Latest / Leaders in Lending / Cognitive AI for Transformative Customer Service

Transcript

Cognitive AI for Transformative Customer Service · Leaders in Lending

0:02
You are listening to Leaders in Lending from Upstart, a podcast dedicated to helping
0:07
consumer lenders grow their programs and improve their product offerings. Each week, here,
0:13
decision makers in the finance industry offer insights into the future of the lending
0:17
industry, best practices around digital transformation, and more. Let's get into the
0:22
show. Welcome to Leaders and Lending. I'm your host, Jeff Keltner.
0:27
This week's episode features my conversation with Peter Voss from AIGO, and Peter's actually
0:32
the guy who back in the day in the early thousands, point the term
0:35
artificial general intelligence. So to be a little outside of our normal thing,
0:38
we're going to go deep on AI in this conversation, and it was really
0:42
fascinating for me because we talked about the recent explosion of interest in generative AI,
0:48
what it does well, what it gets wrong, like hallucinations, and
0:51
why Peter is both optimistic about our ability to create artificial general intelligence, not
0:57
just specific intelligence, but a very generalizable, flexible intelligence, but also that
1:02
the general technologies around statistical AI that underpin most current approaches are really not the
1:07
right way. And so he talks about his approach to what he calls cognitive
1:11
AI, what they're building on that front, what he calls his chat pup
1:14
with the brain, and what the future looks like for generalized intelligence generalized artificial
1:19
intelligence. So I think it's a really fascinating conversation for anyone who's curious about
1:23
what's going on in THEI, what's next, and maybe talk about the edges
1:26
of where some of these things are not working as well and why. So
1:29
please enjoy this conversation, Peter Boss Peter, welcome to the podcast, and
1:34
thanks for making the time to join us. Yes, well, thank you.
1:37
I'm looking forward to this conversation. It's been a little off the beaten
1:41
track compared to many of my conversations with bankers. I don't think we'll go
1:44
down any deep banking rabbit holes. But you've got a little more of a
1:48
background and AI, so give us a little bit of your history to so
1:49
people can have a little context of where you're coming from and what you've been
1:53
doing for the last number of years. Yeah, certainly so. I actually
1:57
started up as an electronics engineer, intarted my own company, and then I
2:02
fell in love with software and my company turned into a software company. I
2:07
developed an ERP comprehensive ERP system for multi medium sized businesses, and that company
2:13
quite well. We went from the garage to four hundred people and did an
2:16
IPO so that that was very exciting. I'd love to do that again.
2:22
And it's when I exited that that company, I you know, I said,
2:25
well, why do I want to tackle next? And what occurred to
2:30
me is that software really isn't inherently very smart. You know, if the
2:35
if the programer didn't think of something, then it'll just crash or give you
2:38
an error message, or you know, do something stupid. So I really
2:44
want to figure out how can we build software that actually has common sense,
2:49
that can learn, that can reason. So that's a journey I've been on
2:53
for the last you know, twenty plus years. And I took off five
2:55
years to study intelligence, all different aspects of intelligence to be able to build
3:01
intelligent system. But I thought I really needed to understand it, and so
3:07
I studied how children learn, how our intelligence differs from animals, what IQ
3:13
tests measure, you know, all the work that had been done in the
3:15
field of AI, and that study culminated and me coming up with a theory
3:20
of how to actually build a thinking machine and then two thousand and one I
3:24
started in my first AI company, and for several years we were just in
3:29
R and D mode, sort of turning my ideas into actual code prototypes,
3:34
and eventually a platform that we then commercialized and automated calls and core center automation,
3:40
and doing that a lot more intelligently than what had been done before,
3:45
because we have a cognitive engine that actually deeply understands what you're saying. Users
3:52
context uses history of what you said earlier in the conversation, so it has
3:55
short term emy long term memory. And that's really the journey I've been on.
4:00
So my current company also commercializes conversational AI, but we also are continuing
4:11
to increase the IQ of a system. You know, we really want to
4:14
get to human full human level intelligence. It's fascinating to me the discussion about
4:20
human intelligence and how close or not, I mean, there seems to be
4:26
this very wide range of opinions and how close some of the you know,
4:29
at least the consumer facing cutting edge technology things like cloud or bard or chat.
4:33
GPT is probably the most popular, and that transformer architecture that seems to
4:38
kind of revolutionize the quality of conversation we're having and I hear everything. I
4:43
think we're like one turn of the crank on that architecture from having like a
4:46
superhuman brain, and people who go, hey, we're just not even on
4:48
the right path. You're kind of doing this thing, and I'm kind of
4:51
curious where you stand on that, and what do you think true human like
4:56
intelligence requires. And it's not just like being able to hold a conversation at
5:00
the touring test is to be out of date in terms of how we think
5:01
about what a truly intelligent machine needs to look like. So what's your perspective
5:05
on how far or close are those kinds of technologies and what do you think,
5:10
if anything, they're missing in terms of how their approach generates or christ
5:14
to mimic intelligence and humans. Yes, it's a very interesting and very important
5:18
question at this point in time, because, you know, check GPT became
5:24
available. I think it really was a major shock to many people and including
5:30
pretty much all AI researchers. Virtually nobody expected the system to be as good
5:36
as it, you know, to be able to do the hold the kind
5:40
of conversation that it can and the amount of knowledge that it has the you
5:44
know, that's one of those fascinating things to me is that people who built
5:47
it really like it's kind I kind of describe, feels like a really fancy
5:50
autocomplete, and then how it got to things like reasoning and understand like,
5:56
they really don't understand how that emmerged to really build a reasoning system. They
6:00
built like an autocomplete, and somehow it has some basic deductive you know,
6:03
it's like they're really surprised by what I can do. Yes, yes,
6:06
absolutely, And you know, I think I think there are ways, certainly
6:11
ways of understanding it, but there are as impressive as it is and as
6:16
useful as it is, there are inherent fundamental limitations to that. And it's
6:23
difficult to look past the sort of impressiveness of it. You know, it
6:28
just seems so intelligent. It's really good at faking intelligence, and of course
6:33
it you know it, it can do really really amazing things. But the
6:41
problem is it's really not reliable. It's not aware of what it's saying at
6:46
all. So there is no metacognition, and so there are a couple of
6:49
components that are really missing to truly make it human like in human level intelligence.
6:58
So and in fact, the name GPT gives you a clue already.
7:01
G stands for generative and which you know another way of saying that it makes
7:06
up stuff and it literally does. I mean you can prompt it and and
7:12
you know ask you know, I can ask it, tell me about the
7:16
white paper I posted on kidney revival or something you know, and you know
7:24
it'll it'll come up with something, you know, complete with One of those
7:27
lawyers recently who asked a question came up with like three precedents and they submitted
7:31
them in a brief and then judge those are those are real cases. You
7:34
can't just so that was the fine I think a couple of thousand dollars for
7:39
profression on this conduct. It's like relied on it, right, that's exactly
7:44
right. So yeah, I mean that's the generative part. And that's inherent
7:46
and generative. It's trained on trillions of pieces of information, you know,
7:50
good, bad, and ugly, and you know what it comes up was
7:55
is really just that it it's sort of coherent in its statistical model, but
8:00
it doesn't it doesn't need to bear any kind of relationship to reality and so
8:05
so, and that's inherent. It's just inherent an inherent limitation of the system
8:11
the way it's built. Now. The second part is GP is a pre
8:15
trained and it's really a read only model. They build a model, it
8:18
costs you know, one hundred million dollars or whatever, you know, some
8:22
enormous amount. Yeah, and it takes you know, days or weeks to
8:26
train the model. And but then it's essentially after that it's largely read only.
8:31
Now it cannot learn interactively. Now, there are tweaks that people you
8:37
know, that that the system allows, that it has some kind of a
8:41
short term buffer that it takes into account. That's basically the PROMPTU you feed
8:45
in. But that information isn't doesn't update the model. So whatever you teach
8:50
it, there is, you know, does not change the model, which
8:54
is very different from the way humans operate, you know. I mean we
8:58
can hear, we can hear one word or two words, and it can
9:01
completely change the way we perceive things. And you know, like Trump one,
9:07
you know, you know, for better or for worse. I mean,
9:09
it changes everything, you know, if you didn't expect it. So
9:13
it updates our model. Basically, you know, certain things just one or
9:18
two words can really update our model significantly. And I think that's what intelligence
9:22
requires. And then what I already mentioned is a lack of metacognition. It
9:28
doesn't know what it's what it's saying, It can't think about its own thinking.
9:33
So these are requirements I think of real intelligence. I mean, imagine
9:37
not hiring an assistant, you know, human assistant for you and that come
9:41
and they oh, yeah, I can do Excel, and I can you
9:46
know, do PowerPoint, and I can you know, do whatever, you
9:50
know, look up things on the internet for you. And then you tell
9:52
them, well, you know, we introduce this new product, we shut
9:56
down one of our branches, we have this new relationship, and so on.
10:00
And next day your assistant comes in and they don't know any they don't
10:03
remember any of that. You know. Yeah, my son had an article
10:07
to respond to today, the middle six, sixth grader, and he had
10:11
to write a summary. And you know, we were talking about this exact
10:13
point because I said, imagine what your history teacher would do if she gave
10:16
you a lesson. And then you said, hold on, give me a
10:20
week and one hundred million dollars and I'll come back and be able to pass
10:22
the TESK. But between now and then, I won't have any memory of
10:24
what she said. He wouldn't past sixth grade very well, let alone a
10:28
job interview or do well in the workforce. So it's it's kind of unbelievable
10:31
when you think about how young a person would have a better level of intelligence
10:35
from that real time learning perspective, it's really quite remarkable, right, And
10:41
it's absolutely crucial for you know, intelligence in the real world. You need
10:46
to be able to learn things on the fly and think through what the implications
10:50
are of the new knowledge that you have. And this technology you simply cannot
10:56
do it. Transform a technology cannot do it. So we need we need
10:58
to different approach, and the approach that I believe we need is something called
11:05
cognitive AI as opposed to statistical or generative and cognitive AI really starts with the
11:11
question of what does intelligence require and to build a system around that, where
11:16
statistical AI started with, Hey, we've got a lot of data. We've
11:20
got a lot of computing power, you know, the big companies like Google
11:24
and Facebook. So what can we do with a lot of data and a
11:28
lot computing power? You know, that's the hammer we've got. So what
11:31
nails can we find? It's not starting with what does intelligence require? And
11:37
that's really the I think the approach that's needed, and you know, that's
11:39
the approach that we've taken. Do you see any kind of recognition? I
11:45
mean, it feels like there's so much money in investment pouring into the generative
11:50
AI space. Now, whether it's you know, everybody trying to catch up
11:52
with chat GPT in terms of anthropic or Google investing in Claude and Bard,
11:58
or whether it's vcs investing in startups, they're all using the same kind of
12:03
statistical AI approach, right They're iterating on the transformer or the diffusion models in
12:07
the space of image image generation. Do you see enough investment going into this
12:13
kind of like new approaches because it feels like we're all iterating on the same
12:16
thing versus trying to build the new thing. There's so much resource going there.
12:20
I wonder what your thoughts are on that versus seeing more investment into different
12:24
approaches that more closely mirror what we think is going on in human intelligence.
12:28
It might be required for a more generalized intelligence capability, right, Well,
12:33
you know, obviously it has a lot of momentum and really for I think
12:37
since really the dot com boom. People really only think in terms of momentum
12:43
investing. You know, it's basically, we are things happening. Hey,
12:46
let me pile in, let me put more. And you know that that
12:50
has this flywheel flywheel kind of effect that the more people jump on, the
12:54
more money comes in. You know. And you know, you can't even
12:58
get a do a PhD computer science if it doesn't involve machine learning, deep
13:03
learning, statistical systems. These days, you know, you're probably not going
13:07
to get funded on a VC if you're not using that, so it you
13:11
know, it hasn't effect. And I think what's happening now is there is
13:16
a realization of hang on, these are real limitations that we have. We
13:20
cannot trust these systems. They cannot learn, they don't have metal cognition,
13:24
you know, and and they're incredibly expensive to run and to train. So
13:31
people are starting to realize it. But you know, the momentum has to
13:35
kind of work its way out, and I think a lot of vcs have
13:39
to lose a lot of money, which I think is going to happen soon.
13:43
You know, as these these startups that try to capitalize on it just
13:48
find well, you know, what are actually applications that are creating value.
13:54
Now, there are a lot of applications that create value, you know,
14:00
such such as as a programming aid and to help you brainstorm staff, or
14:05
to generate images or logos or you know, things where there's a human and
14:09
loop using it as a tool. But there's you know, there's a limited
14:13
amount of value that creates. It doesn't really replace human labor on a large
14:20
scale, and that's really where the real value would be created if it could
14:26
do that. So I think we'll see the tide turning. The problem is
14:31
all of the experts, almost all of the experts in the field, the
14:35
only thing they know is statistical approaches. So you have a whole ecosystem where
14:41
the AI experts in fact, you know, if you talk to somebody now
14:45
who graduated in the last ten years, they actually think that AI is statistical.
14:52
That's it. You know, there's no there's no other approach to AI.
14:56
So you basically have all of all of the experts in the field that
15:01
their background is mathematics, statistics, and logic, and those are kind of
15:07
the only way they can think about the problem. So it's going to be
15:11
pretty difficult, you know, to to have that whole ecosystem really change to
15:16
say no, these approaches are not going to get us to true human level
15:22
AI. So I don't know how long that will take, but certainly seeing
15:28
an you know, an increasing realization of what the inherent limitations are and you
15:33
can't just brute force your way to to human level intelligence. Yeah, it's
15:39
interesting you talk about the evolution of companies and how they invest in this or
15:43
not. And it reminds me a little bit of like the moment when Google
15:46
was founded, because it feels so inevitable looking back what Google became and what
15:50
it is. But like search was considered by many a solved problem and they're
15:54
like, there's not room for other search engine, there's licos, there's alt
15:58
that this like because a starting a search engine right now, it doesn't make
16:00
any sense. Uh, And they were clearly off the mark. And I
16:04
think that feels like the same inevitability that Google looks like in retrospect behind the
16:10
guys you know open AI and chat GPT, for instance. But it is
16:14
often true that that's the moment when the true innovation comes out and new things
16:18
are founded that end up looking inevitable in retrospects. But maybe foolhardy at the
16:22
time, and maybe it's one of those moments for a totally different approach despite
16:26
the large, you know, influence that some of these companies have right now.
16:30
Yeah, now you make an excellent point there, and I mean,
16:33
yes, how could a little startup Google possibly compete against the existing, you
16:40
know, established search engines? You take the same with Amazon. How could
16:44
some you know, little mickey Mouse online books books possibly take on Walmart?
16:49
You know? And we see that again and again. It's, you know,
16:55
at a certain point, the establishment is, you know, it is
17:00
fixed on a particular approach and they really can't break out of it. So
17:04
I want to ask you a little bit about what you guys do You describe
17:07
it to me once at your company as a chat bot with the brain one
17:11
of the products you have, So talk to me about how that plays out
17:15
in a company and what you mean when you say with a brain, Like
17:18
that's an interesting phrase. I think I'd love to understand where you're kind of
17:22
coming at from with that. Right Well, all the other chatbots out there,
17:26
as far as I know, do not have a brain. And what
17:30
we mean by brain is basically a cognitive engine that has deep understanding users context.
17:37
Has short term memory, it remembers what were said earlier in the conversation,
17:41
has long term memory, it remembers previous conversations with each individual. So
17:48
it really has a comprehensive cognitive engine or brain that manages the whole conversation,
17:56
and other chatbots really use in a approach that's been around for thirty years or
18:02
so. So on the one hand, you have an intent classifier where you
18:07
know, whatever the utterance is blah blah blah, weather okay, it triggers
18:11
the weather app. Or you say I hate Uber, don't ever give me
18:18
Uber again, Well, I'll probably still trigger the Uber app because it latches
18:22
onto that. So it's a fixed a fixed list of options that can get
18:26
triggered. And once they get triggered, there's basically little float charity type program
18:30
that somebody writes, you know, where do you want to go? How
18:34
many people are going? And do you want Uber X? So there you
18:38
know. And then obviously people have different levels of sophistication where they may hard
18:45
code certain terms to be remembered, so there may be some very limited specific
18:51
short term memory or long term memory. They look up something in the database,
18:55
but it's not an inherent part of the architecture that there is deep understanding.
19:00
Context is being used. The system reasons about what it hears, and
19:04
you know, and and so with a brain you can have a much much
19:10
more sophisticated conversation. So as an example, one of one of our customers
19:15
is one eight hundred Flowers, a group of companies which includes Harry and David,
19:22
popcorn factory and so on. So they use our system very very successfully
19:26
to provide a concierge type service to their twenty million customers. And so our
19:33
system remembers who you buy gifts for, for what occasion, what kind of
19:37
gifts they're like, it's able to do that because of the deep understanding and
19:41
the memory. So you know, this is where the brain really allows us
19:48
to handle very complex conversations. And the same approach you know, can be
19:55
applied to you know, banking or medical applicat as or whatever. I mean,
20:02
we've all experienced how limited IVRs, you know, voice interactions and chatbots
20:08
are. I mean Bank of America Erica, I think they have like what
20:14
three percent repeat usage or something, you know, very very low number in
20:19
spite of hundreds of millions of dollars being spent on that they don't have a
20:25
brain. That's fascinating. I'm curious. How where do you think you guys
20:29
are at in the taking the architecture from what it's being used for today to
20:33
more generalized tasks. You talk about replacing human labor on a large scale,
20:37
Where are your systems at and the ability to move from a call center,
20:41
a chap type interaction to you know, more generalized cognitive capabilities and types of
20:48
labor. Yeah, So over the last few years, our concentration has has
20:52
been to really build out the infrastructure, the robustness, you know, and
20:56
scalability and easily implementation and all of that. So our focus has been very
21:03
much on commercial aspect of our technology. But you know, we now have
21:08
a very solid system and our focus is actually shifting back again to cranking up
21:15
the IQ of our system. You know, there's still quite a ways to
21:18
go to scale up the amount of knowledge that the system has, the degree
21:25
of complexity of what it can handle, how it can handle, For example,
21:30
theory of mind. You know, what is the other person thinking?
21:33
What does the other person know? So we need to scale up our brain
21:38
in various areas to you know, handle these more complex situations. Also,
21:45
at the moment, there's still a fair amount of work involved in actually teaching
21:49
the system the specifics of each company that we that use our system. And
21:56
as our system becomes smarter, it will be able to self configure to a
22:02
much larger degree where you know, it will be able to ask the question,
22:06
you know, do you want me to handle it this way or that
22:08
way, or let let me read let me read your training manuals. You
22:14
know, at the moment, there's still a human in the loop in many
22:17
of us for many of these training exercises. So as we crank up the
22:22
IQ, the implementation implementations will become much easier and it will be able to
22:30
handle, you know, even more complex situations and conversations. So we we
22:37
just recently sort of rebooted our development focus. We currently have about ten people
22:47
working on that full time, but we are actually in the process to raise
22:52
money to increase it to one hundred. We want a hundred people on that
22:56
project. We know what we need to do to crank up IQ to get
23:00
to human level intelligence, but were currently looking for funding and so we're looking
23:06
for the kind of visionary investors that can look past the what if you know,
23:12
the bandwagon that everybody is jumping jumping on and say, well, what
23:18
does intelligence really require and who has who has potentially a solution for that.
23:22
I'm fascinated by this idea that the smarter version of your iteration of your program
23:29
will start to kind of be able to train itself. I mean, what
23:32
are the things you think? What are the real markers of as the system
23:36
gets more and more intelligent, Like the self configuration asking for the data it
23:40
might need to be able to respond more accurately is a really interesting one,
23:44
kind of like self directed learning, which is very different I think than in
23:48
the way we might use those terms today. But what are the markers you
23:51
see of like how this evolves and how you can tell that it's getting smarter
23:55
and smarter? Like, what are the you know, there's not like a
23:56
ninth grade test to a high school test you want to pass the way GPT
24:00
I think passes the you know, the bar exam or whatever. What are
24:03
the markers for you as you get farther down this line of real real cognition
24:08
and you're ramping up the IQ of the system, what does that look like?
24:11
Yeah, so it's very different for us. It's to us, it's
24:15
this is not a black box. So it's not that we kind of keep
24:19
throwing more data and computing power at it and then wow, you know what
24:25
can it do? Now? Wow that's cool? You know, Okay,
24:27
can we tweak it a bit here and there to to make sure it doesn't
24:32
swear at us or say naughty things, you know, or whatever. That's
24:34
not at all the approach I mean, uh, the with with the cognitive
24:38
architecture that we have, it's it's completely scrutable, so we can see what
24:42
is doing and why it's doing what is doing, and so we have a
24:47
curriculum that we're going through to enable the system to handle more and more complex
24:52
tasks, to handle more and more complex reasoning. So you know, I
24:56
mean, they're little surprises along the way we say, wow, this wor
25:00
a lot better than we thought it might, you know, or this is
25:03
a lot harder than we thought, and you know, it takes more time.
25:07
But we really have a particular development paths that we've mapped out to get
25:12
from where we are to get to full human level intelligence. And it's just
25:17
an incremental incremental paths, and you know, it requires a fair amount of
25:22
effort, not billions or trillions of dollars, but you know, certainly tens
25:27
of millions, you know, I mean for us to you know, we
25:32
need you know, one hundred two hundred people working on this to to really
25:37
get to human level commercial product. But we have a very specific road map
25:45
to achieve that. What kinds of things are on your curriculum? I'm curious.
25:48
I always ask my kids teachers for their you know, curriculum for the
25:52
year. What are the kinds of things just conceptually that are on the curriculum
25:56
for your system? Yeah, So with a new project that that we have
26:00
now is one of the big changes is our commercial product. The language ability
26:07
was really put in by humans. You know, we've fed it, we've
26:11
feded the different language rules and so on. So there was a lot of
26:15
curation being done. And we've always known, and we've known for you know,
26:21
for the last twenty years that we've been developing the technology and commercializing it.
26:26
We've always known that ultimately the system has to be able to learn language
26:30
by itself, essentially with a teacher. So our current focus is on you
26:37
know, finally replacing these human coded rules with a knowledge that the system requires,
26:45
and so that's the curriculum we're right now doing, is to teach it
26:51
the basic language skills interactively. And you know that that curriculum, it has
27:00
certainly has some overlap to the way children learn. You know, there are
27:03
layers of what is the most fundamental knowledge that you need to have. You
27:08
know about things. There are objects, there are activities, there are attributes
27:14
that they are entities. You know, there are general categories and those things
27:19
have to be kind of very deeply embedded. So that that's our focus right
27:26
now. Uh, the curriculum now as we do that, we will we
27:30
will then go back to the kinds of things we've been doing. Is you
27:34
know, how how do you take an order or how do you interact with
27:40
APIs to achieve different to achieve different things, So that that will be the
27:45
curriculum that that we execute as we go along. That's a that's a fascinating
27:51
journey. I can't wait to see what comes out of it. It seems
27:55
like a very interesting and different approach and it'll be really interesting to see how
28:00
hopefully your system doesn't have the same kind of teenage angst and difficulty that humans
28:06
do. But the deuteractive learning a very interesting path. No, no,
28:12
no, it will not. We're not going to put injected with hormones or
28:17
you know, so it's it's not going to have to go through puberty or
28:21
any any of that. So that's thankfully not something we have to deal with
28:25
in our AI. But you know, I mean, I am so excited
28:30
about working on this project, and I have been for as long as I
28:33
have been working on it because of the tremendous potential that it has. And
28:37
you know, I mean I could name just like three categories. The first
28:42
one is imagine having a true PhD level researchers AI researchers. You know,
28:51
you get one that gets trained to be a cancer researcher, for example,
28:56
with all the background knowledge, then you can make a million copies of that.
29:00
You now have a million PhD level cancer researchers chipping away at the problem.
29:06
You know, I mean, we're going to make much much more progress.
29:08
You take that for other diseases. You take it, you know,
29:12
for energy for you know, battery efficiency, for pollution, you know,
29:18
climate change, all of you know, having these PhD level AIS chipping away
29:26
at these these many problems that are facing humanity. That's kind of the one
29:30
area. The second area is obviously dramatically reducing the cost of goods and services
29:38
by having a lot more things automated. I mean, human labor is the
29:44
bottleneck for affordability in in many things. And the third one that that is
29:49
actually, in my mind, even the most exciting one, is what we
29:55
call a personal personal assistant, and told really be called personal personal personal assistant
30:02
because of three different aspects of personal. So the first personal is that you
30:07
own it. It serves your agenda, not some mega corporation's agenda. The
30:12
second personal is that it's hyper personalized to you, so it learns your preferences,
30:18
your history. You know your goals and your likes, your dislikes and
30:22
so on. And the third personal isn't the one of privacy that you decide
30:27
what you want to share with whom. So you know, certain things you
30:30
share with your spouse, other things you share with your coworkers, and yet
30:34
other things you might share with Amazon. So you know, all of these
30:38
things are going to be possible with true human level AI, and I think
30:44
it will be tremendous to improve the quality of life, you know, and
30:49
basically help us flourish. And your second point is things often describe current generive
30:56
AI technologies to some people and other AI technologies is what would you do if
31:00
you had unlimited free interns, like not super talented people, but people who
31:03
could image, process, write basic copy, Like how would you leverage that
31:07
in your business? And you're talking about a world that you have maybe not
31:11
unlimited, maybe not free, but a plethora of highly trained experts in different
31:15
areas. It's a fascinating thing to think about what you could do with that
31:18
capability in the world. Really yeah, and the important thing is the ability
31:23
to learn in real time. You know that they can learn these tasks by
31:30
themselves the way human can. It's not then having to go through a training
31:34
process and whatever you train is that's limited to what it can do what the
31:38
training is. Now, this is truly dynamic, ongoing learning, reasoning about
31:45
it at a high level, and you know that it's a completely different approach
31:51
from from the statistical generative you know AI, and you're not going to have
31:56
these surprises of where it says some with great confidence that is just completely utterly
32:02
wrong. That may be a good mirror of some humans intelligence. Plenty of
32:07
people who also say things with high degree of confidence that are completely wrong,
32:10
but maybe not the one we want a mirror in. RAI. Uh,
32:15
well, Peter, this has been a fascinating conversation. I really appreciate you
32:19
making the time today. That's given me a lot to think about, and
32:22
I'm sure the audience as well. I do have a couple of traditional closing
32:24
questions I'd love to ask you before we before you finish up here, and
32:30
the first of them is what's the best piece of career advice you've ever got?
32:34
Well, for me personally, I can kick myself that I only started
32:37
my first business when I was twenty five. I should have started when I
32:44
was sixteen. Now this is not going to be, you know, true
32:49
for everyone, but for me, it's until you actually if you know,
32:52
if you aspire to run a company or to be part of the you know,
32:58
the core team, you really just need to get the experience. You
33:02
know, it's just completely different working for a boss as an employee than having
33:08
the responsibility of running a company. So if that's something you want to do,
33:13
the sooner you can start. I mean, whether it's part time,
33:15
you know, whether you do it even for free, just to gain the
33:20
experience, but to me that you know, if I have to say one
33:24
regret that I only started at twenty five. Interesting reminds me of advice about
33:30
parenting that you can never really understand it until you do it. And if
33:34
you're waiting for a better time, like there isn't really a better time,
33:37
you just got to go out there unprepared as you are and learn on the
33:39
job, because it's really the only way to learn. My last question was
33:43
maybe we've been talking about I mean, this may be easier one for you
33:45
because we've kind of been going on this all days, but it's always ask
33:49
everybody what's one bold prediction for the future that we are going to have AGI
33:53
and I believe it will be fantastic for humanity. It will really help us
34:01
manage the many difficult challenges that are facing us. Well. I love that
34:06
that's a great place to leave it later, because it's certainly I think possible
34:10
and would be an amazing opportunity and exciting development as we get there. So
34:15
I hope you're right on this one, and we'll be watching carefully because it's
34:17
a very exciting time and lots of stuff going on. So thank you so
34:21
much. For making the time to join us on the podcast today. Yes,
34:22
thanks for having me. It's fun. Upstart partners with banks and credit
34:28
unions to help grow their consumer loan portfolios and deliver a modern, all digital
34:32
lending experience. As the average consumer becomes more digitally savvy, it only makes
34:37
sense that their bank does too. Upstarts AI lending platform uses sophisticated machine learning
34:44
models to more accurately identify risk and approve more applicants than traditional credit models.
34:49
With fraud rates near zero, Upstarts All Digital experience reduces manual processing for banks
34:55
and offers a simple and convenient experience for consumers. Whether you're looking to grow
35:00
and enhance your existing personal and auto lending programs or you're just getting started,
35:06
Upstart can help. Upstart offers an end to end solution that can help you
35:09
find more credit worthy borrowers within your risk profile. With all digital underwriting,
35:15
onboarding, loan closing, and servicing, It's all possible with Upstart in your
35:20
corner. Learn more about finding new borrowers, enhancing your credit decisioning process,
35:25
and growing your business by visiting upstart dot com Slash four dash Banks That's upstart
35:30
dot Com slash four dash banks. You've been listening to Leaders in Lending from
35:36
upstart. Make sure you never miss an episode. Subscribe to Leaders in Lending
35:40
in your favorite podcast player using Apple Podcasts. Leave us a quick rating by
35:45
tapping the number of stars you think the show deserves. Thanks for listening,
35:47
until next time.