Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:05
Today, we're talking all about artificial
0:08
intelligence. Over the past decade or so,
0:10
there's been a slow build in public interest as
0:12
whispers spread about the technology finally
0:14
making massive leaps. But with
0:17
the release of ChatGPT last November,
0:20
well, the excitement and
0:22
alarm about the power of this technology
0:25
has reached something of a fever pitch. In
0:27
our new three-part series, we were interested
0:29
in the moves and maneuvers that help catapult
0:32
AI forward. We're tracking the way
0:34
the big tech companies, specifically Google,
0:36
Microsoft, and Meta, are acquiring
0:38
and developing AI technology, and we'll be getting
0:40
into the ethical quandaries they face
0:42
too. But today, we're taking a look
0:44
at the way AI has affected three industries,
0:47
climate change management, fast food,
0:50
and advertising, both the positive
0:52
and worrisome effects of letting computers
0:54
make choices for us. Here
0:56
to talk about all that and more is journalist Haya
0:59
Camps and comedian Kenny DeForest.
1:02
Stick around for the best of Business Wars
1:04
Daily.
1:09
The Spark Cash Plus card from Capital One
1:11
helps you earn unlimited 2% cash
1:13
back on every purchase.
1:15
And with no preset spending limit, your purchasing
1:18
power can adapt to meet your business
1:20
needs. Jorge Gaviria, founder
1:22
of Macienda, reinvests his 2%
1:25
cash back to help grow his business with
1:27
new products. Macienda is
1:29
a masa supply chain company connecting
1:31
and elevating traditional Mexican heirloom
1:34
corn farmers with restaurants, chefs,
1:36
and consumers.
1:37
Wow, I just love hearing
1:40
about companies like Macienda that have
1:42
an important mission behind their work. Now
1:44
imagine what the Spark Cash Plus
1:46
card from Capital One and unlimited 2%
1:48
cash back on every purchase could
1:51
do for your business.
1:53
Capital One, what's in your wallet?
1:55
Find out more at CapitalOne.com
1:58
slash Spark Cash Plus. terms
2:01
and conditions apply. Millions
2:03
of small business owners like Leanna Lynn
2:05
choose Spectrum Business. Spectrum Business
2:07
services are essential to
2:10
our business. Call 855-282-FAST
2:12
to choose Spectrum Business for your business.
2:14
Get fast, reliable internet starting at $49.99 per
2:16
month and the Spectrum Business
2:19
Connect phone system with video conferencing for
2:21
just $19.99 per month to get a free
2:23
mobile line for one year. I cannot live without it.
2:25
Call
2:25
855-282-FAST to choose Spectrum
2:29
Business today. Restriction supply services
2:31
not available in all areas.
2:40
From Wondery, I'm David Brown and this is
2:42
the best of business wars daily. We
2:44
talk about trends, controversies and how
2:46
they affect you, the consumer. And
2:49
joining us is Kenny DeForest and Haya
2:51
Camps. Thanks so much for
2:53
speaking with us on the best of business wars daily. Thanks for
2:56
having us. Yeah, of course. My pleasure to be
2:58
here. Thank you. Kenny, I understand you just released
3:00
a new comedy album in a special called Don't
3:02
You Know Who I Am? Let's take
3:04
a listen to a clip from your standup.
3:07
Basically what generalizing is,
3:10
I would say, if I could describe it,
3:12
my brain is a bad
3:14
detective. That's what it is. That's how I would explain. I
3:16
got a really confident bad cop living
3:20
in my head. And
3:21
he looks at all the correct evidence and
3:24
then just makes me wildly wrong completely.
3:28
So every day he comes in just all cocky, hot
3:30
cup of joe, just like, alright fellas, let's take a look at this
3:32
case. What do we have here? Okay. Interesting.
3:35
It appears when you enter a room, people tend
3:37
to react warmly. Alright. They're
3:40
often smiling. They say nice things about
3:42
you personally and professionally. Not
3:44
to mention there's always an outpouring of support
3:46
when times get tough. Open and shut case,
3:49
everybody hates you. Everybody
3:50
hates you. And
3:55
it's the only thing they talk about when you're not around. I
3:58
know, it's crazy.
3:59
Do you believe this? They actually get together at coffee shops
4:02
and talk about what a piece of s*** you are. And
4:05
the only way to get him to stop would be if you message
4:08
each of them individually on Facebook, preferably
4:11
at 3 in the morning, and apologize
4:13
for things that happened over 15 years ago
4:15
that only you remember. And I'm like, right on, detective.
4:19
Danny, tell us
4:21
more about this special. What's the focus? Yeah,
4:23
so it's called Don't You Know Who I Am, which
4:25
is a little bit of a play on the closing joke, which
4:27
you will get if you watch it. But
4:30
it basically, I had a friend that
4:33
passed away during COVID and
4:36
he had struggled with drugs and I grew up playing basketball with him. And
4:38
I started talking to friends and realizing
4:41
that there's this sort of post-sports
4:43
drop off that a lot of athletes experience
4:46
that no one really talks about. And
4:48
I definitely experienced it. You know, you
4:50
go from sort of this daily serotonin hit playing ball
4:53
and you don't realize that it's more than a game. It's
4:56
like it's how you deal with stress. It's so
4:59
much of your identity and
5:01
it's just gone. And I think a lot
5:03
of athletes have a dark period. And
5:06
a lot of times, you know, drugs, alcohol, things
5:08
of that nature are kind of used to fill the void. And
5:10
I felt like maybe this is an interesting thing for
5:12
me to talk about in hopes that
5:14
it'll be relatable to other people that have experienced
5:17
the same thing. So it's kind of that. The
5:19
special opens with a story of
5:22
me getting dominated on
5:24
the basketball court and it ends with a story of redemption.
5:27
And then the middle is kind of my
5:29
struggle of putting it all to bed and moving
5:31
on. I can't wait to check it out.
5:33
That sounds really interesting. You're
5:35
a reporter at TechCrunch, but you've also launched
5:38
multiple companies. You're in a book, directed
5:41
a venture capital firm. But
5:43
I understand on the down low you
5:46
have a passion of photography.
5:48
Have you caught anything in your lens this summer?
5:52
Yeah, no, I love photography. My
5:54
favorite thing continues to be people. You
5:57
know, you can take beautiful landscapes. You can do beautiful all
5:59
sorts of things.
5:59
But people have this, like
6:02
we are trained as humans to read other
6:04
people in a way unlike anything else.
6:07
And capturing somebody's
6:08
essence or soul or beauty is,
6:12
it fills me with a joy unlike anything else.
6:15
And yet you cover technology. That
6:19
interesting life choice. Yeah. Yeah. I'm
6:23
guessing AI has been on your radar. There's
6:26
been a lot of AI stuff. Yeah, for sure. It's
6:28
a red-hot industry right now. But
6:31
actually I've been looking a lot at founders that
6:33
look unlike other founders. It
6:35
turns out that,
6:37
like I've been kind of tracking two trends.
6:39
One is that a lot of very successful
6:41
founders tend to be older than
6:43
the narrative seems to support,
6:46
right? The media loves a story about a Zuckerberg
6:49
or somebody who just like quits college
6:51
and becomes a multi-billionaire. The wonderkind,
6:53
right? That's sort of the, yeah. Exactly.
6:56
But it turns out that a lot of really good
6:58
founders have a lot of experience. And
7:00
that experience you get by being in an industry
7:03
for a long time,
7:04
right? It's like if you're a fresh journalist,
7:06
you know how to write. You have nothing to write about. You haven't
7:08
lived. You haven't been anywhere. So the
7:10
same thing for companies, right? If you have been, if
7:13
you're the guy who's been talking to everybody
7:15
in that industry for 20 years, you
7:17
know everybody, you know everybody who might buy something, you
7:19
know everybody you might hire, that's
7:21
a very different circumstance than
7:24
a fresh-faced investor.
7:27
Yeah. I guess there's something of a parallel
7:29
in that AI is considered a relatively young field,
7:32
but it's already being put to work dealing with some problems
7:34
we've been wrestling with for ages. And in
7:36
the case of something like wildfires, they
7:39
seem to be intensifying certainly over the past decade.
7:42
And up until recently, fighting wildfires
7:44
has largely relied on human observation and judgment,
7:47
but not anymore. Let's check out a clip
7:49
from a recent episode of Business Wars Daily.
7:53
A slate of companies have cropped up in recent years,
7:56
promising more efficient wildfire detection
7:58
with, you guessed it,
8:01
One of the companies is Pano AI. Last
8:04
year it raised $20 million and pulled
8:06
in an additional $17 million this year.
8:09
Pano works by mounting high-def remote
8:11
control cameras in areas at high risk
8:13
for wildfires, according to TechCrunch.
8:16
The cameras can be controlled manually or rotate
8:19
automatically, scanning the landscape around
8:21
them. The video footage is then checked
8:23
out by AI which looks for signs
8:25
of smoke or fire, and if it detects
8:28
fire, an alarm goes off. Pano's
8:30
CEO, Sonia Kastner, told
8:32
TechCrunch that humans will double-check the AI's
8:35
findings before alerting nearby fire
8:37
stations, cutting down on potential
8:39
false alarms. But Pano appears to
8:41
be doing its job. The company said it
8:43
had detected a fire in Oregon earlier
8:45
this month, a full 14 minutes before
8:48
the first 911 call. That's
8:50
valuable time saved, but of course, this
8:53
technology will cost you. A fire
8:55
station hoping to license Pano's technology
8:57
will have to pay $50,000. But
9:01
fire stations aren't the only customers Pano
9:03
is hoping to target. Power utilities
9:05
and private landowners like ski resorts
9:07
also use Pano's technology. And
9:10
of course, like any emerging market, consumers
9:13
have options here. FireScout
9:15
is another player in the game, also relying on
9:17
high-def cameras and using AI to scan
9:19
the images. FireScout says it detected
9:22
a 59-acre fire in California
9:24
at the end of June, only several minutes
9:26
after it started.
9:29
Hi, I heard you got a demo
9:31
from Pano AI. What was that like? It's
9:34
really interesting, right? As a consumer,
9:37
you're used to very slick software. Like if
9:39
you use Nest or anything like that, everything
9:41
is slick, everything is easy. And
9:43
the demo, at first I was like, is this
9:45
it? Really, honestly, is this it?
9:48
I was not super impressed. And
9:50
I thought about it for a moment. And I was like, actually, the
9:53
alternative to this is somebody sitting high
9:56
up in a tower and being on a lookout, right?
9:58
I've done fire lookouts in the past. like just
10:00
helping out a friend. And it's kind
10:02
of wild how
10:05
even relatively simple technology, like the actual
10:07
camera technology is off the shelf and super
10:09
simple. The thing that is advanced is
10:11
the AI that actually can look at this
10:14
and say, hey, that that change in
10:16
the landscape since I scanned a minute ago, what
10:18
is that? Is that a cloud? Or is
10:20
that something sinister? And the actual
10:23
cool thing is because they have most of the area covered
10:25
by more than one camera, they can do extremely
10:27
exact triangulation. So they see it
10:29
from one camera, they see it from another. That
10:31
means you can calculate the exact GPS coordinates.
10:34
Like if I'm up in a tower, I can say, well, it's kind
10:36
of to the east, right? And somebody else can go,
10:38
oh, yeah, it's kind of to my north.
10:41
But that means you have an enormous area where this fire
10:43
could be. And you still don't know if it's a real
10:46
fire or not. So you have to send somebody out there. And
10:48
by the time you've done that, you know, half an hour
10:50
has gone by, maybe an hour has gone by. So
10:52
by having much more accurate cameras
10:54
that can zoom much further than you can with a pair of binoculars
10:58
and accurate triangulation, it
11:00
means that you actually get a little bit of a technology edge.
11:02
And this is a world where, you know, a small
11:04
fire, one helicopter can put out a big fire
11:07
that is
11:08
potentially lots and lots of lives lost in
11:10
a huge resource drain.
11:12
Yeah, I certainly see
11:14
the advantages there. On the
11:16
other hand, you were talking a lot about cameras
11:19
and zooming in. I can
11:22
imagine other things getting caught in the crosshairs
11:24
of those cameras. Any concern about
11:26
AI? I don't
11:28
want to say being too good here, but
11:30
maybe even going rogue.
11:32
Yeah, it depends a little bit what you mean
11:34
by going rogue. So these cameras, like any sort
11:36
of buildings and any places
11:38
where people are pixelated out. So
11:40
even as an operator, you can't look there, essentially.
11:43
So if that happens to catch on fire, that's a different
11:45
problem. Well, let's just pull back
11:47
for just a moment. Kenny, I think I have
11:49
to ask, do you have any kind of video
11:52
security, for example? No,
11:54
I live in an apartment in New York City. So it's
11:57
beautiful. Yeah, it's beautiful.
11:59
what's going on out there. It's
12:02
a beautiful thing about a one bedroom, not a lot of ground to cover.
12:05
But I have had a ring system before. I had
12:07
an ex that had a ring system at her place
12:10
and I liked it, but
12:13
at the same time, the alerts
12:17
became too much. And so what ended up happening is I
12:19
would mute the alerts and then I would basically just not
12:21
check the ring because it would
12:23
alert me if a squirrel ran by. And
12:26
at some point I was like, I don't want to look
12:28
at my phone more. So I'm going
12:30
to mute these and then I just ended up basically not
12:32
having a ring system. You know what I mean? Yeah, well, too
12:34
many false alarms certainly get that. But also,
12:36
I think there are a lot of people who are really
12:39
uncomfortable about the ubiquity of
12:42
cameras. I mean, you never really know
12:44
where one might be and the
12:46
lenses can be made so tiny that
12:49
I mean, what is off limits, what
12:51
isn't. And the flip side of that,
12:53
it seems to me, Kenny, is it
12:55
we become sort of inured. We become
12:57
sort of used to this
12:59
level of constant surveillance.
13:02
Do you ever
13:02
get that feeling that you're being watched? I
13:05
mean, yeah, it's everywhere. And the part that's
13:07
scary for me about this, and I was saying this to
13:09
Haya before we started, I'm excited to hear somebody
13:11
who has enthusiasm for the technology because all
13:14
I feel is fear. It's just 100% fear. I grew up watching, I'm like,
13:18
have you people not seen Robocop? Have we not
13:20
watched Minority Report? We're just marching
13:23
towards this, to me,
13:25
this end that's been prophesied since before
13:28
the technology even was close to
13:30
existing. You know what I mean? Philip K. Dick was in the 50s being
13:32
like, this is probably going to be bad. And
13:35
here we go because we're just, we're
13:37
humans. We can't help ourselves. We're little lemmings marching
13:39
towards the cliff. And that's how I feel. And
13:43
it's kind of that old idea, like
13:45
how much privacy or liberty
13:47
do you give up for security? And that's kind
13:49
of the paradigm that we're messing with
13:52
right now. And the facial
13:55
recognition stuff scares me. It's like
13:57
the idea of going cashless. I'm like, what done
14:00
anything illegal before? Don't you want to be able to peel off
14:02
a 20 and not have anyone know why you did it? I
14:04
mean, what are we doing? You know what I'm saying? I don't
14:06
need everyone to know where I am all the time. I
14:08
think that's what the blockchain is for now. Yeah, yeah, there
14:10
you go. Great. That's right. Yeah. Well,
14:14
I was thinking about, you know, you add AI
14:17
to that mix. And it's one thing if somebody can
14:19
see something or make out something through a
14:22
sophisticated camera in Zoom. But
14:24
once you enter data and you're,
14:26
you know, and you're
14:29
able to make the kinds of correlations
14:31
that most human beings wouldn't
14:33
bother to do,
14:35
I wonder, and
14:36
that seems
14:37
to get to the problem, right? That we,
14:39
in a sense, don't know
14:42
what
14:43
to be afraid of. But we have this
14:45
looming sense that the power here
14:48
could really turn on us. Kenny,
14:51
do you think that's fair, Kenny?
14:53
Yeah, well, oh, without a doubt. I mean,
14:55
I had the same thing with the ring. What
14:57
I know, I think it makes us more afraid.
15:00
Like, I don't know if you've heard of the Citizen
15:02
app, but somebody when I was, you know, in New
15:04
York was like, you should download the Citizen app. And I was like,
15:06
man, here's what I learned. If I
15:08
know what's going on outside of New York City, I'm never going
15:10
to go outside because it's insane. Like,
15:13
there's almost like too much information. Now, obviously,
15:16
you know, ignorance is bliss, whatever,
15:18
but you also want to protect yourself. I get it. But
15:21
like, when I would scroll the ring thing, like looking
15:23
through the neighborhood, it just made me feel
15:25
like I lived amongst a bunch of paranoid,
15:27
terrified people. And it would be like,
15:30
is this a package thief? And then you look at the video
15:32
and it's clearly just like a teenage kid
15:34
walking by, you know, and everybody's just
15:36
now everybody's like a potential criminal.
15:39
And I was just like, it was infecting my
15:42
brain in a way where I was like, this doesn't feel worth it, man.
15:44
I don't know. Like, lock your
15:46
doors, have
15:48
a floodlight or, you know, one
15:50
of the alarm systems that just makes a lot of noise.
15:52
I mean, that's more of a deterrent anyway. And
15:55
what are we gaining by putting cameras everywhere?
15:58
Is it really a safer society? I mean, it's... I
16:00
don't know.
16:01
Hayao, what do you think for Pano AI?
16:03
Do you think that there's something of a backlash
16:05
that's sort of pushing against the development
16:07
of this technology, or what
16:09
do you see as a future for the company?
16:11
Honestly,
16:14
I hear everything everybody
16:16
says as scary about
16:18
AI, but at the same time, you can't
16:21
slow the tide of the oceans, right? And
16:24
I think it's one of those things where if somebody says to
16:26
a fire department, hey, for $50,000
16:28
a year, instead of people who are in watchtowers,
16:31
maybe six hours every now and again, you have
16:34
24-hour alerts if something goes
16:36
on. But no
16:38
fire chief in their right mind is going to say no to that.
16:41
And so the technology exists. And at that point, it's like, well,
16:44
we can't shove this toothpaste
16:46
back in the tube. And now I think we're into
16:48
mitigation. How do we make this less harmful
16:50
to people? If you
16:52
like to do naughty things in the forest, how do you make sure
16:55
that that doesn't get caught on camera? That kind of thing,
16:57
right? Right, right, right. But that's a different conversation,
16:59
I think.
17:01
Well, and to jump on what you just
17:03
said, I've been obviously so far
17:05
on this episode sort of more anti, but here's
17:08
the truth too. We are entering
17:10
a part of our development as a society
17:12
where we have some really big challenges and we have
17:15
a massive population and dwindling
17:18
resources. And obviously,
17:20
these are powerful tools. I mean, I lived in California
17:22
for five years. I mean, yeah,
17:25
if you could put out a fire before
17:27
it gets you, I mean, the amount that
17:29
that saves, the environmental impact of
17:31
that, you know, is massive. So
17:34
obviously, we are entering a
17:36
part of human history where we need help. We got
17:38
to figure some stuff out. And so
17:41
if we can have these technologies help
17:43
us, and as long as we can do it in a way that's safe,
17:45
like, obviously, we have to explore
17:47
it. And like you said, the toothpaste is out of the tubes. So at
17:49
this point, let's, I guess,
17:51
try to embrace it and make sure that we're regulating it
17:53
in a way that we can still be naughty in the
17:55
woods. You know what I'm saying? Right.
17:58
Yeah, this can seem some serious. I mean, this is...
17:59
pretty heavy-duty stuff and we're
18:02
about to shift to some lighter fare, I promise.
18:04
Stay tuned as we talk about how AI is shaking
18:07
up the fast food industry as
18:09
the best of Business Wars Daily continues.
18:15
Business Wars is brought to you by Volkswagen.
18:18
Making the switch to an electric vehicle can
18:20
be scary. It's tough,
18:22
right? Because part of the allure of EVs
18:25
is how tech-filled they are. But
18:27
sometimes they can feel overly futuristic
18:30
and, let's face it, a little intimidating. Volkswagen
18:33
saw all this and decided to build an accessible
18:35
EV, the all-electric Volkswagen
18:38
ID.4. It has the look and feel
18:40
of a traditional SUV, so it doesn't
18:42
feel like a spaceship, and it features 30 plus
18:45
cubic feet of cargo space and available
18:48
all-wheel drive. Volkswagen was focused
18:50
on getting the little things right in order to enhance
18:53
and simplify your driving experience. With
18:55
Volkswagen, your transition to an EV
18:57
will be intuitive and joyful.
19:00
Visit vw.com to learn more. The
19:02
all-electric Volkswagen ID.4.
19:05
It's a little things. It's a VW. We
19:08
get support from Airbnb. I
19:11
recently stayed in an Airbnb on a trip
19:13
for work and I liked it so much, I
19:15
scheduled a whole family vacation, in
19:18
some ways around the app, because it made
19:20
vacation so easy. Maybe you've had
19:22
a similar experience staying in an Airbnb
19:25
and thought to yourself, you know, this actually seems pretty
19:27
doable. Maybe my place could be an Airbnb
19:29
too. It could be as simple as starting
19:31
with a spare room or your whole place when you're
19:34
away. What I'm trying to say
19:36
is, you could be sitting on
19:38
an Airbnb and not even know it. Maybe
19:40
you're planning a long weekend with friends or family
19:42
this fall. Well, while you're away, you
19:45
could Airbnb your place and make
19:47
some extra money to help pay for the trip. Whether
19:49
you could use some extra money to cover some bills
19:52
or for something a little more fun, your
19:54
home might be worth more than you think. Find
19:57
out how much at Airbnb.com Hey,
20:06
welcome
20:11
back to the best of business wars daily. Joining
20:13
us is journalist Haya camps and comedian
20:15
Kenny DeForest. We're talking about AI
20:18
today and how it's being used in some interesting
20:21
ways, including in the drive-through
20:24
of your favorite fast food restaurant. Let's
20:26
hear a little more about drive-through AI and a clip
20:29
from business wars daily. Many
20:33
chains see it as a great deal according
20:35
to CNN. They say it can ease burdens
20:37
for overworked employees, improve
20:39
order accuracy, and here's the kicker, upsell
20:42
customers better than any old human can.
20:45
Yep, because sometimes a human asking,
20:47
would you like fries with that just doesn't do the trick.
20:51
Presto Automation is one company that's
20:53
established itself in this burgeoning industry
20:55
and upselling happens to be one of its selling
20:58
points. Presto's CEO
21:00
told analysts on a call last month that
21:02
Presto's AI automatically upsells
21:05
every order and those can be tailored
21:07
to the weather, the time of day, the order itself,
21:10
or even the customer's order history according
21:12
to CNN. Restaurant
21:15
operators sure like the sound of that. Last month,
21:17
CKE restaurants, as the
21:19
name behind Hardee's and Carl's Jr., announced
21:21
that it was partnering with Presto to bring AI to
21:24
its drive-throughs. Meanwhile rival
21:27
AI company, High Auto, has
21:29
become another major player in the drive-through
21:32
industry. High Auto's biggest brag? Its order
21:34
accuracy rate, human cash technology.
21:38
That's a jump that could result in more satisfied
21:40
customers and a higher rate of repeat visits.
21:43
High Auto also partnered with the king of fast
21:45
food, McDonald's, to test AI in some of
21:48
its drive-throughs. AI
21:50
can be great for restaurant owners, that
21:52
appears clear, but customers, workers,
21:55
that may be another story according to CNN. The
21:58
technology can be glitchy. allowed
22:00
drive-through isn't the best place for AI to
22:02
pick up on voice cues after all and
22:05
out of 10 AI orders placed at a White
22:07
Castle in Indiana three folks asked
22:09
to speak with a real live human either
22:11
to correct a mistake or because well simply
22:14
wanted that human interaction because
22:17
that's another thing this AI technology lacks
22:19
at least for now warm and
22:21
the ability to have a quick chit chat
22:23
with another person we
22:26
got White Castle Wendy's McDonald's
22:28
Hardee's all serving up
22:31
AI at their drive-throughs what do you make
22:33
of this I'm not sure we need something
22:36
more impersonal when we're getting our food but
22:39
I mean what say you I mean I don't
22:41
know where you go through your drive throughs where where I
22:44
go order accuracy isn't what
22:46
what they're known for to be perfectly honest I was just
22:48
before we started recording I was complimenting Kenny
22:50
on his radio voice and I drove through my
22:52
phones the other day and this this this clearly
22:54
an AI came on was like hello
22:56
are you do you want to order
22:58
blah blah blah today and then he got cut
23:00
off and instead what we met was with like
23:08
right and I was like okay I'm not being funny I'll
23:10
take Kenny taking my order as a robot
23:12
any day of the week over the alternative
23:15
I'm I don't see a downside to
23:17
be perfectly honest if I want to go talk to someone
23:19
to drive through ain't it just
23:22
saying well
23:24
said Kenny have you come into
23:26
contact with AI at the drive through or
23:28
would you prefer face-to-face interaction I
23:32
don't know if this counts as a drive-through but
23:34
I did recently see the robot barista
23:37
in Green Point Brooklyn and
23:40
boy did I not like that I I
23:43
mean here's the thing I get it but here's
23:45
what I think about who is this good for
23:48
people that are already incredibly
23:50
wealthy you're just automating a workforce
23:52
these are jobs when these are jobs
23:55
for people that probably don't have
23:57
a ton of other options and you're just gonna automate
24:00
that. But
24:02
so then what happens eventually? Like,
24:05
so all the farmers are then at some point
24:07
automated. So then the CEO
24:09
of Burger King is just the only one raking
24:12
in money and every everyone else is robot
24:14
staff. And then what they just give us like,
24:16
just enough money every month so that we can continue
24:18
to buy Burger King like I don't see how the
24:21
future and to be honest with you, I
24:23
guess, you know, maybe not the
24:25
comedic relief role that I'm probably here for. But,
24:28
you know, as someone who advocates for
24:30
mental health and for speaking, you know,
24:32
there are a lot of stories out there of people that
24:34
were like,
24:35
you know, I was suicidal and a
24:38
cashier showed me a simple kindness, and
24:40
it made me rethink. And
24:42
that's what I worry about. I mean, the world is already
24:44
getting more and more impersonal. We're getting more
24:46
and more online. And we
24:48
need human interaction. I mean, if there's one thing COVID
24:51
showed us, we are social creatures,
24:53
we are tribal creatures, we need
24:55
to be a amongst our tribe and we need to have
24:58
interactions and not to be isolated. And
25:00
that's what worries me about it. And I mean, who's
25:02
really messing up the order at White Castle?
25:05
It's sliders and fries. That's what they ordered.
25:09
But the big thing here, and this gets to sort
25:11
of part of the business model, it's not just
25:13
that you lay off a lot of workers, it's upselling.
25:16
And apparently robots, we
25:19
are told, are better at it than humans.
25:21
Although I have to say, Kenny, I would
25:23
have a tougher time saying no to
25:26
an upsell from a robot than a real
25:28
human being. So does
25:31
that sound right to you? No.
25:33
I look a robot right in the eyes and
25:35
say, you don't affect me, sir. Yeah, shove it up your
25:40
tin can pal. But I say, take
25:43
those fries. Hya,
25:47
what are you hearing about this business approach?
25:49
Or is everyone sold on the idea that this
25:51
is really going to be good for business?
25:53
Well, the truth is, you know, every
25:56
drive through employee now
25:58
is trained to do upselling anyway, right? And
26:00
most of the time they don't, they just want to get the job done.
26:03
And so I think there's this interesting thing. I think what they're
26:05
trying to do is trading the
26:07
amount of extra time it takes to try and do an upsell
26:10
with, and then they trade that off with order accuracy.
26:12
And so hopefully it takes the same amount of time, but
26:14
they make more money. I think that's the sell. But
26:17
to Kenny's point, and I think this is actually a really, really important
26:20
one, like whose jobs are getting replaced.
26:22
There's so many people I know who had their first job
26:25
in fast food. I worked in retail as my first
26:27
job, right? But the really scary ones
26:29
to
26:29
me is trucking. Like self-driving
26:32
cars and trucking is
26:34
going to completely change the
26:37
landscape. Almost 6% of every
26:39
working adult in the US works in trucking,
26:42
and all of those jobs are about to go away. Now,
26:45
is it safer to have a self-driving truck?
26:48
Maybe. Is it more efficient for the truck
26:50
operators? Probably. But what
26:52
are you going to do with that 6% of workforce? And
26:55
so you're seeing a massive shift of wealth from
26:57
people who already had a lot of money, away
26:59
from people who didn't have a lot of money, combined
27:02
with a lack of safety net, no
27:04
mental health care. And if you lose your job, you have
27:06
no normal health care either. Like I know
27:08
I'm meant to be the pro side of this argument,
27:10
but it's looking pretty bleak over here.
27:13
Well, nobody's in any corners
27:15
here. Feel free to just sound off as
27:17
y'all are doing. I love it, because this is,
27:20
I think, bringing to the surface a lot of these issues
27:22
that we haven't had a moment to sort
27:24
of pause and think about. But
27:26
I also think that maybe this deserves just a
27:28
wee bit more scrutiny, Hyah, because if
27:31
we're talking accuracy, chat
27:33
GPT itself, that was the introduction
27:38
for a lot of folks, the introduction to
27:40
AI. Well, it certainly
27:42
is not the most accurate source.
27:44
And in fact, if you go on and ask
27:48
chat GPT as I did today, a fairly
27:50
basic question that I figured it would have no trouble
27:53
handling, you get bogus answers.
27:56
Is accuracy really a thing?
27:58
I mean, can you... Can we rely
28:01
on accuracy as a hallmark
28:03
of AI as it stands?
28:05
So actually I think for
28:08
ordering in drive-throughs, AI is kind of
28:10
the perfect use case. And the reason is that you have
28:12
a very small, very predictable
28:14
set of activities, right? If you ask
28:17
chat GPT the lyrics for
28:19
a song or something else, like you expect
28:21
chat GPT to know everything. You don't
28:23
necessarily go through the drive-through and say, hey, what's
28:26
the population of Nepal? They don't have to know that.
28:28
You don't expect that. And so because
28:30
the question space is much smaller,
28:32
the accuracy goes up dramatically.
28:35
You can still ask them what the population of Nepal is, but they can
28:37
just tell you to bugger
28:38
off because they're not expected to
28:40
know. And I think because of that, you
28:42
end up with much higher
28:44
accuracy because you know, like there's
28:47
probably maybe 300 interactions
28:50
in total that you might possibly have with
28:53
an ordering chat bot.
28:55
Now the accuracy thing is true, right? The
28:58
accuracy for chat GPT 4.0, which
29:00
is like the newest one, isn't
29:02
super great. So I think the thing
29:04
about chat GPT 4 that's interesting is that
29:07
we're currently arguing about
29:10
a malfunctioning toddler. Toddlers
29:14
aren't meant to be fully functioning human beings yet,
29:16
right? Chat GPT 4 has been out
29:18
for less than a year and it's learning
29:20
at an incredible rate. Like a year ago, it basically
29:23
didn't exist. But I think there
29:25
is something there. The accuracy will go up and
29:27
there are like different degrees of accuracy.
29:30
What's important if they said, Hey, chat GPT
29:32
is going to be running 911. I'd be like, Hey, that's a really
29:34
bad idea. But if the worst case scenario
29:36
is that they give me a strawberry milkshake instead
29:38
of a chocolate milkshake, I'm like, okay, fine.
29:40
That's a like the error rate
29:42
there and the impact of an error I'm willing to live
29:45
with. Or maybe I'm not, and I'm going to shout as manager,
29:47
but either way, kind of the problem space
29:49
is limited enough that it seems safe enough
29:51
to try.
29:53
Yeah. Kenny, you, you raised some really
29:55
serious concerns here. I'm curious though, have you
29:57
actually tried chat GPT or any of these?
30:00
AI services yet?
30:03
I Mess with it a very
30:05
little bit, but I'm just like there's
30:07
just part of me That's like I don't want to participate
30:09
in this at all I'm just kind of
30:11
waiting and seeing I did mess with it a little bit.
30:14
It is fun I have a
30:16
couple friends, you know that are writers and
30:19
they they had they made some interesting points They were like, yeah,
30:21
Chet Che GPT is not gonna be able to you know
30:23
bang out a pilot for you or whatever But
30:25
it's actually can be a good writing partner and
30:27
a friend of mine was telling me this he was working on a short film
30:30
And he'll feed jet chat GPT like okay
30:33
two people want to date one's a germaphobe
30:36
and one is really messy Give
30:38
me six examples of something
30:40
that could happen at the dinner and like for
30:42
those kinds of prompts It seems
30:44
to be really helpful It's kind of like a writing
30:47
partner that you can bounce ideas off of so,
30:49
you know I think where I'm trying to land with
30:51
this stuff is
30:53
It's here
30:55
Learn to be responsible with it. Don't be Don't
30:58
just shut it out and pretend it's not there because it's not going
31:00
anywhere so I think
31:02
at some point I will sort of test the waters
31:04
a little bit more and Open
31:06
myself up to it, you know, but it
31:09
is hard because I think about it, you know with the drive-through
31:11
thing It's like a cost-benefit analysis, right?
31:14
So the benefit is what more accurate fast
31:16
food orders and the cost is
31:18
Tons of jobs and even more
31:20
economic inequality
31:22
Like I don't know man. I don't really care if
31:24
a human gives me the wrong milkshake it's fine at least
31:26
that person other job and So
31:30
to me it's like the cop but like again with
31:32
the wildfires That's a great application for
31:34
it that clearly benefits society I
31:36
just don't think more accurate Wendy's orders
31:38
is I don't think the juice is worth the squeeze When
31:41
it when it when it comes to that, you know, yeah
31:44
Comedian Kenny DeForest is talking
31:46
with us as well as journalist hiya camps
31:49
of tech crunch You're listening the best of business
31:51
wars daily and coming up We'll be talking
31:53
more about how AI is disrupting industries
31:55
from firefighting to advertise
31:58
stay with it
32:03
Everyone knew Mike
32:05
and Denise Williams were meant
32:07
for each other. It was the perfect match.
32:09
High school sweethearts
32:10
married after college and
32:13
they just started a family. It
32:15
was a picture perfect life until
32:17
one cold December day when
32:19
Mike Williams disappeared.
32:21
Mike never returned
32:23
from a hunting trip yesterday. Everyone
32:26
had their own theory of what happened. Alligator
32:29
got him and took him somewhere. He did not drown
32:31
in that lake. How could somebody just
32:33
poof off the surface of the earth? But
32:36
there were some close to Mike who were
32:38
hiding a secret. From Wondery
32:40
comes a story about the extremes
32:43
of love. How it can consume us.
32:46
How it can make us turn a blind eye.
32:48
And how it can ruin us. Enjoy
32:50
Over My Dead Body Gone Hunting on the Wondery
32:53
app or wherever you get your podcasts. You
32:55
can listen to Over My Dead Body early
32:57
and add free on Wondery+. Get
32:59
started with your free trial at wondery.com.
33:10
Hey, welcome back. We're talking with journalist Haya
33:12
Camps and comedian Kenny DeForest
33:14
about AI. Just
33:17
do it. Because you're worth it. Diamonds
33:19
are forever. Where's the beef? I
33:21
bet you can name the companies that go along with those
33:24
iconic slogans, right? They were the
33:26
product of lots of late nights and work shopping.
33:28
In other words, human creativity. But
33:31
that may not be the case much longer. Artificial
33:34
intelligence is already changing advertising.
33:36
And, well, rather than trying to fight it, ad
33:38
agencies are embracing AI. Let's
33:41
hear a bit more about that with a clip from Business Wars
33:43
Daily. Last
33:46
week, WPP, the world's largest advertising
33:48
agency, announced it was going, quote,
33:51
all in on AI, CNN
33:53
reports. The agency is teaming up
33:55
with NVIDIA, which is one of the leading suppliers
33:58
of AI hardware and software. The
34:00
two companies are creating a platform that
34:02
allows clients to create ads and other content
34:05
simply by typing in a description
34:07
of what they want. Oh
34:09
boy, watch out Draper. This
34:11
new platform will allow brands to create
34:14
two-dimensional images and videos and
34:16
three-dimensional simulations for advertising
34:18
purposes, Ad Age reports. It'll
34:20
connect generative AI tools from places
34:23
like Adobe and Getty Images, giving
34:25
clients access to sweeping libraries
34:27
of images and graphics. WPP's
34:30
project will launch soon, although the
34:32
company has not yet said when.
34:35
WPP isn't the only agency giant
34:37
embracing generative AI. Ogilvy,
34:40
which reps clients like BMW and
34:42
Burger King, has launched its AI
34:44
lab in Paris. The goal,
34:47
the company said in a statement, is to harness
34:49
and build on its expertise in using AI
34:51
and give brands access to AI art
34:53
generators so clients can use
34:55
the platform to create their own ads
34:58
and videos for various purposes. All
35:00
of this access to creative materials
35:03
on demand raises questions about the future
35:05
of the agency model. After
35:07
all, when an AI can show a luxury
35:09
car or a designer ensemble on
35:11
a London street or any desert in the world
35:13
by typing a few words, the
35:16
need for pricey shoots in exotic locations
35:19
plummets. A representative from WPP
35:22
told CNN that the new platform outperforms
35:25
having people quote manually create
35:27
hundreds of thousands of pieces of content using
35:29
disparate data coming from disconnected tools
35:32
and systems. That sounds
35:34
like smaller creative teams will be needed
35:36
to service client accounts. Jenny,
35:40
I want to play a little game with you here because I know this
35:42
sounds far fetched, but let's imagine you're
35:45
a comedian with a comedy special that's
35:47
being released and you need to get some eyeballs
35:50
on that. Turn to
35:52
AI for PR material, perhaps? And
35:54
say you.
35:56
See, this is why I'm talking. I mean, yeah,
35:58
obviously that'd be amazing.
35:59
You know, I can streamline it, it's gonna
36:02
be faster, I don't have to pay anybody to do
36:04
it. But here's the thing, my album that just came
36:06
out, the album cover. One
36:09
of my close friends and high school teammates
36:11
is an incredible artist, and we paid
36:13
him to make the cover
36:16
of my album. And now I'm selling stickers
36:18
and I'm selling t-shirts and him and I are splitting the profits
36:21
of that. And that feels so much more gratifying
36:23
than if I had just used AI. And here's the thing,
36:25
I know AI would have done a good job, I would have made something
36:28
cool. I don't pay anybody, I'd make more money,
36:30
but
36:31
I don't know, human collaboration, especially in
36:33
art, that's what art is. I mean, it's
36:38
always in some respects
36:40
collaborative. I mean, obviously there might be a visual artist
36:42
that works by themselves, whatever. But
36:44
even so, the human element in art
36:47
is so important, I think. I
36:49
hope, I want to believe that. Maybe
36:51
I'll be wrong, you know, maybe in a few
36:53
years, there'll
36:54
be a 100% AI movie that blows
36:57
my mind, I don't know. You
36:59
know what this reminds me, at this moment reminds
37:01
me a lot of what was happening with
37:05
recorded music, you know, at the turn
37:07
of the millennium. I mean, where
37:09
we have a lot of people involved in
37:12
the, and there's a lot of ways too, in the
37:15
making, not just the creative
37:17
content, but the marketing
37:20
of that. It disappears and
37:22
what you are left with is, yes,
37:24
the end of the gatekeepers, but also a kind
37:26
of grand leveling in that you have
37:29
more creators who are now able
37:31
to access platforms
37:35
and ears that they might not have been
37:38
able to reach were
37:40
it not for the proliferation
37:42
of technological tools, like, you
37:44
know, they could pro tools and that sort of thing, right?
37:48
What are you seeing in your reporting? I mean,
37:50
is there a kind of parallel here, parallel
37:53
experience that AI represents?
38:00
play with AI a lot and I have
38:02
tried, I've really tried to make it write
38:04
my newsletters for me. And in theory,
38:07
I feel like it should be able to, right? Feeded all the
38:09
articles, like turn it into a newsletter. And
38:11
it just falls flat. It
38:13
has no personality, it has no soul. And
38:16
I think the stuff Kenny touched
38:18
on is like, I think for art
38:20
to make sense to humans, it needs to have
38:23
a message. It needs to have something that goes underneath,
38:26
that understands humanity to
38:28
be able to connect with it. Like I am not,
38:30
I'm not Hemingway, right? But I would like to think
38:33
that my writing is funny, engaging, ads
38:36
analysis, ads context, and
38:38
people read it for those reasons. Like
38:41
a lot of the stuff that a lot of journalists cover,
38:43
you get a press release, you mush it
38:45
around a little bit, you add a picture, you hit publish,
38:48
please AI, take that away from us. Humans
38:51
don't need to be involved with that. But for
38:53
real like fun reporting, talking
38:55
to people, engaging, like
38:58
it's about the core of what makes us human.
39:00
I think that talks about art. I think that talks about anything
39:03
we're like, I'm sure you've had this, right?
39:05
You sit down with a magazine and you read an article, it's like, whatever,
39:07
I've got some new facts. Or you have one where
39:09
you're like, oh, that was engaging. I feel like
39:11
I knew the story. I understood
39:14
the
39:15
characters. I understand why this impacts
39:17
the emotional element there. Yeah. Right.
39:20
Right. And so far, AI
39:22
hasn't been able to do that. And I think like
39:25
there are some advertising type forms
39:27
where AI can probably do well. Like it can
39:29
crank out 6000 different Facebook
39:32
ads that are super targeted to Kenny's
39:34
interest, to my interest, your interest. Right.
39:37
And so I can see that if it's just
39:39
about crunching numbers, like
39:41
an AI can
39:43
crank out 10,000 pictures and 10,000
39:45
ads and see what I look at for
39:47
a millisecond longer than any other ad and then
39:49
reformulate that ad. Great. But humans
39:51
should never be involved with that anyway.
39:53
Hiya, you mentioned something that
39:56
made me think of a kind
39:59
of dilemma that I think the
40:01
rise of AI hints at.
40:04
And it's not just the, there's
40:06
been so much written and talked about plagiarism,
40:08
for example, getting AI to do your homework and
40:10
all that. But if you take it to
40:13
that next level where you have professionals
40:15
say in the creative business,
40:17
and for you as a journalist, of course, it's
40:19
writing that story and it's coming up
40:21
with that narrative. On the one hand,
40:24
I can see how being inspired
40:26
by something that chat GPT or
40:28
something might spit back out at you would
40:30
be a benefit. Do
40:33
you credit that at all?
40:35
publicly? Do you do you? Do you
40:38
concede to your readers
40:41
that you had assistance from AI
40:44
or not? How do you how
40:46
do you try to, you know, square that
40:48
because these are some real issues facing a lot
40:50
of folks in the creative arts right now.
40:53
Well, I think there's a couple of truths in
40:55
there. And I'm like, one of them is that I don't think
40:57
original art exists. And
41:00
what I mean by that is that every everybody
41:02
like everybody is inspired by everything.
41:05
The idea of simultaneous invention, like
41:08
the crossbow was invented in six different
41:10
places all around the world at roughly the same time. That
41:12
wasn't because everybody had this lightning
41:15
of genius idea was just like it was the next
41:17
logical step. And I think a lot of that happens for
41:20
creativity too. Like I
41:22
at some point wrote an article which was like, Hey, am I
41:24
an AI? And what I meant was
41:26
that I'd written books that were that
41:28
were based on ideas
41:31
that were brought by other writers, other
41:33
people. So I was adding something,
41:35
but I was also collating a lot of information.
41:38
And that's pretty wild.
41:40
I'm hearing something very comfortably analog
41:42
in the background that train going
41:44
past your window. Right?
41:47
Yeah, yeah. There's something soothing
41:50
about the old school. And I
41:52
think that that's a big part
41:54
of the way that the lens through which we're
41:56
understanding AI
41:58
that it is something for which We don't really have
42:01
a perfect parallel. I think everyone
42:04
is concerned about what this means writ
42:07
large. But sticking with what
42:09
we were talking about there, I know recently Ogilvy,
42:12
the PR company, launched this thing called the AI
42:14
Accountability Act calling for
42:17
influencers to disclose their collaborations
42:20
with AI-generated content. They
42:22
disclosed that with a hashtag like, hashtag
42:25
powered by AI or some kind of watermark.
42:28
One of the leaders of the Ogilvy campaign
42:30
said, I want to quote here, if we aren't
42:32
honest with our end consumers, then
42:34
we lose the effectiveness of influencer marketing
42:37
as a channel. And I think in a way this gets to
42:39
this human aspect of
42:41
it. And what, if anything,
42:43
it reveals when we're
42:46
trying to be honest with those
42:48
that we're communicating with. Kenny,
42:51
do you trust influencers in the first place? Or are you
42:53
on social media much? Or would you appreciate something
42:55
like an accountability act like that?
42:58
Yeah. To answer your first question,
43:01
absolutely not. I
43:04
don't know if you've met an influencer, but they
43:06
basically are already robots. So
43:09
I assume. But yeah,
43:12
I think people should absolutely have to disclose
43:14
that. I think just especially as,
43:16
again, it's like you said, it's brand new. We don't know what this
43:18
is. There's a lot of optimism. There's a lot
43:20
of concerns. And I think one
43:23
way to help with that is to disclose
43:25
so that the public knows. I think people should absolutely
43:28
have to disclose that. I think just
43:31
especially as, again, it's like you said, it's brand new.
43:33
We don't know what this is. There's a lot of optimism.
43:35
There's a lot of concerns. And I think one
43:38
way to help with that is to
43:40
disclose so that the public knows, okay,
43:42
this is a situation where AI is being used. Now we
43:44
can decide, does this seem like a good
43:46
thing in this particular arena? Does it not? I
43:49
think the public being informed about
43:52
the ways in which it's being used is going to be paramount
43:54
to sort of figuring this thing out in a way that could
43:56
be beneficial and not as harmful.
44:00
think like in my world, you know, right now, a lot
44:02
of my, I myself am not a Guild
44:04
member, but we have a lot of my close friends are on
44:06
strike. And, you know, we're looking
44:08
at the effect on, you
44:11
know, television and movies and
44:13
things of that nature. And, you know, and
44:15
in the future, like, listen, there's people that have come
44:17
up through stand up, I've been doing this a long time. And
44:19
there's just certain people that have a charisma
44:22
and a look, and they're maybe not
44:24
great joke writers, but people are like, no,
44:26
that person has something. And they'll
44:28
try to hire a different comedian to write jokes
44:30
for that person. Well, what if that person could just use
44:33
an AI joke generator and be like,
44:35
all right, I grew up in this place
44:38
in a trailer park, write
44:41
a joke about my childhood, and then
44:43
like, you know, they might be able to go
44:45
on tour
44:45
with that. And I personally hate that and think that
44:48
stinks. And I would love it if they would have to tell people
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More