't have to solve all identity life cycle, you can
27:48
just focus on some of it. OK, Yeah, yeah.
27:53
And I. And I think perhaps, like I
27:56
said, to put a ball on it is thinking about the scope of what
27:59
you're working with. And, and and that's it can be
28:03
kind of kind of surprising. But when I speak with clients
28:06
and, and leaders, the scope is so much, it's, it's even hard to
28:10
grasp how big is the problem. But OK, now let's, let's create
28:15
compartments and, and start fixing this.
28:17
You're right, we don't need like the whole thing for the whole
28:21
depth of capabilities for everything.
28:23
And, and I think having that a realistic approach to the
28:27
problem, it can be less overwhelming.
28:29
Otherwise you're going to be pulling your hair.
28:32
I don't want to do anything. Hands in the air.
28:35
I, I don't know, I give up kind of thing, right.
28:37
So I hope people feel more motivated.
28:40
There's a, perhaps a prescriptive step that we can do
28:44
and, and, and, and measure OK, improvements.
28:47
I, I use this, this knowledge with my son, right.
28:50
So sometimes I say, oh, I, I don't think I can win races in
28:54
skiing. I didn't hate you, man.
28:56
But the whole point is #1 are you having fun?
28:59
Why are you practicing and #2 are you improving compared to
29:03
what, what you did last time, right?
29:05
Or last year? So same thing I'm telling my
29:08
kids, perhaps to the CSO's out there, are you having fun?
29:12
Why are you doing this? And are you improving compared
29:16
to yourselves, right? So having those metrics are
29:18
those KP is improving. Forget about the outside word
29:21
for a little bit, but are you improving?
29:23
I think that's a that's a perhaps a refreshing way of
29:26
thinking about it. Well, even if you tried to do
29:29
everything, you would never be done anyway.
29:31
So you kind of already have to draw a line to say, OK, we're
29:34
going to take this capability this far because that's all that
29:37
makes sense for us. Like it doesn't make sense to be
29:40
like perfect across the board and have every single use case
29:43
solved for 99% of the companies that are out there.
29:48
You have to draw a line somewhere.
29:49
Anyway, you're never, you're never done.
29:50
Absolutely. I agree.
29:51
I agree, Jeff. I want to pivot the conversation
29:54
a little bit into analyst territory because you've got
29:57
experience in this space, you know, formerly with, with
29:59
Gartner before you joined the dark side and went to product,
30:02
right? In the same way that I was an
30:04
enterprise before I joined the dark side, I went to consulting,
30:06
right? I'm, I'm curious what your
30:09
thoughts are on how that traditional approach of, you
30:15
know, I look at these analyst reports and say, OK, here's
30:19
where the information is coming from.
30:22
And now we've got chat bots that will surface a bunch of
30:26
information. Chad, you know, Jim kind of
30:28
mentioned it up at the very beginning here of, you know,
30:31
building like search engines to surface content, etcetera.
30:35
I'm curious where you stand on sort of that analyst model
30:39
because I think there's a shift that's sort of in progress
30:42
that's happening. I'm seeing the consulting.
30:44
I'm also seeing it maybe at the analyst level.
30:47
What are your thoughts on where AI is maybe starting to change
30:50
that business model a little bit?
30:53
I think the analyst world is changing.
30:56
I agree and I disagree that AI will destroy the profession of
31:02
analysts. Of course, Full disclosure for
31:05
the ones I don't know, I'm a former analyst, right?
31:07
So coming from that world, I have tremendous respect for the
31:11
profession. Today, I use them as a client.
31:15
So I, I became a client of analyst firms and, and so coming
31:20
from that perspective, I think AI can help a lot in, in the way
31:27
and the speed research can be written because there's a lot of
31:31
the, of the processes, taking an idea and digesting that idea,
31:35
research writing. I think a lot of the editing, I
31:38
remember when we were publishing magic quadrants, there was like
31:41
weeks of that process. Like I'm not creating new data.
31:45
It wasn't it was just the editing, looking for English
31:49
grammar corrections and and other factual things.
31:52
But so absolutely it, it can help there, I think, even for
32:00
perhaps easy to find answers. I was researching the other day
32:04
say, Hey, I'm I'm my fridge here at my house.
32:09
It's it's kind of getting old. What's the best fridge that can
32:12
fit here? And I think a lot of CSO's, they
32:15
are doing that. Sometimes they come with our
32:18
sales reps as saving and say, hey, the client Googled this or
32:23
asked Claude for who is the best IGA tool.
32:27
I think, well, they're going to get what is publicly available
32:30
out there. I think the AI will be
32:33
interesting and helpful in just finding data that is available.
32:37
But to Jim's point, not all available, not our knowledge is
32:42
like readily available, right? Not yet.
32:45
But what is not going to be there is the actual analysis.
32:50
And by that I mean to find strengths and weaknesses.
32:54
That's easy, very easy. You can quantify.
32:57
You can even have a scoring system and say this is stronger,
33:00
this is weaker, this is better, this is not so good.
33:04
So you you can't quantify things.
33:06
And I think AI can even help you quantify things.
33:09
Now, what's best for you if I'm buying a refrigerator, do you
33:16
like French doors? Do you like style?
33:18
Do you like price? Do you like space?
33:21
So those type of things will. Influence a lot.
33:24
The analysis here and since my, my two years in this role, which
33:32
is strategy and includes competitive intelligence analyst
33:35
relations, I think more and more I, I, I use the analyst
33:40
information to, to help me with, with that part of the inference
33:45
of the data, right. So I had the data, everybody has
33:48
access to data today. I think the data is very cheap.
33:51
So the thing today, so, and by the way, just a little
33:55
parenthesis. If you're building a business
33:57
model in accumulation of data, even if you're like an
33:59
individual, Oh, I told my data I'm you're done that that is
34:06
going to be killed by AI yes. And your your peers, they're not
34:11
going to be your friends. They're going they're not going
34:13
to have fun working with you because if you're protected, no,
34:16
bring them share data. But now the inference, right.
34:19
So the analysis on this is the best fridge for you, You guys,
34:24
you aren't consulting right. So the how many times a client
34:27
would approach us and say I have this problem, What's the best
34:31
solution for me? Right?
34:33
It could be anything. It could be guitars, it could be
34:35
cars. The.
34:36
First answer is always. It depends.
34:39
That is the first consulting answer.
34:41
It depends. Right.
34:43
And so that that that is my long answer to say no, I don't think
34:48
AI will replace them, but also because I think humans like to
34:52
pay to see humans on stage doing human things.
34:58
So, yeah, so from the conference perspective, yeah.
35:02
You know, I I keep hearing this thing that companies that are
35:07
going to succeed in the world of AI are ones that have built a
35:11
Moat. And so when I think about
35:14
analyst firms, I wonder what is their Moat?
35:17
What do they have that will come that will prevent AI from eating
35:23
their lunch? I think the other thing that I
35:26
here a lot is that the value will be in the data, having the
35:31
proprietary data. The analyst firms have that and
35:35
yet Wall Street's been punishing them because it's, you know
35:39
what, what does this AI going to do?
35:42
It's going to eat their lunch. But it seems to me that, you
35:46
know, the analyst firms are also providing net new data.
35:52
They're, they're creating the data by, you know, not
35:56
evaluating the. Analysis, the analysis, right.
35:59
So I'll give you an example, yeah, Jim, which which is quite
36:03
perhaps this is a secret world in, in, in analyst firms, but
36:08
it's not a secret per SE. But people don't they, they look
36:11
at the reports, they look at the events or at the conferences,
36:14
right. But while I was a gardener, I
36:16
was taking more than 1000 increase a year, which means I
36:20
was talking to another human on the other side of the phone.
36:23
They're telling me problems. I was listening to problems and
36:25
recommending solutions that become data points that we can
36:30
extract. Oh, there's like half of them.
36:32
They were concerned about IGA. Half of them were and that how
36:37
can you automate that with AI, right?
36:41
That type of data point, by the way, right.
36:43
If I ask the question this year's, they're going to say
36:46
something. If I ask this question next
36:48
year, that that's why it's a machine of increase.
36:51
So and next year the trend is going to be different than
36:53
shifting to other areas. But because I'm a human talking
36:57
to a human. So that part of the equation
37:00
here, I think it's very difficult to replicate.
37:03
That's why the Moat for Gartner, Forrester, the other firms that
37:08
those there are based on inquiry, just the scale, but
37:12
also the data, they're fresh and, and, and, and to have this
37:17
type of competitive intelligence for me today to compete,
37:20
freshness is super important. So I think, yes, I imagine one
37:25
day because answering to NQS sucks.
37:28
I don't like spending time answering 300 questions.
37:32
I meant maybe I could have an agent within my company and say
37:36
this is everything that our product can do.
37:39
Then Garnick would have an agent.
37:41
So they'll collect everything and they will calculate and then
37:45
a human on the control plane, right?
37:47
Not in the loop. A human, the control plane say
37:51
interesting, but this analysis, it's a little bit off.
37:54
The evaluation criteria here is a little bit different.
37:58
I want to put a higher weight on this versus that.
38:02
So I think there are a lot of opportunities for optimization
38:07
of the, the process of the analyst world.
38:10
But I think what is also interesting, Jim, and, and, and
38:14
this is a counter argument. What I've just said is that
38:16
there's a lot of individual and boutique players today, right?
38:20
So I, I, I don't think Garner is alone by themselves or Forrest
38:24
or copied your coal or IDCI think there's a lot of amazing
38:28
people, right? I mentioned some of them already
38:31
like Simon Moffat with a Cyber Hut or Francis Odom or
38:38
Lalitschoda. There's a lot of interesting.
38:41
It's almost like a new crop of boutique influencers,
38:46
individuals that are also creating this very pointed type
38:52
of advice. And I think, I think buyers are
38:55
listening to them more as well, at least using them in in
39:00
addition to the traditional model of the analysts.
39:05
I think what it comes down to is trust.
39:08
Do you trust that the information you're getting is
39:11
coming from a reputable source? Do you like the source?
39:16
Whatever like means, right? So, you know, when we talk
39:19
about, you know, whether it's an analyst firm and there's many of
39:22
those, right? You've got Gartner, Forrester,
39:23
Gomer, Cole, right, kind of the big ones.
39:26
You've got individuals, you know, maybe smaller analyst
39:28
firms that also put out great research.
39:31
You've got practitioners who are also doing things where they
39:33
are. You know, they're speaking at
39:35
Ideniverse and they're speaking at sort of like vendor neutral
39:37
conferences and bringing up topics.
39:40
You've got podcasters. I mean, like it or not, you
39:43
know, there's this whole influence for society.
39:46
You know, do people trust the people who are bringing in
39:48
information? And maybe it's not trust of
39:51
person. You know, I, I like to think
39:53
that we've built up trust over the years here on the podcast,
39:56
right? And through our consulting
39:56
careers and say, hey, you know what, these guys are straight
39:59
shooters and they're not, you know, skeevy, right?
40:01
Or whatever, you know, you want to want to do.
40:04
But that trust is something that takes a long time to earn, and
40:08
you can disappear in a flash. I I, I, I agree.
40:12
And especially when, well, going back to the example, a
40:15
refrigerator, right? A silly thing like that, but
40:18
who, who do we consult when we're buying things, right?
40:22
So we love to get reviews, we love to get opinions, right?
40:27
So we go on Amazon, we're going to read the reviews.
40:30
We're going to look how many stars this thing has.
40:33
We trust that system that Amazon has put in, right?
40:36
We trust the reviews of of especially other buyers, right?
40:41
Other buyers, what they're saying, the independent as well.
40:45
I think I, I, I come from an era of people that walked into Best
40:49
Buy and they have those what they call the genius.
40:52
What? What is it?
40:53
The Yeah. The well, they have the what is
40:55
it? The G Squad?
40:56
Apple G. Squad, the genius is Apple,
40:58
right? Yeah.
40:59
So hey, I need like a printer, right?
41:04
And and I kind of missed that day.
41:06
So it, it could be perhaps a little bit nostalgic here, but I
41:09
think humans, we, we make a lot of decisions based on trust and,
41:14
and who we like, right? And, and it could be this firm
41:18
or this other firm, especially if I think I'm thinking about
41:25
what means to gain trust. And I think it's the
41:29
repeatability, right? So if, if I watch your podcast
41:32
once, that was a great episode. But if the second one is, you
41:37
see, I'm not going to watch again, right?
41:39
But it is consistent showing that, hey, or if, if you go to a
41:45
big analyst firm or even if, if you're going to giga on or QKS,
41:50
there's a lot guys and, and, and you keep saying, wow, this
41:54
applies to me. I, I read this and I, I bought
41:57
this and it, it worked well. Maybe I'll go back and, and, and
42:01
listen to them to them again. So I think it's the
42:04
repeatability as to trust to your point, right?
42:07
And it, that takes time. How many episodes do I need to
42:10
watch? I didn't understand and say OK,
42:12
I think I like them. Well, you should be watching all
42:16
421 episodes of The Identity Serpakis.
42:20
You know, you bring up recommendations, and that was
42:22
kind of where I was leading This next was, OK, so we've built
42:25
trusts, you know, whatever, whatever source you trust,
42:28
right? You've built up that trust.
42:29
And people rely on recommendations.
42:30
You mentioned star ratings and reviews, right?
42:33
All that stuff. And I think that's where really
42:35
a lot of the unsurfaced intelligence comes from.
42:41
It comes from, Hey, let me talk to another one of your customers
42:45
or let me talk to another one of your, you know, another company
42:48
who has experience with this product or service or whatever
42:50
may be. And you start to develop these
42:54
internal, you know, professional networks.
42:56
So I can say, hey, Enrique, give me the scoop on this, right?
42:59
What do you think about this? Here's kind of what I'm dealing.
43:01
We have the, we have a, you know, a trusted relationship and
43:04
you're sharing trusted relation, you know, information to me.
43:07
I'm sharing trusted information to you.
43:09
The same thing is happening right now, you know, obviously
43:12
with, you know, all the different analyst firms and all
43:14
the things like that. I don't think we're too far away
43:16
where where we start to develop that same trust relationship
43:20
with AI chat bots. For example, maybe there is a
43:26
proprietary chat bot and you know, look, everybody's working
43:29
on a chat bot of some sort. It's trained on their own data.
43:31
But do I trust the identity at the center chat bot or do I
43:36
trust the, you know, the identifiers chatbot or the
43:39
Gartner or the coal or, you know, cyber Hut, you know,
43:42
whatever it may be. I don't think we're too far away
43:45
from like being in a state where we're not only starting to trust
43:48
people, we're starting to trust, you know, an AI that is coming
43:52
through this data because for whatever reason, we like what it
43:57
says or we like the output or it has proprietary information that
44:01
no other chatbot has. And I think that's yes.
44:04
And then I think that's kind of talk about.
44:06
That, that, that is the key point, Jeff, because if you
44:09
think about, yes, I, I might like your chatbot better than
44:13
somebody else's because of the data it exposing to me.
44:17
So similar to the stars on an Amazon review, right?
44:20
Those stars, they're, they're humans who clicked.
44:24
I think right there there might be bots, but that's a whole
44:27
different type of AI. But.
44:29
Well, that is a problem, right? A lot of fake reviews and
44:32
there's even rousers. And those imagine that 04000
44:37
people bought this same refrigerator and they like them.
44:41
And the same concept we Garner as well.
44:44
Gardner is saying this vendor is a leader because blah blah,
44:48
blah. A lot of a lot of the inquiries.
44:50
So those thousands of inquiries with people that really like
44:54
this product. So I think when we we we look at
45:00
at that I think trust would also be related to the data.
45:04
There might be chat bots providing similar answers and
45:08
say, hey, it could be an interesting idea for business,
45:12
right? So Can you imagine Amazon you,
45:15
you're buying a hammer And they're like star reviews from
45:18
humans. And this is my review from an
45:20
AI. From the from the AI that was
45:23
using the hammer on the humans. No, it's getting dark, but yeah,
45:28
but well, I think this is the SO.
45:31
I think this is the business model that changes first or
45:34
maybe in in the earlier parts is we're all familiar with the
45:39
exchange of your e-mail address for a report, right?
45:43
Every single analyst firm does a thing where hey, download the
45:46
new Leadership Compass, download the new, you know, Magic
45:50
Quadrant or Buyer's guide or whatever report that is like the
45:53
thing that everybody looks at. You type in your e-mail address,
45:56
you know, you go to a vendor site, you type in your e-mail
45:58
address and you get it for free. Now you're exchanging your
46:01
contact information, etcetera. That is the model that we're in
46:05
today. I can very clearly see a spot
46:08
where that is no longer available.
46:10
You are That information is gated behind something that
46:14
makes me say, OK, I need to go to that analyst firm directly
46:19
and either have a subscription to that bot or whatever to get
46:22
that research because. Yeah, exactly.
46:25
Yeah. Because now, now my bot has
46:28
proprietary information that nobody else has.
46:30
It's not just all over the Internet that I can get for
46:33
free. What do you think about that?
46:36
It'll be an interesting business model.
46:38
I I think especially talking with smaller firms, for example,
46:44
Gardner doesn't do Commission research, but the ones who don't
46:48
know that Commission research is anybody, it could be my company,
46:52
your company. You approach an analyst to say,
46:55
I want to write about electric guitars and to generate the
47:01
demand or or the interest for people to buy more guitars,
47:04
right. And then of course, you sell
47:06
Fender or you sell keeps that could be a lead generation for
47:10
your company. But so this is a this is a
47:13
business model that a lot of analysts, they do.
47:15
So they they charge, they charge the vendor that wants to talk
47:20
about a topic. So those are Commission
47:22
research. So Garden doesn't do that.
47:24
They, they charge their clients subscriptions to have access to
47:29
this increase to. So they, they it, which is a
47:31
very, it's a very intelligent model.
47:35
So I'm a client, I pay Gartner, I ask them questions, they take
47:41
data points from our conversation say, wow, that was
47:43
another client who mentioned this vendor.
47:46
So that is also feeding their engine of data they sell in
47:51
their reports. That's why I think it works so
47:54
well. So there are different models
47:56
for monetizing in analyst world, right?
47:59
So some would charge for contracts in consulting.
48:03
So I, I, I do consulting and based on the things that I
48:07
learned during my consulting gigs, I'll write this report.
48:12
A lot of analysts, they do that. And there are influencers too.
48:16
Some, it could be a former Tier 1 analysts, they leave and they,
48:21
and they do lectures, right? So they, they will make their
48:24
money on the events and, and, and as a speaker and, and stuff
48:28
like that. What what you described about
48:33
the gated content or not, I think is, is just another way
48:37
of, of, of revenue or how revenue could flow into the
48:40
analyst world. It is an interesting thought.
48:45
I I, I never thought about I would be, you know, this now
48:51
imposter syndrome time, right? I never thought, OK, maybe
48:54
Enrique by myself. Can I, can I be an analyst,
48:57
right? And, and having my firm like a
49:00
Kupinger, Martin Kupinger as well.
49:01
He was an individual once until Kupinger Cole became Kupinger
49:05
Cole or Francis Odom or a Simon Moffat or El Alicia ODA.
49:11
So that's why I say, well, those guys are amazing.
49:14
I admire the, the courage. I think any and anybody that has
49:18
the courage to launch a an organization or a launch a
49:22
company, I think well, you guys, what you did with podcast, I
49:27
never launched my own company. Maybe one day, who knows.
49:31
So what I was going to say earlier when I brought up the
49:35
motes, I think the motes that analyst firms potentially have
49:39
is IP rights to their information, especially to their
49:44
analysis. So, you know, I don't think it's
49:48
legally kosher for an AI large language model that's owned by a
49:54
company to take that information, you know, go out
49:59
and procure it. However, like create a fake
50:01
e-mail address, get that data and plug it into their model and
50:05
then start spitting out like, hey, you know, Savient is in the
50:09
upper right quadrant was like, what are you talking about?
50:12
If you could prove that essentially is taking that IP
50:17
and passing along as its own, that is totally not kosher,
50:21
right? I think also when you think
50:25
about it though, there are people who businesses and
50:28
individuals who are paid to train AI models.
50:31
I learned about a business where.
50:33
I know people doing that. Yeah, they're lawyers and they
50:36
train AI models on how to review legal documents and things like
50:41
that. And I know I've gone to LLMS and
50:45
uploaded files and said how, you know, TLDR basically summarizes
50:52
document for me or find something.
50:54
And now that's part of the model, right?
50:57
I just trained the model with IP that belong to another company.
51:02
So I haven't heard anything about lawsuits about that yet,
51:05
and I don't know what it would be based on, but I really feel
51:09
like. Well, you you may have the whole
51:11
yeah, artists there was a whole controversy around copyright and
51:17
infringement when they say, well, this is AI generated
51:20
picture. Wow, no, this was a real
51:22
picture. They just you add another
51:23
finger, right, But it was an existing picture.
51:27
Maybe they swap the faces, but it it is art that somebody else
51:31
created right. So there's a there's a whole
51:33
interesting discussion. I am a big advocate to talk more
51:36
about that because I think it's a big risk.
51:39
I think the risks of AI in in culture and appropriation of
51:44
things, but even more the effects and the negative effects
51:49
of AI in our, in the perhaps the erosion of culture itself.
51:55
I think it's, it's big. I don't want to perhaps be very,
52:01
very, let's say controversial here, but I think yes, we are
52:04
afraid of AI replacing jobs. Yes, I think this is this is a
52:08
real risk, but I think the risk of AI eroding culture is bigger.
52:15
So give us an example, because I kind of feel like technology
52:18
overall, not just the Internet, right?
52:21
Even before the Internet, technology overall has been
52:24
eroding culture. I mean, let me use an example
52:28
that comes to mind, which is when I was growing up, there
52:31
were like 5 TV channels, right? And you can go to school the
52:36
next day and say, oh, did you watch this show?
52:38
Everybody, you know, half the people watch that show.
52:41
If it was the popular show, I can tell you I've got like 500
52:45
TV channels now. There's no way that we, you
52:48
know, and we have streaming. I've got all these apps and the
52:52
technology has just made it so that we, you know, people don't
52:57
enjoy the same cultural experiences from ATV
53:00
perspective, but you. You, you are describing what is
53:03
called a, a shared experience, right?
53:05
And, and we as humans, a big portion of our culture of humans
53:11
is to have those shared experiences.
53:13
It, it could be artistic, like you go to a concert, right?
53:17
Or you, you see, you watch music together.
53:20
It could be sports or religion. And I know this part of the
53:24
discussion may sound like tinfoil type of fear mongering,
53:29
but hear me out. I think the risk that we're
53:31
talking about here is instead of having that shared experience
53:35
which generates cultural moments.
53:38
For example, the last episode of Game of Thrones, like it or not
53:42
right, everybody was watching the drop of that episode.
53:45
We are all talking about the same thing with AI technologies
53:50
becoming powerful to the point where I can create not only
53:53
pictures, I can create the whole movie.
53:56
Think of Sora that just got shut down.
53:59
Sora was this AI application. It could build smart short
54:03
clips, but it's not too far fetched to imagine that we could
54:07
create a whole episode of ATV series, say, hey, I like Back to
54:11
the Future, I like Star Wars, I want to mesh up with Rick and
54:14
Morty and then make a series. I'm not a lot awesome by the
54:18
way. I would totally sign up for
54:19
that. I would too, but the risk is
54:24
then we create this individualized content that is
54:26
amazing for me, maybe for me and Jeff only, right?
54:30
Maybe we share that experience, but then billions of users will
54:33
never heard about my crazy mash up story, right?
54:37
And I think we missed that social experience of OK, we we
54:44
like the same thing or even we hate the same thing.
54:47
If we go to a concert, Justin Bieber or or oh, there's this
54:52
new band. They call this the anti AI
54:54
movement with the AADP. It's Angina de Quatrin.
54:58
I had to, I had to mention that as a, as a mandatory item, if
55:01
you if you heard them. So this I don't know.
55:03
Have you guys heard this band before?
55:05
I did check it out after we talked about it the other day.
55:07
Yes, it was cool. Yeah, it is micro tonal math
55:11
rock and they just came out of nowhere, but they they've been
55:14
playing together for 18 years or something crazy like that.
55:17
But they they just have this they say they're they're aliens,
55:20
right. They arrive on Earth.
55:21
So this all this is a perfect example of something outside of
55:25
AI because it this the music looks like AI because so nuts.
55:29
And then the costumes are crazy, like they're dressed up like
55:33
aliens and whatever. And what happened was this
55:37
cultural movement that everybody was talking about it.
55:39
So Rick Viado influencers talking about this thing.
55:43
So I think that type of the AI eroding culture, the risk is if
55:49
we go that route of creating content and it could be a song,
55:54
it could be art, paintings, music, movies, we're getting to
56:00
that focus of just me enjoying this.
56:03
And then we missing on that bigger picture of the shared
56:06
experiences. I think we, we, we started with
56:08
that, by the way, Jim, when we think of streaming for music,
56:15
right? So the ability of creating
56:18
playlists, right, So we were not listening to albums.
56:22
They were meant to be listened anymore, right.
56:24
So I have my playlist, which is different from yours, which is
56:28
so we, we, we share that that perhaps the the song is Oh, I
56:31
like that song or I like that band, but.
56:33
But we were doing that a long time ago with mixtapes, right?
56:37
I don't know if it was when you were young, the whole world was
56:41
maybe not. Maybe it's just me, but you drop
56:44
a cassette in your stereo, listen to listen to the radio
56:48
station, then hit record when the new song started, and if it
56:52
wasn't the song you wanted, you'd rewind back and wait for a
56:55
song you'd like. Now it's like you've got this
56:57
golden mix tape. And if somebody would?
56:59
Say some bullshit that made us that was.
57:02
Linky Park, I said. Oh man, yeah, yeah, you get
57:07
those tidbits. And now we pay for that classic
57:09
experience of tapping that stuff in, right?
57:12
But that, that's what I meant on, on, on the erosion of
57:15
culture. And so I'm not too bothered
57:18
about the AI generated music or the AI generated art.
57:22
I've been using electronic drums, AI drummers for making
57:27
music for decades, perhaps 10 years.
57:30
I remember Fruit Loops was one of the software that for, for
57:34
creating drum tracks. And it's OK.
57:37
I think that that's fine. I I don't think we all there's
57:40
AII don't people, I don't think people care.
57:42
I don't think the next generation will care if it's AI
57:44
generated music or not, But I think the shared experience is
57:47
going to be important. Even if you're Justin Bieber,
57:50
right, Justin Bieber on stage, sharing a screen YouTube, it's
57:55
still a human doing silly human things on stage.
57:59
And I think I just found our next speaking engagement thing,
58:02
Jim, at a conferences, we'll just hop on a stage and just
58:04
play YouTube videos of of. Our of episodes, then you do a
58:08
this could be a great idea. You you just play an episode of
58:12
yours and and you do commentary. Yeah, and it would be a kind of
58:16
a. History sign a good question.
58:17
People can throw their tomatoes and you know, stuff it like
58:20
these guys are don't know anything what they're talking
58:22
about. I don't think that's a crazy
58:24
idea. I like it.
58:25
I don't know if I'm brave enough to try it.
58:27
We've definitely strayed into what I would call like lighter
58:29
note territory. I want to keep talking about the
58:31
AI stuff because I think he brought up something that I've
58:34
been thinking about recently and that is that personalization of
58:38
entertainment. And so you talk about, you know,
58:40
music and this has been around for a while, you know, if you
58:44
remember Pandora from a very long time ago, which
58:48
theoretically is still out there.
58:49
I can't remember the last time I used it, but it was, you know,
58:51
kind of one of the the originals of like, oh, let me create a
58:54
music station and let me seed it with a couple of artists or
58:58
bands that I like. And then let me use the thumbs
59:01
up, thumbs down to train, you know, the algorithm that we
59:04
existed to say, oh, Jeff likes this.
59:07
So play more songs like this. And they had the Music Genome
59:10
Project to do all that kind of stuff.
59:11
And it was OK Cool, right. And I know a lot of people that
59:15
have very, very meticulously tuned Pandora stations because
59:21
it is exactly they've spent years, right?
59:24
And hundreds, thousands of songs, yes, no, you know,
59:27
whatever it may be. And they probably have multiple
59:29
stations that they. But was there like a social
59:31
element to that as well right there?
59:33
Was a little bit because you could share, you know, to some
59:35
degree, but it was still very personal.
59:37
It was this is Jeff's station and this is what he likes.
59:41
And it happens to be a whole lot of EDM and new metal.
59:44
It's like, OK, that's just, that's just me, right?
59:47
But now we're in this era of AI and generative capabilities.
59:53
You know, we're very close to the point where there's a lot of
59:55
these, you know, music services that will generate music based
59:58
on whatever you want. You give it some prompts and
1:00:01
based on your prompt or you know your specific direction, will
1:00:04
generate an entire album. Not album, an entire track for
1:00:09
you, complete with. Singing all the musical anal
1:00:13
things lyrics right etcetera. And so instead of having a spot
1:00:18
where I need to see the station to say OK, play Korn, Linkin
1:00:23
Park, Nirvana and Rez and whatever it may be, right, It's
1:00:28
going to become always hey, I want to hear this style of music
1:00:31
and instead of playing I. Want all original?
1:00:36
It's generating A continuous stream of AI generated music
1:00:40
that is tuned to my tastes. That that, that is what I I
1:00:45
think will be collapse of the humanity, I think.
1:00:48
That's and it may be good and maybe not, but it'll be 100%
1:00:51
personalized. Yeah, exactly.
1:00:53
And the same thing for. Movies, right, I want to see the
1:00:55
risk is that people people going to love it as a wow, this
1:00:59
station, this look, I created a another record for Nirvana,
1:01:03
right, a banner that that doesn't exist anymore.
1:01:05
What if was a new album and then boom, AI will generate that
1:01:08
album. I will love it.
1:01:11
Maybe it could become something that you can share and and more
1:01:15
people. OK, I'm going to buy that album.
1:01:19
Food for thought right. So there just this week I saw,
1:01:21
for example, a movie they they recreated Volk humor using AI
1:01:28
and he's there acting. Of course, everything is legal,
1:01:31
his estate approved and and but I, I think that kind of
1:01:40
contradicts what I said AI in that sense, even though it's
1:01:42
like ethics weird, it, it's still OK from a shared
1:01:46
experience. It's going to be on on theaters
1:01:49
may right, maybe running for Oscars and stuff, but it's AI
1:01:54
generated, but it's very different from something
1:01:57
personalized. I think they used Val Kilmer in
1:02:01
a mainstream way to sell tickets for this movie The traditional
1:02:05
way, right the the the movie industry likes to do.
1:02:08
So I think it might be kind of a controversial topic, but I think
1:02:15
the individualization of the content allowed by AII think
1:02:19
that's the risky part of it. I wonder if I've heard that what
1:02:27
we pay for AI is not even close to what it costs to do the AI,
1:02:32
right. And so maybe that switch has
1:02:36
flipped some point in the future where energy gets much cheaper,
1:02:40
the processors are much more powerful.
1:02:44
But I'm kind of wondering if it becomes, you know, right now
1:02:48
it's like, Oh yeah, just create me an album.
1:02:52
Well, what if that cost you 2999 to create yourself an album?
1:02:55
Yeah. So you want an album?
1:02:56
Yeah. Do I really want to create an
1:02:58
album or not? Yeah, you know, like.
1:03:02
I like that. I like that business model as
1:03:04
well. Jim.
1:03:05
Hey, you, you, you. You want a new song?
1:03:08
OK. How much would be to hire like
1:03:10
humans to to record like a custom song for you, Right.
1:03:14
So, yeah, I think there's some interesting things to be
1:03:16
discussed here. You want.
1:03:18
The novelty in, you know, 20 years.
1:03:19
Like what do you mean a human created this?
1:03:22
Yeah. Bunch of hipsters and hats and
1:03:25
mustaches and all kinds of stuff.
1:03:29
But I think we're, you know, there's a couple of ways to look
1:03:31
at this from a content perspective is, you know, you
1:03:33
bring up a good one, Enrique, where, you know, there are the
1:03:35
estates of people who have passed on that they have
1:03:38
licensed their likeness to be using that scenario.
1:03:40
I think Harrison Ford might be another one.
1:03:43
I think Morgan Freeman is not the one his voice can be used in
1:03:45
certain scenarios for, like Star Wars.
1:03:46
Content yeah for the for the voice on Star Wars I think I
1:03:49
think he he's signed already right yeah to authorize his use
1:03:53
of this voice yeah as I think and that's a whole different
1:03:56
discussion right which I think it's more in the legal realm of
1:04:02
what what's ethics are that makes sense and would you Jeff
1:04:09
sell your avatar to present podcasts in the.
1:04:12
Afterlife. Nobody wants this avatar, that's
1:04:15
for sure. Or this.
1:04:16
Nobody wants this book or this voice.
1:04:18
So I feel it. How would you make you feel that
1:04:21
that whole idea I, I don't know how I, I don't know how I feel
1:04:24
about that. You know, I don't know, I think
1:04:26
this is a Black Mirror episode where, you know, you upload your
1:04:29
consciousness into like being an AI and you know, is that is that
1:04:35
really you? I think I, I'm undecided on it,
1:04:39
but originally I was like, Oh yeah, that sounds cool.
1:04:41
Make me an AI. And, you know, people can still
1:04:43
talk to me after I'm after I'm dead.
1:04:46
And then I, I don't remember where.
1:04:47
I think it was a Black Mirror episode that I saw.
1:04:48
I was like, oh, maybe I don't want that because they brought
1:04:51
up some really interesting kind of negative things that I hadn't
1:04:54
really given a lot of thought. To so I don't know, Yeah.
1:05:00
It might, yeah. It might have been that one, I
1:05:02
don't know. All I know is we put out a
1:05:04
public podcast and so there is nothing stopping anyone in the
1:05:08
world from taking the, I think, almost 400 hours of content that
1:05:13
we've produced and doing something with it.
1:05:17
I mean, I do it with our AI episode every April first.
1:05:20
You know, this year I feel was the best representation of my
1:05:23
voice yet as a voice clone. Jim's not so much.
1:05:27
Now I paid for a little extra, Jim did not.
1:05:29
And it it shows at the end of the episode you see kind of the
1:05:31
difference of it. But as of it stands right now,
1:05:35
today is April 16th, 2026. It is at least capable it would
1:05:41
ever be right now. It is only going to get better
1:05:43
and better and better. Sorry, not better.
1:05:45
More capable over time and easier to use and more available
1:05:49
for everyone and cheaper. I yeah.
1:05:52
And yes, yes. But to your point, I, I think I
1:05:56
like the idea of the finality, right.
1:06:00
So, OK, I, I, I have a mission here on earth and then boom.
1:06:03
And, and it becomes kind of this commodity rarity as well.
1:06:10
So if you want to watch episodes that Enrique was there or Jeff
1:06:13
was there, it's it's going to be there.
1:06:15
But I don't like the idea of more episodes of me and then I'm
1:06:18
kind of diluting the value of where it was in a sense.
1:06:24
Well, I think that's where the, you know, the, the provenance of
1:06:28
the content comes into play. And I think this is an area
1:06:30
that, you know, to bring it back to our identity standpoint is
1:06:33
can you trust that this thing is either real or authenticated or
1:06:39
authorized or whatever may be, right, some sort of identity
1:06:41
verification thing? You know, I, I, you know, the
1:06:44
Blade Runner person in me says like, OK, there will be a
1:06:48
tombstone that will be, you know, the dates that I was alive
1:06:52
and then kind of like a slash. And any content after that is,
1:06:56
yeah, you know, he died in whatever year it was.
1:06:59
But there is still these things may be coming out.
1:07:02
If that's if the person, well, the person may not agree to it.
1:07:06
Somebody could absolutely do this against someone else's will
1:07:08
and recreate things. But, you know, we could be in a
1:07:13
scenario where it's like, OK, the mortal life was this, the
1:07:16
digital life was this. How's that for a lighter note?
1:07:22
That was deep. That was interesting.
1:07:25
And I think, yeah, I want to watch that again because did I
1:07:30
make sense? I think.
1:07:33
I think. Much sense, I think, is any of
1:07:34
us trying to figure this out in real time as it's happened.
1:07:39
Yeah, I I think I I would not license my content afterlife.
1:07:43
That's what I'm I just so you're not.
1:07:45
I thought about that right now. Traded on a on a BitTorrent or a
1:07:50
blockchain somewhere. Hey, I got some of that,
1:07:52
Enrique. Enrique doing crazy stunts.
1:07:54
I don't know, but it's, it is something I I haven't thought
1:07:58
about just now. You asked me the question.
1:08:01
Yeah. I I think my content ends with
1:08:02
me. OK.
1:08:04
That's fair. I I will honor it.
1:08:06
How about that? Yeah.
1:08:08
We all take your episodes down low, sorry.
1:08:11
No, that they'll live forever. No, that can remain.
1:08:13
I'm, I'm talking about new content, yeah.
1:08:16
Gotcha. Yeah, yeah.
1:08:17
All right, let's leave it there. You know, I, I enjoyed the AI
1:08:21
conversation. I think we probably spent like
1:08:23
20 minutes on actual identity, but that's fine.
1:08:26
You know, this is something that's in the consciousness for
1:08:28
us all. And I think we got a little bit
1:08:29
deeper, which I enjoy. So yeah, maybe this last 40
1:08:32
minutes was more for me than anyone else.
1:08:34
AI for me and you've now drank like 5 bottles of whiskey if
1:08:39
you're doing a shot for each time.
1:08:41
Yeah, you might actually be need to be converted into AI if
1:08:44
you're still if you're still alive after that one.
1:08:47
And AI at the center does take you to, you know, our website.
1:08:50
So yeah, I have that domain. Let's go ahead and call it here.
1:08:55
Enrique, thank you so much for being part of this.
1:08:57
It's always a pleasure to have you on the show.
1:09:00
I'll have your LinkedIn connection information in our
1:09:02
show notes for people check out Looking forward to see you in
1:09:06
EIC. I'm sure I'll see you at
1:09:07
ideniverse in A. Couple months.
1:09:10
And yeah, you can connect with Jim and I on LinkedIn.
1:09:13
Tell us how crazy we are. Tell us you want more or less AI
1:09:16
content, all that stuff. We're always open to to that
1:09:19
feedback. And don't forget to like and
1:09:22
subscribe. And if you didn't like this,
1:09:24
send it to somebody you hate. That's fine too.
1:09:25
If they can listen idacpodcast.com for all the
1:09:29
information, especially all of our discount codes we've got for
1:09:31
the for the conferences that we have.
1:09:33
So yeah, that's it. Thanks everybody for watching
1:09:36
and or listening and we'll talk with you all in the next one.
1:09:41
You've been listening to Identity at the Center.
1:09:45
We hope you've enjoyed the show. Make sure to like, rate and
1:09:49
review, and we'll be back soon. But in the meantime, hit the
1:09:52
website at identity@thecenter.com.
1:09:56
See you next time on Identity at the Center.