Latest / Founder Mode / AI as a Financial Co-Pilot with Shain Noor

Transcript

AI as a Financial Co-Pilot with Shain Noor · Founder Mode

0:00
People are adding to Sylvia assets that they would never
0:03
even tell their French advisor. And a lot of this stuff is like
0:06
they don't want to get judged for it, right?
0:08
And so they'll ask a bunch of dumb questions, which are really
0:11
just questions that you want to know answers to about your own
0:14
financial life, which they don't get to to a financial advisor
0:16
that they call once 1/4. You're listening to Founder
0:35
mode. Where we talked to builders
0:36
turning complex problems into usable systems.
0:38
Today's episode is about money, decisions and how AI might
0:42
finally help people think instead of react.
0:44
Yeah, our guest is building tools that don't just crunch the
0:47
numbers, they help people reason.
0:48
So Kevin, back in the day, I was a Big Mint user, I tried
0:51
Personal Capital, I tried all of them.
0:53
And and you've bounced around a bunch of these tools too.
0:55
Now that we're in the age where AI gets applied to everything,
0:58
what actually stands out to you in finance?
1:00
What's what's differentiated and what's just a wrap?
1:02
My sense of all of these tools have always been like, men.
1:05
Finance is just messy. Like very rarely you just have
1:08
like one bank and one credit card.
1:10
Like it just becomes overtime and especially if you've been
1:12
working and, and you know, thinking about like retirement
1:14
accounts like that, it constantly just sort of spews.
1:16
It should be that AI makes it easier for all those alternative
1:20
asset classes to come in right, Because it basically it you
1:22
know, it's it's just language we're saying, hey, I have this
1:24
collection go value, help me value it help me kind of track
1:27
it and put it in into My Portfolio.
1:29
The. Fact that it doesn't do things
1:30
and it's how it is like this view only sort of like reasoning
1:33
only like, hey, it just looks and gives you advice like, well
1:36
you're like, well, even if and gave it logins like it's it's
1:38
whole design principle is like do nothing like from an action
1:41
point of view. And so like at least initially,
1:43
I think that to me is like giving me more sort of, you
1:46
know, sort of trust to sort of like.
1:48
Yeah. And it's not going to do
1:49
anything, it just has the data. Yeah, awesome.
1:52
What about you, man? Like are there other things that
1:54
you think that you would want for this conversation?
1:55
Like, I'm curious what you know, what kind of magic is Shane
1:59
going to bring? Us I don't know for me it's like
2:01
is somebody going to finally build the financial advisor app
2:04
that tells me things that every other financial advisor has told
2:07
me before that I actually listen to that that tension feels like
2:10
a core part of this is like is what Shane and team building
2:13
something that makes me go OK, finally I'll listen to this
2:15
this. AI, I think there's some truth
2:17
to that. I mean, they always say like,
2:18
you know, bet on what you use. So I don't know.
2:20
Shane's a founder and engineer and really just focused on using
2:23
AI to improve financial decision making.
2:26
You bring them out. Let's do it.
2:38
Shane, welcome to founder Mode. You're building an AI copilot
2:41
for money decisions. Today, we're talking about how
2:43
AI can actually help people think more clearly about their
2:46
finances without adding noise or stress.
2:48
Thanks for having. Me.
2:49
Yeah, we're we're glad to have you here, Shane.
2:51
And you know, AI money, it's a fun time to be alive.
2:53
You you've worked deeply in AI and systems engineering.
2:56
What problem finally pushed you to build the product around
2:59
personal finance? Yeah.
3:00
I mean, I think the main thing here was if you look at the
3:03
history of all these personal financial products, it's all
3:07
about clicking, right? So Robin Hood made it easier to
3:10
buy a stock. You know, there's low trading
3:13
fees, all that stuff. And so the entire history of how
3:16
personal finance has worked, I don't think has focused on the
3:19
decision factor of any of this stuff, right?
3:22
And I think that is where most users stall is, OK, that's cool.
3:25
I know how to do these things. I'm learning about retirement,
3:28
I'm learning about investing. But when do I actually do these
3:32
things throughout my life? And I think that decision factor
3:35
is where AI comes in. When you first started, what did
3:38
you realize that most people get wrong about managing their
3:40
money? Because I mean, like, I've used
3:41
a ton of these tools like Mint and Personal Capital,
3:44
Wealthfront, like all these different things.
3:46
And they've each kind of come at it from a different direction of
3:48
like, you know, it's the neo bank or the a new way of
3:51
managing money. Yeah, And going back to
3:52
mint.com, all this stuff was aggravating.
3:54
It was all budgeting focus, transaction focus.
3:57
If you look at Personal Capital, I mean, then it upgraded to
4:00
looking at your net worth. If you look at Copilot, then
4:03
they upgraded to having a better design and dashboards and graphs
4:07
and visuals. And there's robo advisors like
4:09
you said, Wealthfront and all that stuff.
4:11
But at the end of the day, none of this still is tackling
4:13
reasoning, right. And so just because now we're on
4:16
the AI native way for this, you can actually get a tool that
4:20
thinks about your situation, gives you concrete next steps
4:23
basically on what to do that that action behind it and why.
4:27
You've described AI as this kind of a copilot, not an autopilot.
4:31
What does that distinction mean in practice, in the way you've
4:34
built Sylvia? Yeah.
4:34
I mean, we think about this a lot in totally how much do you
4:38
want to go down the autopilot route, right, versus keeping the
4:42
human in the loop. I think there's three aspects of
4:44
this and all it comes down to trust, transparency, like
4:46
credibility and all that stuff, right?
4:48
So how much trust can you show the user?
4:50
And that's all going to be about accuracy, being transparent.
4:53
So again, it comes back to the reasoning of when you're showing
4:57
a user recommendation, can you show the traces of the AI,
5:00
right, of like how it even came about that decision and also
5:03
referencing back to sources, right?
5:06
So if there are values that you are referencing to or if there's
5:08
tax codes that you are mentioning in the responses, can
5:11
you link out to the sources so the user knows that this is much
5:15
more accurate than something else?
5:17
I think all that plays a factor in how much of this is copilot
5:21
versus autopilot. Obviously the execution factor
5:24
of this, you can either have guardrails, right?
5:26
So if someone sets up an agent that says, you know, buy and
5:29
sell on, you know, market open and close or whatever it is or
5:32
has certain thresholds, those sorts of things are are fine.
5:35
But if you are trying to make the agent to like irreversible
5:39
big decisions and obviously there's going to be some sort of
5:42
factor of trust and credibility that needs to be involved there,
5:45
right? And so the way we think of it as
5:47
small, simple, reversible decisions can be fully on
5:51
autopilot. What are those things that you
5:53
think about today? What is it that today's kind of
5:55
AI does well? And then what do you think that
5:58
humans should still be in control?
5:59
And then if you want to pull back the curtain, like what LLMS
6:02
do you guys use the most? Anthropic by far, that's the
6:04
main one we use. Obviously we have like a zero
6:06
day retention policy, all that stuff.
6:08
We sort of focus on obviously the financial context, memory
6:11
and having all that data, you know, on the platform and
6:14
personalizing, right? And so obviously building on top
6:17
of Anthropic has been great. They're great at, you know,
6:19
finance is what it seems like and also coding.
6:22
And so I think these general alums, I mean, I think there's
6:25
probably like 3 things that matter the most.
6:28
Even when it comes to like the what when real money is
6:30
involved, it's obviously the memory of your financial life.
6:33
So of course you can go and chat to until you can upload all
6:35
these documents, like you said, I mean, it's a lot more
6:37
hardwired. There's not enough structured
6:39
data there. The whole point of Sylvia was
6:41
that you can and do all this stuff in your preferences and
6:44
your profile page. You have direct connectivity,
6:46
real time access to all your bank accounts through Plaid.
6:50
Obviously we look at tickers, we do all that stuff.
6:52
And so it's all about reasoning on your own personal situation,
6:55
not just a general LLM tackling this.
6:58
And then obviously the verification layer, right?
7:00
And so we have dedicated agents behind the hood that kind of do
7:05
sort of like this ping pong method of making sure the agents
7:08
are even critiquing the agents of what tool calls they're
7:10
using, critiquing the data, the numbers that are coming back.
7:13
And it's good because you can do that.
7:15
You can build those guardrails when you know it's your domain
7:18
right instead of a generalized LM.
7:20
I love that you're talking about personalization because I
7:22
completely agree. Given how personal finance is,
7:25
how do you build trust in an AI system that's making sense of
7:29
someone elses financial life? Actually, we have distribution
7:32
through through my cofounder and that obviously builds a layer of
7:36
trust with the independent investor audience.
7:38
I think you you want to explain to them how these AI agents can
7:42
actually go to work for you and benefit you in the long run,
7:45
right? The whole goal here is how can
7:47
you increase the independent investors net worth?
7:49
That's the KPI, right? If you use Sylvia, you should be
7:52
able to grow your net worth. And if you can't do that, then
7:55
the job is not being done. That's one thing that surprised
7:58
you about how people actually use the product.
8:01
Like what kind of what's they have this sort of context to
8:03
where AI because I agree with you.
8:04
Like again, I can upload all the statements, but when I look at
8:07
Sylvia, it's just easier, right? Like it just solved the context
8:10
thing and keeps it up to date. I think like there's more work
8:12
to do. Like I have a ton of like Angel
8:14
investments, right? And like even finding out like,
8:16
you know, what the current status of that company and that
8:19
human is like is the person that I signed this with even working
8:21
at the company anymore. Like, you know, like, but what's
8:24
something that surprised you? Because I'm sure there's stuff
8:26
that you guys have seen that you're like, wow, that wasn't
8:28
what we expected. I think the biggest thing was,
8:32
you know, financial advisors, right?
8:34
And so most of these users either had financial advisors or
8:37
currently had one and they were still using Sylvia, whether it
8:40
be back in the day, financial advisors would be like, no,
8:42
don't show me the crypto. I don't care about it, all that
8:45
stuff. And it's like people are adding
8:46
to Sylvia assets that they would never even tell their financial
8:50
advisor. And a lot of this stuff is like
8:52
they don't want to get judged for it, right?
8:54
And so they'll ask a bunch of some questions which are really
8:58
just questions that you want to know answers to about your own
9:00
financial life, which they don't get to to a financial advisor
9:02
that they call once 1/4 right. And so this is like an always on
9:06
24/7 financial advisor, your personal CFO that's with you.
9:10
You can ask judgmental, you can ask whatever you want, right?
9:14
And so there's a lot more context being thrown in there
9:16
like, you know, users are asking, should I buy this purse
9:19
for my what? Like it's a very different
9:21
question than you would ever go to a financial advisor for,
9:24
right? And so that's one aspect is like
9:26
the emotional, judgmental aspect, which we saw a big trend
9:29
in AI regardless. So everyone using chat, TPT and
9:32
all that, almost like a friend. The other aspect was the
9:34
proactive nature, right? And so Sylvia first launched as
9:37
almost like a Q&A, right? You you can chat with her, you
9:40
can ask questions, but we notice a lot of people would actually
9:44
prefer the other route, which is you set up some of the stuff for
9:48
me, right? And so we had a lot of users
9:50
reaching out to us of like, Hey, can I get another call with you
9:53
and you help me sub Sylvia for like, how can I become a power
9:56
user, right? I don't know which questions to
9:58
ask. It's like an entire black box
10:00
similar to chat to be seen all that stuff.
10:02
I mean, obviously on Twitter and stuff it comes out that ho, hey,
10:05
like try this prompt or, you know, the tricks of like each me
10:08
like I'm five years old, all these sorts of things that we
10:11
just didn't have embedded into the product, especially when it
10:14
is about your financial life. Are these workflows, It's like
10:16
how can you do one click workflows of personal financial
10:20
statement, all that stuff which are you know, slowly we are
10:22
adding to the product. Do you use?
10:23
This term personal CFO, which I think is interesting.
10:26
I'm a, I'm a marketing and growth guy.
10:27
So I'm, I'm very fixated on that positioning because I think for
10:30
a lot of people, you know, that are maybe less sophisticated in
10:35
the business world, they don't even know what a CFO does.
10:37
So talk to me about kind of how you landed on we're, we're
10:40
calling this kind of like a personal CFO versus A, you know,
10:44
AI financial advisor or one of the many other ways you could
10:46
have described it. Exactly.
10:48
We didn't want the consumer to think of concrete stuff like
10:52
what does a financial advisor do and go into this thinking.
10:55
That's the capability of Sylvia. We almost want them to think
10:58
broader of come to us with a wish list of things that you
11:02
would imagine a personal CFO would be able to do.
11:05
Look at your entire portfolio, look at the news, track all this
11:09
stuff for you and give you insights and advice just like a,
11:13
you know, would that's tackling your personal financial life.
11:16
And so that's kind of how we went about it.
11:18
Yeah, I like that a lot and I think that I've enjoyed that
11:21
part of the product the most. You know about like, you know,
11:24
even random things like, you know, how does the current macro
11:28
events of the world impact My Portfolio, right?
11:31
And I think, you know, it's your Co founder pop posts these kind
11:34
of like snippets and screenshots of these kind of things.
11:36
You know a lot in his Twitter account.
11:38
And I've asked some of these questions myself, just like,
11:40
hey, this thing's going to happen or if this thing happens,
11:43
what should I shouldn't do or how would that impact My
11:45
Portfolio? And then like I said, these,
11:47
some of these deeper research ones where it's like, Hey, go
11:49
figure out what's up with my like Angel portfolio, Like
11:51
here's all my safe notes. Go figure this out.
11:53
But to me, like the piece where I feel like when you say CFO, I
11:57
think of the person that like mahogany office sitting at a
11:59
desk, like, you know, giving me the advice, but they're
12:02
definitely not going to like go do the work.
12:04
Like there's no shovels in the CFO's office.
12:06
But it's like, I know what that entails.
12:08
What, how do you guys think about that?
12:10
And are there certain things where you're like, damn, we, we
12:13
would start to take action on those if, if users wanted it or
12:16
trusted it and maybe those are the places to start versus some
12:19
of the big ones where it's like, hey, go rebalance My Portfolio,
12:22
you know, to survive this latest job change, right?
12:26
Or something like that. I truly.
12:27
Do still believe that building in the kind of verifiable
12:31
reasoning layer 1st and making sure that that is at 99% of
12:37
accuracy is much more important because the execution layer has
12:41
been solved? Yeah, talk, talk about the
12:43
measurement framework that you're using to kind of
12:44
understand, you know, stickiness and engagement in the platform.
12:48
What it what are those signals that tell you that people are
12:51
like, this is, you know, this is the thing that I'm going to use
12:54
for the next X number of years to manage my finances and think
12:57
about mental health and Wellness.
12:59
So when we first launched with just Q&A, we noticed that right
13:04
when there was a bug in our daily summary e-mail and for
13:06
whatever reason in the beginning of Sylvia, it didn't go out to
13:09
like 50% of our users that night.
13:12
We got so many complaints, so many emails inbound of like,
13:15
hey, that's the thing that I look at every day.
13:17
Once we introduced those proactive alerts, we noticed
13:20
that the high net worth user base has a much higher retention
13:23
on radar than they do with anything else in the entire
13:26
product. Yeah.
13:27
And and so, but to me, it's that, it's that I agree.
13:29
Like it. Yeah, that e-mail I see every
13:31
day, I'm like, what are the first 3 digits?
13:33
And I'm like watching those. And then what are the two things
13:34
that move to the bottom? Like I'm just kind of like, you
13:36
know, playing with that game. It is incredible.
13:39
Like how well that thing keeps the context up.
13:42
But I also think there's like this, you know, to your point,
13:45
like the radar thing I've touched on a little bit, but I'm
13:47
kind of like the lazy, but I'm like, and then I wish radar
13:49
would like self recommend radars.
13:51
Like it's, you know, it's only like, you know, just like it's
13:53
like the ChatGPT question where it's like, man, I wish ChatGPT
13:56
would just tell me like how the best users are using it.
13:58
And like, what's the thing that it's the, you know, when I
14:00
interview people, like vendors always like, what are you the
14:02
best at? Because I'm like, look up agency
14:04
people, vendor people. They're like, oh, I can do all
14:06
these things. I'm like, no, no, that's fine.
14:07
But what are you absolutely the best at?
14:09
Let's just start with that. And then I know that I can kind
14:11
of grow from there. And I'm curious like, and maybe
14:14
that's the question, what is Sylvia the best at today?
14:17
And like you know, what are your best users getting the most
14:21
value from and what are they doing?
14:22
A couple. Things on the radar thing I did
14:24
tweet the other day of hey, ask Sylvia, give me radars that are
14:28
topical to what's going on in the news right now and what I
14:31
should be following. And then she came back with, you
14:33
know, 7 radars and then she had my created it for me.
14:36
And then someone, you know, one of our users had commented, I
14:38
want her to do this automatically every day.
14:41
And so that is a huge aspect of this.
14:43
And it goes back to like, hey, I don't want to go in and learn
14:45
how to set things up or like figure out which questions I
14:48
should be asking or what I should set up.
14:49
She should be able to do that for you.
14:51
And I think that's the next step is the entire proactivity layer
14:54
of of Sylvia. Yeah, as you're saying that, I'm
14:56
thinking like there's just be a green button at the bottom of
14:58
that summary that says, do you want me to set up some writers
15:00
for you tomorrow? Like I I would be shocked if a
15:03
ton of people don't click that button. 100% and and that's the
15:05
power users are using radar to the advantage and using Sylvia
15:08
as almost like a yeah, almost like a partner, right?
15:11
It's like, what should I do? Hey, you are my, you know,
15:14
wealth advisor, my personal CFO, whatever.
15:16
Ask me questions of what you need answers for to learn more
15:20
about me and then suggest me which radars, which research,
15:23
like what should I be doing everyday?
15:24
And those users are by far, you know, the most, you know, user
15:28
heavy chatting heavy users. How do you guys handle ensuring
15:33
that the guidance that Sylvia's giving and the kind of way that
15:38
you're structuring its understanding of a user's
15:41
financial data and inputs such that it doesn't make stuff up,
15:45
it doesn't give the wrong answer, right?
15:47
You mean a lot of this is storing everything in a
15:49
structured database, to be honest, right?
15:51
And then having that Fact Check loop, right?
15:54
And then there's a verifiable, you know, check loop agent that
15:57
essentially goes back and checks on all this stuff, right?
16:00
If there is something that Sylvia doesn't know, she should
16:03
say, I don't know. Even if, you know, there has to
16:05
be some sort of confidence metric there, whether it be over
16:08
90% or whatever, that she can then say like I don't have the
16:11
answer to that, then the user can kind of, you know, guide her
16:14
in the right direction. Preferences is another aspect of
16:16
this, right? And so when there is something
16:20
that she tends to get wrong, often the user can actually go
16:23
in on their profile page and set things up in preferences, which
16:26
are essentially what should she keep top of mind every time she
16:29
responds to you. And if it is something that
16:31
you've talked to her about always in other chats, but she's
16:35
not saving that into preferences.
16:36
So these are like grounded, you know, facts that you have a
16:40
spouse or what are the ages that I'm thinking of, you know, a
16:43
year from now to buy a house, all that stuff that you would
16:46
probably tell, tell a wealth advisor.
16:48
Those are things that should be grounded in truth in your
16:51
preferences. And in that way, she always
16:53
keeps that stuff top of mind because the most important thing
16:56
is context, right? And obviously the context gets
16:59
cluttered. There's going to be more and
17:00
more chances of hallucinations. Happening.
17:02
Yeah, it's great. I want to talk a little bit
17:04
about the future before I do that.
17:05
I'm just curious, like we talked to a lot of kind of founders in
17:08
AI, like how big is the team? Like because I think one thing
17:11
that's kind of always impactful is like when we first talked,
17:12
like the team was very small. I'm guessing the team is still
17:15
relatively small, but I'm curious like how big is the team
17:17
today? How do you see that evolving?
17:20
Now we have, you know, 6 engineers and so it's still
17:23
fairly small team. We have a designer, we have, you
17:25
know, OPS now that we've obviously merged with pro cap,
17:28
there's obviously that exact team.
17:30
But I think the the main thing is, yeah, we're up against these
17:33
legacy, legacy companies that have huge, huge amount of
17:36
employees, which I see as a, a huge advantage to us, right?
17:39
I mean, we can move fast, we can push code to production so
17:42
quickly. And so, you know, it's a great
17:44
advantage for us. Yeah, there's no way Fidelity's
17:47
adding that feature this, you know, for a long time.
17:52
It's on that point. Like how do you see AI sort of
17:56
replacing that traditional advisor or fundamentally just
17:59
changing how their role? Like do you imagine like the
18:01
wealth people that I don't know, Vanguard, Charles Schwab,
18:05
Fidelity sort of a using this as a tool in their portfolio?
18:08
Or do you imagine this just sort of running away and people just
18:11
moving out of these financials? Like where do you sort of see
18:15
from your seat, you know where you are today?
18:16
I think it's. Going to be hybrid right and so
18:18
similar to engineers, right and so if you're not using AI today
18:21
as an engineer, it's probably it's probably too late.
18:24
But what you're saying with junior engineers with you know
18:27
they're not able to get a job or you know stay up to date with
18:30
everything that's going on. I think this similar thing is
18:32
happening with financial advisors, right?
18:34
Most of them are cookie cut of what the portfolio plan looks
18:38
like for the range of 1 to 3,000,000, for the range of five
18:41
to 10. Like these are just plans.
18:43
And then they have maybe a set number of clients that they can
18:46
only interact with because they only have so much time.
18:48
And then there's once a quarter phone calls, right?
18:50
And so the relationship is lacking there.
18:52
The collaborative nature is lacking there.
18:55
And I think that's where all these independent investors are
18:57
going towards. So of course, there's going to
18:59
be financial advisors who leverage the AI, increase their,
19:02
you know, client base of client lists and they'll be able to
19:05
tackle, you know, a lot more clients at once and also give
19:08
the power to the client of being able to chat to an LLM.
19:11
And then maybe the advisor can, you know, see what's going on
19:14
and go chime in. And they need to, right.
19:16
But it's always going to be a different type of person for.
19:18
Maybe you're going to build like an agency or advisor mode where
19:21
they can go in and manage, you know, 10 clients versus, you
19:24
know, just login. Well, the interesting thing is
19:26
one of the most requested features right now is that a
19:28
multi portfolio view which would tie into that anyways.
19:31
So it's just building out the tenant, tenant architecture,
19:33
right? So if I can add my spouse, I can
19:34
add my children to Sylvia and then they can have some sort of
19:37
read, write have been privileges of course, and I can add anyone.
19:41
Oh, you, you've kind of got your finger on the pulse of AI and
19:44
finance right now. And this is a weird question
19:47
that we ask a lot of folks building an AI that, you know,
19:49
every time we ask it, I feel like the time horizon gets
19:51
closer and closer. Oh, what do you think is going
19:53
to happen in three or five years, like 15 minutes from now?
19:55
It feels like it's hard to know what's going to happen.
19:57
But from your perspective, zooming out from what what
20:00
humans might be replaced with AI?
20:02
Like how do you see the broader kind of macro landscape changing
20:06
with with the power of AI and how that can impact the finance?
20:10
Everyone turns into a manager is kind of my, my, my look and feel
20:14
of all this stuff, right? I would just be managing agents.
20:17
So that's what I can speak. On this has been awesome.
20:20
Shane, where can people follow your work and, you know, see
20:23
what you're building? Yeah, everything on
20:24
cfosylvia.com. Go sign up, get a personal CFO
20:28
and then at Shane Codes on Twitter.
20:30
Awesome, Shane. This was clear, practical,
20:32
grounded. Thanks for joining us.
20:33
Thank you guys. Appreciate it.
20:35
Have a good one, That was awesome.
20:36
What stuck with me was that Sylvia's not trying to build
20:38
another dashboard. Every other app in this space is
20:41
like logins. See your net worth.
20:42
Cool, now what? Like he skipped past that and
20:45
went straight to the reasoning. It's like the product actually
20:47
thinks about your situation. Yeah, and the part that hit me
20:49
was when he talked about users asking Sylvia things that they'd
20:52
never asked their financial advisor, like should I buy this
20:54
purse for my wife? That's not a portfolio question.
20:56
That's a trust signal. People had this thing like a
20:59
judgment free partner and not a tool.
21:01
Yeah, the retention insight was wild.
21:03
Like they had a bug where the daily summer e-mail didn't go
21:05
out and get flooded with complaints.
21:07
And I get it was funny, I don't think I complained, but I
21:09
definitely noticed when it didn't work.
21:10
You're, you're maniacal about your inbox, though, So it's, you
21:13
know, it's not really an app, right?
21:14
It's this push the proactive layer.
21:17
Kevin, what are the top five takeaways from today's
21:19
conversation with Shane? Yeah, there was a bunch.
21:21
It was hard to pick 5, but I think, you know, finance apps
21:24
have optimized for clicking and now, like AI really optimizes
21:27
for deciding. And then this notion of like,
21:29
show your reasoning, like the trust that he's building comes
21:32
from the traces, the sources and saying, I don't know when the
21:35
confidence is low. And 3rd, it's like your best
21:37
users are won't really set that up for themselves.
21:40
Like they want help. They're like just, you know, how
21:42
can I one click this thing and like get all my data in there
21:45
and recommend like what things I should be doing or watching and
21:48
then forth, like nail that reasoning layer.
21:51
So really understanding and making sure that the reasoning
21:53
behind the things that Sylvia was saying, you know, is not
21:56
been solved by robo advisors. But it's something that's hard
21:58
to go fix later if you're if you're like thinking isn't smart
22:02
upfront. And then finally, like this
22:04
notion that small teams like that is the Moat man shipping
22:07
fast, being able to roll out new tech and like the, you know,
22:10
fidelities of the world aren't going to be able to add things
22:13
like this, you know, very quickly.
22:14
Probably not. Yeah.
22:15
If this episode changed how you think about money or AI, share
22:18
it with someone who's trying to make better decisions.
22:20
And make sure you're subscribed to Founder mode so you don't
22:23
miss conversations like this. That's it for today.
22:26
The real question isn't whether AI can manage your money, it's
22:29
whether you'll put your money where your AI is.