WEBVTT

00:00:00.859 --> 00:00:05.219
Hey, what's up, Toby? We are at the 16th podcast

00:00:05.219 --> 00:00:10.179
on process debt. I love it. Today, we are going

00:00:10.179 --> 00:00:13.240
to go a little contrarian and we are going to

00:00:13.240 --> 00:00:19.079
talk a little bit about how generative AI might

00:00:19.079 --> 00:00:23.600
lead to degenerative human intelligence. That's

00:00:23.600 --> 00:00:26.379
the topic. That's not what's been promoted. Hey,

00:00:26.420 --> 00:00:28.359
let me interrupt you before we jump into this.

00:00:29.180 --> 00:00:32.020
in degenerative discussion around degenerates.

00:00:34.000 --> 00:00:36.619
You know, one of our listeners from the middle

00:00:36.619 --> 00:00:39.500
of the Atlantic Ocean, which may be a cruise

00:00:39.500 --> 00:00:42.520
ship, maybe a working ship, it could just be

00:00:42.520 --> 00:00:46.960
bad Google Maps, but if they could ask a question

00:00:46.960 --> 00:00:50.060
from the middle of the ocean out there, my recent

00:00:50.060 --> 00:00:52.280
thought is they would be asking, how in the hell

00:00:52.280 --> 00:00:55.399
do I find you guys? Ah. I stumbled on you once.

00:00:56.020 --> 00:00:58.990
Yes. So I don't know how these people find us.

00:00:59.149 --> 00:01:03.070
Besides just good luck, you know, being in the

00:01:03.070 --> 00:01:05.629
right crowd. So there you go. Besides carrying

00:01:05.629 --> 00:01:07.569
your pigeons, how else are people accessing this?

00:01:07.629 --> 00:01:12.849
All right. We we are we basically launch our

00:01:12.849 --> 00:01:16.629
podcast Fridays at 5 a .m. Mountain because we're

00:01:16.629 --> 00:01:19.030
out of, you know, the beautiful Boise, Idaho.

00:01:19.629 --> 00:01:23.489
And you can search us. Just look up pop process

00:01:23.489 --> 00:01:27.269
debt on Spotify. We're on Amazon Music. Also

00:01:27.269 --> 00:01:29.290
founded in Audible, which I was very excited

00:01:29.290 --> 00:01:33.090
because I am a total Audible nerd. We were on

00:01:33.090 --> 00:01:39.150
Apple Podcasts, Pandora, and Deezer. Those are

00:01:39.150 --> 00:01:41.310
all the places that you can look us up. Just

00:01:41.310 --> 00:01:43.849
do a search. It's so easy. You'll finally came

00:01:43.849 --> 00:01:47.739
through with even Deezer? Our license deal. Thank

00:01:47.739 --> 00:01:51.200
you. We're official because I bought a microphone

00:01:51.200 --> 00:01:54.219
and we're losing money. That's why we're official

00:01:54.219 --> 00:01:56.939
podcast. I don't want to be the guy that that

00:01:56.939 --> 00:02:01.299
drops names, but my friend, my my casual acquaintance,

00:02:01.400 --> 00:02:04.079
Joe are invited us to come on in. And I'm like,

00:02:04.099 --> 00:02:07.040
oh, yeah, we're kind of a big deal right now

00:02:07.040 --> 00:02:10.460
in northern Canada. Yeah. And I don't think you

00:02:10.460 --> 00:02:14.580
can afford us on your little show. So. Yeah.

00:02:14.819 --> 00:02:17.620
Well, hey, fantastic. So find us on your favorite

00:02:17.620 --> 00:02:22.400
podcasting platform. If you can't find us, you

00:02:22.400 --> 00:02:25.039
know, open up a window and yell at your neighbors.

00:02:25.199 --> 00:02:28.280
Or just or go to LinkedIn and hopefully figure

00:02:28.280 --> 00:02:32.159
out how to automate my posts. So when it comes

00:02:32.159 --> 00:02:33.900
out, then you can get a link and you can listen

00:02:33.900 --> 00:02:37.159
there. That works, too. All good options. If

00:02:37.159 --> 00:02:40.639
they had if they had some some automation. Tools

00:02:40.639 --> 00:02:43.039
that they could use to slurp awesome podcasts

00:02:43.039 --> 00:02:44.500
out of linkedin. That would probably be there.

00:02:44.500 --> 00:02:47.479
Yeah, there you go. I just need to set that up.

00:02:48.659 --> 00:02:50.460
I think that's a consumer problem. I don't think

00:02:50.460 --> 00:02:55.060
that's Hey, anyway, let's go back degenerates

00:02:55.060 --> 00:02:57.099
using ai. Is that what you wanted to talk about?

00:02:57.400 --> 00:03:02.180
Well, I That's a bold brushstroke. Yeah, it is

00:03:02.180 --> 00:03:04.020
a big brushstroke. All the dependents are just

00:03:04.020 --> 00:03:08.300
some of them. No, I'm just saying like we I think

00:03:08.300 --> 00:03:11.590
we underestimate humans. That's really the point.

00:03:12.469 --> 00:03:16.250
Okay, that's a pretty strong position to take.

00:03:16.449 --> 00:03:20.650
Tell me more. Generative AI, LLM specifically,

00:03:21.030 --> 00:03:23.389
are very good at summarizing, they're very good

00:03:23.389 --> 00:03:26.509
at classification, they're very good at telling

00:03:26.509 --> 00:03:31.349
stories, which we talked about last time, and

00:03:31.349 --> 00:03:39.599
I think by getting lazy and not thinking, We're

00:03:39.599 --> 00:03:43.960
just we're going we're we become human doings

00:03:43.960 --> 00:03:48.919
instead of human beings, right? Yeah, and so

00:03:48.919 --> 00:03:51.319
like I've just I've thought about that like like

00:03:51.319 --> 00:03:55.180
You know, I think you know thought a lot about

00:03:55.180 --> 00:03:58.939
like getting reports and getting analyzing things

00:03:58.939 --> 00:04:03.439
and thinking about stuff and We are really good

00:04:03.439 --> 00:04:06.960
in the weeds and figuring stuff out like we're

00:04:06.960 --> 00:04:09.919
really good Like our bodies are fascinating.

00:04:10.900 --> 00:04:14.580
And we have this generative AI thing. Yeah. And

00:04:14.580 --> 00:04:17.240
I wonder if we're just going to get lazy and

00:04:17.240 --> 00:04:22.399
not think. Well, when cell phones came out, I

00:04:22.399 --> 00:04:27.600
got awfully lazy, awfully fast. Oh, yeah. Excuse

00:04:27.600 --> 00:04:32.500
me. Toby is sneezing as we speak and doing play

00:04:32.500 --> 00:04:36.360
by play. Right. There he is, lines up over the

00:04:36.360 --> 00:04:42.779
ball. He's back. I think if I still have memorized

00:04:42.779 --> 00:04:45.100
more than a dozen phone numbers, I would be shocked.

00:04:45.939 --> 00:04:48.480
Oh, it's crazy. You know, and that was just,

00:04:48.480 --> 00:04:52.199
that was the advent of just mobile repository,

00:04:52.720 --> 00:04:57.180
mobile Rolodex. Those that are only fresh since

00:04:57.180 --> 00:04:59.920
this new century began, Rolodex was a physical

00:04:59.920 --> 00:05:03.060
thing, right? Little index cards, and you could

00:05:03.060 --> 00:05:07.230
spin it. write dirty things on them. But yeah,

00:05:07.689 --> 00:05:10.649
what if AI increasingly does more for us, right?

00:05:10.689 --> 00:05:13.850
And I think it will obviously do more, which

00:05:13.850 --> 00:05:17.009
hopefully frees us up to solve harder problems.

00:05:18.129 --> 00:05:25.350
And my thought there is that it will allow us

00:05:25.350 --> 00:05:28.709
to not think as much and thinking is where the

00:05:28.709 --> 00:05:34.220
magic is. Chris, you're totally laughing. I don't

00:05:34.220 --> 00:05:35.660
even know what you're laughing. I don't know

00:05:35.660 --> 00:05:38.819
if I'm being funny. Yeah, I'm laughing because

00:05:38.819 --> 00:05:42.019
I'm trying to imagine a world in which in which

00:05:42.019 --> 00:05:44.560
having some machine do your thinking for you

00:05:44.560 --> 00:05:49.540
is a net positive. Like people that are already

00:05:49.540 --> 00:05:52.800
responsible for their own thinking and feel like

00:05:52.800 --> 00:05:55.000
some semblance of accountability for what they

00:05:55.000 --> 00:06:00.259
produce aren't always very accountable. And so.

00:06:01.219 --> 00:06:05.779
this world of more content creation. Think about

00:06:05.779 --> 00:06:08.439
the new wave of visualization tools and data.

00:06:09.079 --> 00:06:11.079
When Tableau first launched, all of a sudden

00:06:11.079 --> 00:06:13.699
I had executives say, well, this is great. We

00:06:13.699 --> 00:06:17.160
don't need analysts anymore. And I'm like, well,

00:06:17.480 --> 00:06:21.120
you know. It's probably true. You could just

00:06:21.120 --> 00:06:25.949
click around and make guesses. And so. We you

00:06:25.949 --> 00:06:28.470
know, is there a is there a natural limit to

00:06:28.470 --> 00:06:31.649
how far artificial intelligence and inference,

00:06:31.670 --> 00:06:36.430
you know Can go it feels like it's a does it

00:06:36.430 --> 00:06:38.910
get into the fuzzy logic or is that kind of this

00:06:38.910 --> 00:06:41.850
next step? Yeah, and I mean, I'm definitely not

00:06:41.850 --> 00:06:44.310
an expert But I know like what's funny is like

00:06:44.310 --> 00:06:47.149
people don't think about what a large language

00:06:47.149 --> 00:06:50.750
model is No, they're like, well, what is it?

00:06:50.769 --> 00:06:56.480
It's I mean and you know, I I have an idea, but

00:06:56.480 --> 00:06:59.800
I'm not a scientist, right? But they spend millions

00:06:59.800 --> 00:07:02.379
of dollars training these things on words that

00:07:02.379 --> 00:07:05.459
are like each other. And it basically communicates

00:07:05.459 --> 00:07:07.639
just like a human does in which it basically

00:07:07.639 --> 00:07:11.920
takes that first word, takes a sentence, it writes

00:07:11.920 --> 00:07:14.160
a word, and then it basically says, well, what's

00:07:14.160 --> 00:07:17.300
the most probabilistic next word? And so it goes

00:07:17.300 --> 00:07:21.550
down this path. In order for it to not sound

00:07:21.550 --> 00:07:24.649
like a human, it basically doesn't always pick

00:07:24.649 --> 00:07:26.810
the top word. Otherwise, it would sound very

00:07:26.810 --> 00:07:29.529
robotic. And so there's this randomness that

00:07:29.529 --> 00:07:31.490
makes it sound unique, but it's not thinking.

00:07:31.529 --> 00:07:36.350
It's just doing statistics. And so I just think

00:07:36.350 --> 00:07:39.769
from a like, are we going to lose? Are we going

00:07:39.769 --> 00:07:42.449
to from a from a business process standpoint?

00:07:43.910 --> 00:07:47.810
Are we going to basically do more activities

00:07:47.810 --> 00:07:52.370
of less value? Because it sounds because it sounds

00:07:52.370 --> 00:07:54.810
pretty or the PowerPoints good. Yeah, it's great

00:07:54.810 --> 00:07:58.529
Or like are we gonna like are we are we saw what

00:07:58.529 --> 00:08:03.230
problem are we solving? Yeah Yeah, Chris, I think

00:08:03.230 --> 00:08:07.449
I love this thought because Already and it's

00:08:07.449 --> 00:08:09.949
probably just me, you know, it may not be the

00:08:09.949 --> 00:08:13.089
entire world But there's a lot of functions I've

00:08:13.089 --> 00:08:15.029
worked with over my career where I've thought

00:08:15.029 --> 00:08:17.670
you know what a tippy bird could easily replace

00:08:17.670 --> 00:08:24.079
you The accuracy of your answers and so If I

00:08:24.079 --> 00:08:26.579
do that quick AI for tippy bird substitution,

00:08:26.660 --> 00:08:29.019
that's like 20 % of these corporate functions

00:08:29.019 --> 00:08:33.820
already. Yeah But yeah, I mean increasingly it's

00:08:33.820 --> 00:08:36.860
sounding very human it can look very humans for

00:08:36.860 --> 00:08:41.080
sure even in a video Session, you know necessarily

00:08:41.080 --> 00:08:44.600
sure it is interesting. So what happens when?

00:08:44.840 --> 00:08:47.360
We have tools that do all of our thinking for

00:08:47.360 --> 00:08:52.460
us Yeah, do we have like we did generally I would

00:08:52.460 --> 00:08:58.379
be a I Think because of the way that I grew up

00:08:58.379 --> 00:09:03.019
Like I would I would that would not be a very

00:09:03.019 --> 00:09:07.820
encouraging life Yeah, yeah, right like like

00:09:07.820 --> 00:09:10.159
give me hard work. Give me something difficult

00:09:10.159 --> 00:09:15.250
to do. That's within within my capabilities Yeah.

00:09:15.509 --> 00:09:18.429
And give me problems to solve. Well, and this

00:09:18.429 --> 00:09:21.750
is where I totally feel that, Chris. I hear you.

00:09:22.110 --> 00:09:24.629
I don't know if it's growing up with that kind

00:09:24.629 --> 00:09:27.649
of puritanical work ethic, you know, with a lot

00:09:27.649 --> 00:09:31.549
of us. Coupled with whether you're Catholic or

00:09:31.549 --> 00:09:34.190
not, there's also a good dose of Catholic guilt

00:09:34.190 --> 00:09:37.090
that accompanies that for me. And so it's like,

00:09:37.169 --> 00:09:39.149
I have nothing to do, but God, should I really

00:09:39.149 --> 00:09:41.450
be laying around on my couch on a perfectly nice

00:09:41.450 --> 00:09:44.580
Sunday afternoon? The answer is always yes, especially

00:09:44.580 --> 00:09:47.399
if golf's on, because it's the best way to take

00:09:47.399 --> 00:09:51.720
a nap. But to you, stepping back to our overarching

00:09:51.720 --> 00:09:55.759
theme, what happens when people are disassociated

00:09:55.759 --> 00:09:58.200
or disenfranchised from the work that they're

00:09:58.200 --> 00:10:01.320
accountable for, because now they have tools

00:10:01.320 --> 00:10:03.779
doing everything for them? Well, I think what

00:10:03.779 --> 00:10:08.080
happens is you just end up being farther away

00:10:08.080 --> 00:10:13.970
from the customer. And the more abstracted you

00:10:13.970 --> 00:10:18.149
are away from the customer, the more you're kind

00:10:18.149 --> 00:10:23.710
of making stuff up. And that's where the solutions

00:10:23.710 --> 00:10:27.110
are. That's where the problems are solved. That's

00:10:27.110 --> 00:10:29.649
right. Well, so let me take that a step further

00:10:29.649 --> 00:10:32.929
then. So if our goal is to continue to have not

00:10:32.929 --> 00:10:35.509
just generative tools, but generative people.

00:10:35.799 --> 00:10:38.940
They can think on their own. They can take a

00:10:38.940 --> 00:10:42.019
problem and formulate a hypothesis and a solution.

00:10:43.200 --> 00:10:45.460
Would it be fair to say that the motion we need

00:10:45.460 --> 00:10:47.379
to be considering is how do we keep ourselves

00:10:47.379 --> 00:10:50.580
close to customers and then put AI between us

00:10:50.580 --> 00:10:54.919
and our repositories? Because I see AI getting

00:10:54.919 --> 00:10:57.720
injected between us as humans more often right

00:10:57.720 --> 00:11:02.340
now. Yeah, I think, you know, obviously AI is

00:11:02.340 --> 00:11:04.980
going to be really great. Like to be able to

00:11:04.980 --> 00:11:08.539
surf to be able to go and ask a question And

00:11:08.539 --> 00:11:10.820
for it to go and look through a knowledge base

00:11:10.820 --> 00:11:13.320
and provide a solution Yeah, that's gonna get

00:11:13.320 --> 00:11:15.620
better and better and better right like you can

00:11:15.620 --> 00:11:18.720
search. Yeah, like it's it's it's gonna be it's

00:11:18.720 --> 00:11:23.000
gonna be phenomenal but The problem that usually

00:11:23.000 --> 00:11:26.899
they're asking is about design in which the design

00:11:26.899 --> 00:11:33.120
is not good. Yeah That's interesting And so that's

00:11:33.120 --> 00:11:35.860
where the human thing is. It's like, oh, look,

00:11:36.220 --> 00:11:38.980
the AI did such a good job because our design

00:11:38.980 --> 00:11:45.840
is so crappy. No, think about it. Watch the people

00:11:45.840 --> 00:11:50.720
do the work. Watch what happens. Pay attention.

00:11:51.120 --> 00:11:54.639
Go and be curious. And maybe that's what I'm

00:11:54.639 --> 00:11:58.000
getting at is I think people that really care

00:11:58.000 --> 00:12:01.000
about process and keep process debt to a minimum,

00:12:01.309 --> 00:12:04.190
are just ridiculously curious because they're

00:12:04.190 --> 00:12:07.929
always asking, why do we do this? And they're

00:12:07.929 --> 00:12:10.669
not saying, well, what's the solution to the

00:12:10.669 --> 00:12:13.389
problem? Well, we should click the button faster.

00:12:13.450 --> 00:12:15.610
You're like, no, why do we even click the button?

00:12:16.210 --> 00:12:21.570
Like, it's the loss of curiosity. And I think

00:12:21.570 --> 00:12:27.149
the other thing that AI is going to cause a problem

00:12:27.149 --> 00:12:32.149
is just the lack of practice. Yes. And you're

00:12:32.149 --> 00:12:39.389
not gonna be able to differentiate between, like

00:12:39.389 --> 00:12:41.570
it's gonna be hard to differentiate between talent.

00:12:43.850 --> 00:12:47.090
Because talent can hide behind it, right? That's

00:12:47.090 --> 00:12:50.450
right. And you're experts, there's gonna be a

00:12:50.450 --> 00:12:53.929
perception of a smaller gap between your two

00:12:53.929 --> 00:12:57.340
superstars and someone who's just loafing. It's

00:12:57.340 --> 00:13:01.019
the Dunning -Kruger, right? Yeah, yeah. Strong

00:13:01.019 --> 00:13:04.059
presentation of a limited fact set gives the

00:13:04.059 --> 00:13:06.519
appearance of expert, but it takes more time

00:13:06.519 --> 00:13:09.480
to really discern whether that person has substance

00:13:09.480 --> 00:13:13.539
or they just, they read the covers of all the

00:13:13.539 --> 00:13:16.500
books in the library and call it good. Yeah,

00:13:16.500 --> 00:13:17.899
and I think that that's the thing that scares

00:13:17.899 --> 00:13:20.480
me the most about the LLMs is like they just

00:13:20.480 --> 00:13:24.679
tell a confident story. And you know who's susceptible,

00:13:24.860 --> 00:13:29.820
I can't even say the word, susceptible to confident

00:13:29.820 --> 00:13:33.200
stories? The humans. Half of our voters. Yeah.

00:13:33.399 --> 00:13:43.980
Half of our voters. The non -generative half

00:13:43.980 --> 00:13:47.580
of our voters. Well, I mean, the reality is like,

00:13:47.919 --> 00:13:49.779
we want it to be simple and we want to be able

00:13:49.779 --> 00:13:53.019
to understand it. And you know what? the LLMs

00:13:53.019 --> 00:13:57.000
and the chat GPTs are really good at is simplifying

00:13:57.000 --> 00:14:01.059
it. So that there's a perception of expertise,

00:14:01.879 --> 00:14:08.779
even though you might just be confirming someone's

00:14:08.779 --> 00:14:12.299
pre -existing bias. Which is always something

00:14:12.299 --> 00:14:15.029
that we're open to hearing. Yeah. Oh, you think

00:14:15.029 --> 00:14:18.470
I'm looking younger and more fit? Of course I

00:14:18.470 --> 00:14:23.250
am. Of course I did. That's funny. I'm personally

00:14:23.250 --> 00:14:26.509
trying to be less defensive. Like, what is my

00:14:26.509 --> 00:14:28.409
ego? Why does my ego care so much? Like, why

00:14:28.409 --> 00:14:31.889
is it so important? God, my little snowflake

00:14:31.889 --> 00:14:35.250
of an ego, you know. You're right. You're right.

00:14:35.629 --> 00:14:38.549
Like, Chris, this is a great conversation. And

00:14:38.549 --> 00:14:41.850
this idea, too, about curiosity. Yeah, whether

00:14:41.850 --> 00:14:45.450
AI is good bad or other yeah that innately human

00:14:45.450 --> 00:14:48.450
thing that is Curiosity that an answer isn't

00:14:48.450 --> 00:14:50.870
just the end of the conversation. I think it's

00:14:50.870 --> 00:14:53.730
a unique trait So well, I think some of it too

00:14:53.730 --> 00:14:57.210
is like if you partner with the AI Yeah, I mean

00:14:57.210 --> 00:14:59.090
you could do some pretty stinking cool stuff.

00:14:59.389 --> 00:15:01.970
That's right So how many scared of the cotton

00:15:01.970 --> 00:15:05.149
gin? Yeah discovered that it would take your

00:15:05.149 --> 00:15:08.049
job Yeah, and I think I think the key here is

00:15:08.049 --> 00:15:12.590
like, you know just for you know for this Friday

00:15:12.590 --> 00:15:17.389
is specifically just being curious. Find a process

00:15:17.389 --> 00:15:21.610
and just ask why. I think sometimes we're afraid

00:15:21.610 --> 00:15:24.590
to not do something because we might see what

00:15:24.590 --> 00:15:26.409
happens. Well, what happens if I don't get that

00:15:26.409 --> 00:15:29.490
report out? Sometimes no one's watching the report.

00:15:30.529 --> 00:15:35.629
And it's just being intellectually curious and

00:15:35.629 --> 00:15:37.769
really figuring out what are the real problems.

00:15:38.319 --> 00:15:40.379
I love it. Yeah, if you don't have a restraining

00:15:40.379 --> 00:15:44.779
order with a minimum Feet setback be curious

00:15:44.779 --> 00:15:47.840
ask questions. Yeah, open that door grab the

00:15:47.840 --> 00:15:51.220
door before it shuts You know that TSE doing

00:15:51.220 --> 00:15:53.080
something back there. That's probably really

00:15:53.080 --> 00:15:55.879
interesting. There you go I don't know. Sometimes

00:15:55.879 --> 00:15:57.279
I don't know if I'm always picking up what you're

00:15:57.279 --> 00:15:59.860
putting down, but yeah, be curious. I like that.

00:15:59.860 --> 00:16:03.820
Just just be curious All right, man. All right

00:16:03.820 --> 00:16:06.080
We'll talk to you later. Happy curious Friday.

00:16:06.139 --> 00:16:07.700
All right. See you man. See ya