1
00:00:00,775 --> 00:00:02,951
>> Anthony Weaver: This episode is sponsored by

2
00:00:03,023 --> 00:00:04,080
podvantage AI.

3
00:00:04,080 --> 00:00:06,495
>> Sumedha Rai: Uh, to a certain extent, it's a supervised learning

4
00:00:06,535 --> 00:00:09,391
algorithm, which means that you tell the machine, you take this

5
00:00:09,423 --> 00:00:12,399
type of content, you put it into this category, and it's going

6
00:00:12,407 --> 00:00:15,263
to do it really fast. So those type of jobs I do feel

7
00:00:15,319 --> 00:00:17,911
are going toa get replaced. But what can you do? You can

8
00:00:17,943 --> 00:00:20,480
see how to upskill yourself in that,

9
00:00:20,480 --> 00:00:23,395
um, domain. Can you do something,

10
00:00:23,620 --> 00:00:26,567
um, similar to the domain, similar to the expertise, but

11
00:00:26,591 --> 00:00:29,111
can you do it using tech? Because if you're a person

12
00:00:29,263 --> 00:00:32,001
who is now using thei solution

13
00:00:32,103 --> 00:00:35,061
to enable your job, you're already an expert in that

14
00:00:35,093 --> 00:00:38,045
field and people need that. People need more

15
00:00:38,125 --> 00:00:40,973
workers who understand AI so that they can

16
00:00:40,989 --> 00:00:42,905
integrate it. That's in demand.

17
00:00:45,205 --> 00:00:48,149
>> Anthony Weaver: Welcome back everybody, back to another exciting show of day.

18
00:00:48,197 --> 00:00:51,045
About that Wallet podcast. I'm, um, your host, Anthony Weaver, where

19
00:00:51,085 --> 00:00:53,885
we are focusing on the Saich generation,

20
00:00:54,005 --> 00:00:56,909
building strong financial habits so that they can spend

21
00:00:56,957 --> 00:00:59,933
money, talk about money and enjoy their money with

22
00:00:59,989 --> 00:01:02,733
confidence. Today I have somebody who has

23
00:01:02,829 --> 00:01:05,787
been in and out, out of the financial

24
00:01:05,851 --> 00:01:08,739
realm, but also with the AI spin

25
00:01:08,827 --> 00:01:11,787
and I cannot wait to get this started. So how

26
00:01:11,811 --> 00:01:13,495
you doing today, Samita?

27
00:01:13,640 --> 00:01:16,547
>> Sumedha Rai: Uh, I'm doing very well. I'm very happy

28
00:01:16,571 --> 00:01:19,507
to be on the show and um, I'm really

29
00:01:19,571 --> 00:01:22,499
looking forward to these conversations because I've enjoyed your podcast

30
00:01:22,547 --> 00:01:25,395
in the past. So it's an honor to be one of the guests.

31
00:01:25,555 --> 00:01:28,467
>> Anthony Weaver: O thank you so much, so much. Um, and I'm

32
00:01:28,571 --> 00:01:31,379
really dying to really dive into your story

33
00:01:31,467 --> 00:01:33,415
because you've been,

34
00:01:34,220 --> 00:01:36,495
um, I would say internationally

35
00:01:37,275 --> 00:01:40,051
exposed to so many different things in the world.

36
00:01:40,163 --> 00:01:42,731
And what brings you here today

37
00:01:42,883 --> 00:01:45,775
is what brings this question up is,

38
00:01:46,740 --> 00:01:49,395
um, so many people are pretty much

39
00:01:49,435 --> 00:01:52,251
scared of how powerful AI

40
00:01:52,283 --> 00:01:55,179
is getting and what is your take

41
00:01:55,347 --> 00:01:57,495
on the impacts on the human psyche?

42
00:01:58,500 --> 00:02:00,175
>> Sumedha Rai: Um, so I think

43
00:02:01,115 --> 00:02:03,923
with any new change that comes into

44
00:02:03,939 --> 00:02:06,135
our world, there is

45
00:02:06,875 --> 00:02:09,771
a lot of excitement, be a lot of

46
00:02:09,803 --> 00:02:12,495
panic, see a lot of misinformation.

47
00:02:13,155 --> 00:02:16,131
And I see this with AI as well.

48
00:02:16,203 --> 00:02:19,035
People are really excited to try out the new tech. Every

49
00:02:19,115 --> 00:02:21,490
business wants it as a, ah,

50
00:02:21,611 --> 00:02:24,443
solution so that they can go out there and

51
00:02:24,499 --> 00:02:26,739
tell everyone, tell their competitors or the

52
00:02:26,787 --> 00:02:29,715
consumers're powering our solutions through AI

53
00:02:29,795 --> 00:02:32,681
and generative AI. Um, the

54
00:02:32,713 --> 00:02:35,609
other hand, there are also people, uh, really panicking

55
00:02:35,657 --> 00:02:38,633
about job security and they're saying it's going to take

56
00:02:38,689 --> 00:02:41,050
away all the jobs and um,

57
00:02:41,065 --> 00:02:43,945
replace the human jobs that we have

58
00:02:44,025 --> 00:02:46,929
with like, I don't know, some robotic jobs or A.I.

59
00:02:46,977 --> 00:02:49,905
jobs. Um, I think the

60
00:02:49,945 --> 00:02:52,817
first step to any New change or any

61
00:02:52,841 --> 00:02:55,097
new tech in this case is

62
00:02:55,201 --> 00:02:58,121
understanding. So since everybody

63
00:02:58,153 --> 00:03:00,969
is talking about it, oftentimes what happens

64
00:03:01,017 --> 00:03:03,739
is that we get really affected

65
00:03:03,827 --> 00:03:06,699
by other people's opinions. And we see this

66
00:03:06,787 --> 00:03:09,095
one article that talks about

67
00:03:09,395 --> 00:03:12,331
how, you know, the next 10 years are gonna be really grim

68
00:03:12,363 --> 00:03:15,123
for us, and another article that talks about how it's

69
00:03:15,139 --> 00:03:17,987
gonna completely take over all the human jobs and

70
00:03:18,011 --> 00:03:20,787
we start feeling the sense of panic. So at

71
00:03:20,811 --> 00:03:23,571
this point what I'd like to do is I'd like to tell the

72
00:03:23,603 --> 00:03:26,547
person, um, is this your

73
00:03:26,571 --> 00:03:29,507
opinion, uh, which has been formed because of

74
00:03:29,531 --> 00:03:32,265
some research that you have done, or is this just

75
00:03:32,345 --> 00:03:35,273
something that you've been hearing a lot? And I totally understand that if

76
00:03:35,289 --> 00:03:38,169
you hear something repetitively, um, it will

77
00:03:38,257 --> 00:03:41,169
make an impact on you. So the first thing that you need to do

78
00:03:41,217 --> 00:03:43,945
is understand what the technology is.

79
00:03:44,065 --> 00:03:47,065
Understanding is the absolute key. And in

80
00:03:47,145 --> 00:03:49,777
this case specifically, knowledge will empower

81
00:03:49,841 --> 00:03:52,817
you because um, when you're introduced

82
00:03:52,841 --> 00:03:55,761
to new topic and you don't know anything about it, you go online

83
00:03:55,793 --> 00:03:58,665
and you read five articles, um, you feel really

84
00:03:58,745 --> 00:04:01,209
confident and you're ready to have a conversation about it.

85
00:04:01,377 --> 00:04:02,409
>> Anthony Weaver: There'five articles.

86
00:04:02,497 --> 00:04:05,431
>> Sumedha Rai: Y. I mean I've seen people who go like,

87
00:04:05,463 --> 00:04:08,311
yeah, I'm not too sure about it. And then like one

88
00:04:08,343 --> 00:04:11,191
week later they'll come back to me and say I did my research

89
00:04:11,383 --> 00:04:14,167
and uh, they want to talk about it now. So there's a sense

90
00:04:14,191 --> 00:04:16,975
of confidence that you get when you understand

91
00:04:17,055 --> 00:04:19,823
something. And I'm not saying that you're going to understand all the technical

92
00:04:19,879 --> 00:04:22,351
aspects of it, exactly how the model's working,

93
00:04:22,503 --> 00:04:25,367
but just having an understanding of what this

94
00:04:25,391 --> 00:04:28,223
solution looks like. What can AI actually do,

95
00:04:28,359 --> 00:04:31,343
what kind of jobs can it replace? Because at the

96
00:04:31,359 --> 00:04:34,359
end of the day there are certain jobs that are taking

97
00:04:34,447 --> 00:04:36,785
say like 8, 8 or 10 human hours

98
00:04:37,205 --> 00:04:40,149
and they're repetitive in nature and they can be automated.

99
00:04:40,317 --> 00:04:43,260
And if those jobs can be automated, um,

100
00:04:43,260 --> 00:04:46,173
people and businesses will want to do

101
00:04:46,189 --> 00:04:49,173
it like that. It's going to be done at a fraction of a cost with less

102
00:04:49,229 --> 00:04:51,749
error because there is no human fatigue

103
00:04:51,917 --> 00:04:54,265
or um, there is less

104
00:04:54,420 --> 00:04:57,381
um, um, human error due to

105
00:04:57,453 --> 00:05:00,445
monotony. So they will become more efficient.

106
00:05:00,565 --> 00:05:03,525
But we also need to understand that there is like this whole

107
00:05:03,605 --> 00:05:06,573
new uh, spectrum of jobs that is not going to get

108
00:05:06,589 --> 00:05:09,315
affected just because we're not there yet.

109
00:05:09,355 --> 00:05:12,320
AI is just not there yet. So um,

110
00:05:12,320 --> 00:05:15,155
if you're panicking because you've been hearing

111
00:05:15,235 --> 00:05:18,163
about this a, ah lot and you don't understand

112
00:05:18,179 --> 00:05:20,651
it completely, go out there, read about

113
00:05:20,683 --> 00:05:23,539
it and um, if you're scared

114
00:05:23,587 --> 00:05:26,467
about your jobs, understand what is the nature of

115
00:05:26,491 --> 00:05:29,363
your job and how is ainna affect it. So I

116
00:05:29,379 --> 00:05:31,970
was talking to someone recently and um,

117
00:05:32,075 --> 00:05:34,707
she mentioned that as a part of

118
00:05:34,731 --> 00:05:37,539
her team, um, there was one

119
00:05:37,587 --> 00:05:40,289
person who was categorizing a lot of data

120
00:05:40,417 --> 00:05:43,377
into uh, 150

121
00:05:43,441 --> 00:05:46,281
categories. And she used to take 10

122
00:05:46,313 --> 00:05:49,233
hours, 15 hours to go through one sheet because like just

123
00:05:49,249 --> 00:05:51,897
the sheer amount of data that was on the sheet needed that much amount of

124
00:05:51,921 --> 00:05:54,609
time. And then she said, we started talking to this

125
00:05:54,657 --> 00:05:57,577
vendor and they said they can categorize it, which makes

126
00:05:57,641 --> 00:06:00,457
complete sense to me to a

127
00:06:00,481 --> 00:06:03,321
certain extent. It's a supervised learning algorithm, which means that

128
00:06:03,353 --> 00:06:06,241
you tell the machine, you take this type of content, you put it

129
00:06:06,273 --> 00:06:08,859
into this category and it's going toa do it really

130
00:06:08,907 --> 00:06:11,483
fast. So those type of jobs I do feel

131
00:06:11,539 --> 00:06:14,291
arenna get replaced. But what can you do? You can

132
00:06:14,323 --> 00:06:16,840
see how to upskill yourself in that,

133
00:06:16,840 --> 00:06:19,775
um, domain. Can you do something,

134
00:06:20,540 --> 00:06:23,307
um, similar to the domain, similar to the expertise, but can you do

135
00:06:23,331 --> 00:06:26,163
it using tech? Because if you're a person who

136
00:06:26,219 --> 00:06:29,155
is now using thei solution to enable

137
00:06:29,235 --> 00:06:32,011
your job, you're already an expert in that field and

138
00:06:32,043 --> 00:06:34,939
people need that. People need more workers

139
00:06:34,987 --> 00:06:37,851
who understand AI so that they can integrate

140
00:06:37,883 --> 00:06:40,617
it. That's in demand. So I

141
00:06:40,641 --> 00:06:43,393
think, um, TLDR as we call

142
00:06:43,409 --> 00:06:43,965
it.

143
00:06:44,265 --> 00:06:45,005
>> Anthony Weaver: Yes.

144
00:06:45,070 --> 00:06:47,325
>> Sumedha Rai: Um, don't be scared

145
00:06:47,625 --> 00:06:50,377
immediately when you read something.

146
00:06:50,561 --> 00:06:53,409
Please go and do your research. That's gonna give you a little

147
00:06:53,457 --> 00:06:56,209
more confidence, that's gonna give you some more knowledge and see

148
00:06:56,257 --> 00:06:58,910
if you can take some steps, um,

149
00:06:58,910 --> 00:07:01,737
to go in a different direction. Or maybe

150
00:07:01,761 --> 00:07:04,161
you're just gonna realize, okay, um, I'm a

151
00:07:04,193 --> 00:07:07,193
nurse, this is not going toa get replaced anytime

152
00:07:07,249 --> 00:07:09,459
soon. So you'renn do better.

153
00:07:09,577 --> 00:07:12,551
>> Anthony Weaver: So what are your thoughts on working harder, not

154
00:07:12,623 --> 00:07:15,367
smarter first and then work

155
00:07:15,431 --> 00:07:16,155
smarter?

156
00:07:16,530 --> 00:07:18,715
>> Sumedha Rai: Uh, okay, I think

157
00:07:19,695 --> 00:07:22,655
it's a cool thing, uh, to ask. I think

158
00:07:22,735 --> 00:07:25,727
even with the tech that we have, the tech

159
00:07:25,751 --> 00:07:28,671
has already worked very hard. There are humans

160
00:07:28,743 --> 00:07:31,335
behind it that have worked hard

161
00:07:31,495 --> 00:07:34,115
to enable some humans to work smart.

162
00:07:34,775 --> 00:07:37,753
So, um, personally I feel

163
00:07:37,809 --> 00:07:40,297
any solution, any tech, has both

164
00:07:40,361 --> 00:07:43,097
these already embedded in it, whether you see it or

165
00:07:43,121 --> 00:07:45,929
not. So you can say that with the

166
00:07:45,937 --> 00:07:48,681
solutions you're going to work smarter, but also making the solution.

167
00:07:48,753 --> 00:07:51,673
Somebody worked hard to create a solution and now we're

168
00:07:51,729 --> 00:07:54,673
working hard to improve the solution because you

169
00:07:54,689 --> 00:07:57,569
might have seen it. Um, a lot of companies

170
00:07:57,617 --> 00:08:00,537
are coming up with new solutions every day. Not all

171
00:08:00,561 --> 00:08:03,305
of them are vetted. Some of them might be trained on

172
00:08:03,345 --> 00:08:05,889
biased data. Some of them might have problems

173
00:08:06,057 --> 00:08:09,031
related to knowledge gaps. Some of them might have been trained

174
00:08:09,183 --> 00:08:11,895
till a certain date so they don't have a lot of recent

175
00:08:11,935 --> 00:08:14,639
data. Some of them are not showing you the sources.

176
00:08:14,807 --> 00:08:17,295
We're still putting in a lot of hard work to make sure

177
00:08:17,335 --> 00:08:19,887
that um, it is a

178
00:08:19,911 --> 00:08:22,300
solution that is promoting uh,

179
00:08:23,087 --> 00:08:25,999
a better society, so to say. So

180
00:08:26,127 --> 00:08:28,999
we're still working hard and smart, but

181
00:08:29,087 --> 00:08:31,711
once we get to a point where we need to

182
00:08:31,743 --> 00:08:34,431
work a little less hard and a little more

183
00:08:34,503 --> 00:08:37,186
smart, that's when I would say, yeah,

184
00:08:37,250 --> 00:08:40,170
concentrate towards working smarter. Don't be one of those

185
00:08:40,202 --> 00:08:42,610
people who says, I'm not going to touch it. I'm very

186
00:08:42,642 --> 00:08:45,430
comfortable. Because at some point uh,

187
00:08:45,430 --> 00:08:48,370
you're going to be pushed in that direction, whether it's by your

188
00:08:48,442 --> 00:08:50,854
company, by your peers, whether by

189
00:08:51,634 --> 00:08:54,418
the sheer fact that your uh, job might

190
00:08:54,466 --> 00:08:57,374
get um, obsolete. You will get.

191
00:08:57,754 --> 00:09:00,546
>> Anthony Weaver: Yeah, because what I'm thinking about is when it

192
00:09:00,570 --> 00:09:03,402
comes to job replacement and

193
00:09:03,498 --> 00:09:06,299
promote promotability, one of the things that I

194
00:09:06,347 --> 00:09:09,067
always promote or say to

195
00:09:09,091 --> 00:09:11,907
myself or anybody who's looking to get into their

196
00:09:12,051 --> 00:09:14,763
field for the first time is to

197
00:09:14,899 --> 00:09:17,891
understand the business as a whole. Write the processes and

198
00:09:17,923 --> 00:09:20,619
procedures in place and streamline it so that you

199
00:09:20,667 --> 00:09:23,635
can promote or uh, replace yourself

200
00:09:23,795 --> 00:09:26,659
in that sense. Uh, so say if somebody is

201
00:09:26,707 --> 00:09:29,035
looking at their job and they're like, hey, this is really

202
00:09:29,075 --> 00:09:31,911
monotonous. Is it a way that,

203
00:09:32,043 --> 00:09:34,395
like what tool or program

204
00:09:34,895 --> 00:09:37,615
should they look into learning if they're looking

205
00:09:37,655 --> 00:09:40,551
to replace some of those repetitive things? So like, the company

206
00:09:40,583 --> 00:09:43,071
might not be on go about this

207
00:09:43,223 --> 00:09:46,215
AI thing but say if they can just write a

208
00:09:46,255 --> 00:09:48,847
small Python script or

209
00:09:48,911 --> 00:09:51,759
small. I mean, I know c, I'm aging myself

210
00:09:51,807 --> 00:09:54,727
here, But a small JavaScript to

211
00:09:54,751 --> 00:09:57,620
kind of do what they do day to day. Uh,

212
00:09:57,620 --> 00:09:59,075
what are your thoughts on that?

213
00:09:59,520 --> 00:10:02,431
>> Sumedha Rai: Um, I think if you're a person who can actually

214
00:10:02,463 --> 00:10:05,079
take the initiative to do it, I would support it

215
00:10:05,167 --> 00:10:08,135
completely. Um, um. A lot of solutions

216
00:10:08,255 --> 00:10:11,119
nowadays are built on Python and there

217
00:10:11,127 --> 00:10:13,915
are so many different environments that

218
00:10:14,110 --> 00:10:16,795
um, will enable you to do it free of cost.

219
00:10:17,535 --> 00:10:20,391
Google, um, Colab, for instance, is going to give you GPUs

220
00:10:20,423 --> 00:10:23,055
for free. So you can even try out computer vision

221
00:10:23,095 --> 00:10:26,031
models or you can try out language models which are very

222
00:10:26,063 --> 00:10:28,839
heavy in nature and sometimes cannot be done on someone's machine.

223
00:10:28,967 --> 00:10:31,917
So if you're that kind of a person, I, um, would say

224
00:10:31,941 --> 00:10:34,530
the first thing that you need to do is uh,

225
00:10:34,530 --> 00:10:37,253
to understand what your problem statement is.

226
00:10:37,429 --> 00:10:40,277
So um, you break it down into parts

227
00:10:40,421 --> 00:10:43,157
and you see which part needs to Be automated

228
00:10:43,341 --> 00:10:46,037
and then you do a little more research about it. If

229
00:10:46,061 --> 00:10:48,300
it's ah, a general question, um,

230
00:10:48,845 --> 00:10:51,749
like I don't know, classifying it into certain categories

231
00:10:51,797 --> 00:10:53,385
as I was talking about before,

232
00:10:54,885 --> 00:10:57,757
the chances are the likelihood is that there is already

233
00:10:57,821 --> 00:11:00,781
an open source library that exists to do that

234
00:11:00,813 --> 00:11:03,733
work for you. If you're someone who already knows a little bit

235
00:11:03,749 --> 00:11:06,453
of Python, that's actually a language that's used a lot

236
00:11:06,549 --> 00:11:09,077
in um, AI machine learning, some flavor of

237
00:11:09,101 --> 00:11:11,797
Python. It could be Piparark, which is used

238
00:11:11,861 --> 00:11:14,757
for uh, making sure that you can handle a lot of

239
00:11:14,781 --> 00:11:17,749
data. It could be Pytorch, which is used for

240
00:11:17,797 --> 00:11:20,797
more deep learning algorithms. It could be something

241
00:11:20,821 --> 00:11:23,685
else. But people have put in a lot of hard

242
00:11:23,725 --> 00:11:26,573
work already to create open source libraries for you.

243
00:11:26,709 --> 00:11:29,549
And if your solution, um, the problem

244
00:11:29,597 --> 00:11:32,587
that you're looking to automate um, is, is something

245
00:11:32,611 --> 00:11:35,507
that is, that has been done before, there is a chance that

246
00:11:35,531 --> 00:11:38,419
somebody's already created a library for it. So then your job becomes

247
00:11:38,467 --> 00:11:41,411
even easier. You import that library, you

248
00:11:41,443 --> 00:11:44,419
use it, you test it out and you see if things are getting

249
00:11:44,467 --> 00:11:47,255
better for you from then on. If this is something

250
00:11:47,320 --> 00:11:49,720
uh, that you fancy, uh,

251
00:11:50,715 --> 00:11:53,315
you're going to organically get pulled into it. Trust

252
00:11:53,355 --> 00:11:55,971
me, if you're a person who enjoys

253
00:11:56,083 --> 00:11:58,965
seeing the automation work, there is no

254
00:11:59,005 --> 00:12:00,861
chance that you're not going to want to do more.

255
00:12:00,973 --> 00:12:03,973
>> Anthony Weaver: Yes, I love it, uh, because you

256
00:12:03,989 --> 00:12:06,909
have a talk coming up, um, in New York

257
00:12:06,957 --> 00:12:09,933
City say in June, uh, talking about

258
00:12:09,989 --> 00:12:12,877
the navigating the bias in AI, just

259
00:12:12,901 --> 00:12:15,653
kind of promoting the ethical decisions making

260
00:12:15,709 --> 00:12:18,565
in fintech. Can you just kind of give us a quick

261
00:12:18,605 --> 00:12:21,373
synopsis for those people who have

262
00:12:21,429 --> 00:12:24,137
no clue about the bias or

263
00:12:24,201 --> 00:12:26,257
the non bias in AI?

264
00:12:26,441 --> 00:12:29,073
>> Sumedha Rai: Right. Um, I think that talk was last

265
00:12:29,129 --> 00:12:32,081
year, uh, but it was a very interesting talk.

266
00:12:32,233 --> 00:12:35,129
I tried to capture some of the ideas. No, you're good.

267
00:12:35,177 --> 00:12:37,441
I'm so glad that you actually looked it up. I'm

268
00:12:37,473 --> 00:12:40,225
thrilled. Ah, there's another one coming up in April

269
00:12:40,265 --> 00:12:41,449
that's about fraud detection.

270
00:12:41,577 --> 00:12:44,457
But let me just talk about the bias in AI and why I think

271
00:12:44,481 --> 00:12:47,405
it's a topic that we should be talking about. So

272
00:12:48,150 --> 00:12:51,090
um, most of the training data, um,

273
00:12:51,090 --> 00:12:53,759
as we call it, and I'm just going to explain what training data is real

274
00:12:53,807 --> 00:12:56,375
quick. And AI model needs to learn

275
00:12:56,495 --> 00:12:59,415
from something, um, it needs

276
00:12:59,535 --> 00:13:02,495
to predict something for you and for

277
00:13:02,535 --> 00:13:05,391
that it needs to understand what it's trying to

278
00:13:05,423 --> 00:13:08,407
predict. So as a simple example, if you want the

279
00:13:08,431 --> 00:13:11,119
AI model to understand numbers,

280
00:13:11,247 --> 00:13:14,239
handwritten numbers, you need to show it a lot of Pictures

281
00:13:14,287 --> 00:13:16,875
of handwritten digits and then it's going to understand

282
00:13:16,915 --> 00:13:19,731
writing styles, it's going to understand a two

283
00:13:19,763 --> 00:13:22,587
is with a curve, but a one is a straight line, stuff like

284
00:13:22,611 --> 00:13:25,227
this. So it needs a lot of training data

285
00:13:25,291 --> 00:13:27,975
to understand what it's supposed to predict.

286
00:13:28,315 --> 00:13:31,067
Now a lot of the training data that we

287
00:13:31,091 --> 00:13:33,947
had in the past, um, which

288
00:13:33,971 --> 00:13:36,851
has already been generated was based on

289
00:13:36,923 --> 00:13:39,855
decisions that could have been biased

290
00:13:40,235 --> 00:13:42,090
because you know, um,

291
00:13:43,095 --> 00:13:46,047
it could be because of human cognitive errors, it could be

292
00:13:46,111 --> 00:13:48,735
because of, you know, the society that existed at that

293
00:13:48,775 --> 00:13:51,703
time. But we're using any and all of data

294
00:13:51,759 --> 00:13:54,415
that we can find right now to train models that they

295
00:13:54,495 --> 00:13:57,175
understand a lot. So ah, the AI

296
00:13:57,215 --> 00:14:00,031
model becomes better if it sees a lot more

297
00:14:00,063 --> 00:14:02,535
examples. It's like a child

298
00:14:02,655 --> 00:14:05,195
saying, if you show me an orange five times

299
00:14:06,055 --> 00:14:08,823
I um, will understand that it's an orange. But if you show me an

300
00:14:08,919 --> 00:14:11,903
orange a hundred times, I will definitely understand that it's

301
00:14:11,919 --> 00:14:14,893
an orange. So uh, we try to give the

302
00:14:14,909 --> 00:14:17,225
model as much data as possible. Unfortunately,

303
00:14:17,790 --> 00:14:20,789
uh, we might be feeding it biased data. And as

304
00:14:20,837 --> 00:14:23,733
part of our systems we're not always

305
00:14:23,829 --> 00:14:26,629
very careful that the results

306
00:14:26,677 --> 00:14:28,893
that we're getting are not

307
00:14:28,949 --> 00:14:31,685
biased. Um, if we use such

308
00:14:31,725 --> 00:14:34,045
models to make automated decision

309
00:14:34,125 --> 00:14:37,109
making, it might give out biased results and

310
00:14:37,117 --> 00:14:39,305
it might affect actual human lives.

311
00:14:39,735 --> 00:14:41,959
So for me the question

312
00:14:42,087 --> 00:14:44,919
of understanding the results of a model

313
00:14:44,967 --> 00:14:46,530
from an ethical and um,

314
00:14:47,295 --> 00:14:49,951
unbiased perspective is very important.

315
00:14:50,143 --> 00:14:53,110
But a lot of companies that are integrating AI, ah,

316
00:14:53,215 --> 00:14:55,675
solutions are not making it as

317
00:14:56,095 --> 00:14:59,087
an important, like an absolute necessary

318
00:14:59,151 --> 00:15:01,995
part of their pipelines. Right now they,

319
00:15:02,295 --> 00:15:05,191
you fine tune the model or train the model, they

320
00:15:05,223 --> 00:15:07,967
see how it performs on a uh, on a set of

321
00:15:07,991 --> 00:15:10,831
data, testing data, um, and they

322
00:15:10,863 --> 00:15:13,599
deploy it. But there needs to be this other

323
00:15:13,647 --> 00:15:16,551
chunk that needs to go before the results

324
00:15:16,583 --> 00:15:19,447
are deployed which somehow overlays the

325
00:15:19,471 --> 00:15:22,175
demographic data on top of your results to

326
00:15:22,255 --> 00:15:25,223
understand is it somehow producing biase results?

327
00:15:25,399 --> 00:15:28,391
If it's producing biase results, what can we do?

328
00:15:28,503 --> 00:15:30,951
We recalibrate it for a while and then refeed

329
00:15:30,983 --> 00:15:33,943
it. Do we not use the solution as an automated

330
00:15:33,999 --> 00:15:36,555
solution? We could have a human in the loop.

331
00:15:36,855 --> 00:15:39,711
And is this something that is going to really critically affect

332
00:15:39,783 --> 00:15:42,495
the lives of people, for instance in credit decisions

333
00:15:42,655 --> 00:15:44,783
or in some medical decisions?

334
00:15:44,959 --> 00:15:47,475
So um, for certain things

335
00:15:47,935 --> 00:15:50,559
this is almost an irrepraceable part of your

336
00:15:50,607 --> 00:15:53,015
pipeline and we need to have more

337
00:15:53,055 --> 00:15:54,515
conversations about it.

338
00:15:54,855 --> 00:15:57,631
>> Anthony Weaver: Yeah, because I'm thinking about the models back

339
00:15:57,663 --> 00:16:00,423
in like reading some of the books of

340
00:16:00,479 --> 00:16:02,999
how credit was actually dispersed

341
00:16:03,047 --> 00:16:05,719
to uh, different neighborhoods or people

342
00:16:05,767 --> 00:16:08,423
of color during that timef frme

343
00:16:08,559 --> 00:16:11,295
was like, they call it redlining and based

344
00:16:11,335 --> 00:16:14,271
on, you know, the credit score or they have

345
00:16:14,303 --> 00:16:17,047
the valability of their business

346
00:16:17,111 --> 00:16:20,087
actually being successful or the banks are willing to

347
00:16:20,111 --> 00:16:22,967
take that type of risk. So is this kind

348
00:16:22,991 --> 00:16:25,935
of where we getting back into, hey, they don't

349
00:16:25,975 --> 00:16:28,280
know enough information about us, uh,

350
00:16:28,735 --> 00:16:31,543
meaning us as people of color because we

351
00:16:31,559 --> 00:16:34,511
were always left, it seems like we always left behind and it was always

352
00:16:34,543 --> 00:16:37,039
trying to play this catch up game. How can

353
00:16:37,087 --> 00:16:39,831
we now be part

354
00:16:39,903 --> 00:16:42,559
of that um, that

355
00:16:42,607 --> 00:16:45,551
conversation or be part of that data set so that we

356
00:16:45,583 --> 00:16:47,515
aren't left out of it?

357
00:16:47,815 --> 00:16:50,155
>> Sumedha Rai: Right. I think this is a very, very important

358
00:16:50,250 --> 00:16:53,191
uh, topic that you're raising right now. Very important question

359
00:16:53,263 --> 00:16:55,647
to address. And you're right. There were certain

360
00:16:55,751 --> 00:16:58,583
subsections of the population that were not even given the credit. So there

361
00:16:58,599 --> 00:17:01,239
is no data about them or you know, the

362
00:17:01,287 --> 00:17:03,989
decisions were all adverse, which means there is

363
00:17:04,037 --> 00:17:06,770
negative data about them. So uh,

364
00:17:06,770 --> 00:17:09,717
one of the things that we start with is to give it

365
00:17:09,741 --> 00:17:12,105
more representative data, which means

366
00:17:12,565 --> 00:17:15,093
if you pick out a certain subsection of the

367
00:17:15,109 --> 00:17:18,101
population, you give it both positive and

368
00:17:18,133 --> 00:17:20,805
negative cases and it needs to be more

369
00:17:20,845 --> 00:17:23,821
balanced. So for instance, in your data set, if you

370
00:17:23,853 --> 00:17:26,569
have say five different

371
00:17:26,569 --> 00:17:29,333
um, geographies, you give

372
00:17:29,349 --> 00:17:32,109
it enough positive and negative data in each of those

373
00:17:32,157 --> 00:17:34,890
geographies, uh, four for

374
00:17:34,890 --> 00:17:37,787
um, the model to understand that this should not be

375
00:17:37,811 --> 00:17:40,667
the primary factor that I can make a decision

376
00:17:40,731 --> 00:17:43,531
on. And this could be something to do with

377
00:17:43,603 --> 00:17:46,230
age, it could be something to do with um,

378
00:17:46,310 --> 00:17:48,930
um, racial bias, um,

379
00:17:49,363 --> 00:17:52,323
that we've been seeing. It could be something to do with geography,

380
00:17:52,459 --> 00:17:55,235
it could be something to do with um, the credit

381
00:17:55,275 --> 00:17:58,243
score. So unfortunately, credit scores have

382
00:17:58,259 --> 00:18:01,019
been linked with um, geographies and races

383
00:18:01,067 --> 00:18:04,051
for some time. And this is the link that we're trying

384
00:18:04,083 --> 00:18:06,875
to del link now. Now, assuming you still see the

385
00:18:06,915 --> 00:18:09,747
results, I understand in certain cases you actually do not have

386
00:18:09,771 --> 00:18:12,403
data, so you need to train it. Assuming you see some

387
00:18:12,459 --> 00:18:14,815
results, once you see the results,

388
00:18:14,870 --> 00:18:17,699
um, AI models are inherently predictive in nature,

389
00:18:17,747 --> 00:18:20,739
which means they're giving you probab scores. Can you do

390
00:18:20,787 --> 00:18:23,563
something to the probability scores to bring these results

391
00:18:23,619 --> 00:18:26,443
on the same um, level so you

392
00:18:26,459 --> 00:18:28,855
can recalibrate the scores to say

393
00:18:29,195 --> 00:18:31,255
this negative is equal to,

394
00:18:31,915 --> 00:18:34,667
to the same negative in terms of like the absolute

395
00:18:34,731 --> 00:18:37,517
number. In terms of the probability number that I just gave

396
00:18:37,541 --> 00:18:39,785
you, these two negatives look equal to me.

397
00:18:40,190 --> 00:18:43,037
Um, and then you use this data again to

398
00:18:43,061 --> 00:18:45,973
train your model so then it understands that I'm looking for

399
00:18:46,029 --> 00:18:48,709
patterns right now. I'm not going to just concentrate on these,

400
00:18:48,757 --> 00:18:51,133
like probability measures. Another thing that's really

401
00:18:51,189 --> 00:18:54,065
important is to quantify the bias.

402
00:18:54,645 --> 00:18:57,453
So, um, let's imagine

403
00:18:57,549 --> 00:19:00,053
that, uh, you're training a model and

404
00:19:00,229 --> 00:19:03,117
there's a person who's going toa evaluate it. The person needs to evaluate

405
00:19:03,141 --> 00:19:06,041
it on a certain metric. Now what I'm

406
00:19:06,073 --> 00:19:08,537
saying is that when you overlay the

407
00:19:08,561 --> 00:19:11,369
demographic data on top of your results, the false

408
00:19:11,417 --> 00:19:13,889
positive rate, like a person

409
00:19:14,017 --> 00:19:16,401
being denied credit, for instance, but

410
00:19:16,433 --> 00:19:19,361
wrongly, should be equal in both the

411
00:19:19,393 --> 00:19:22,257
segments. Like if the model is denying something

412
00:19:22,281 --> 00:19:25,217
incorrectly, it should deny it in both the segments equally.

413
00:19:25,361 --> 00:19:28,209
And now if it's not, let's work towards getting this

414
00:19:28,257 --> 00:19:30,841
number correct. So there are things that you can do and

415
00:19:30,873 --> 00:19:33,857
hopefully, you know, we're going towards a society

416
00:19:33,921 --> 00:19:36,863
that's more equal. So the data that's going to be generated now, now

417
00:19:36,879 --> 00:19:39,623
it's going toa be more equitable. So when we use this to retrain the

418
00:19:39,639 --> 00:19:42,415
models, things are going toa look better. And again, for this

419
00:19:42,455 --> 00:19:44,943
thing, we can t have the solution be automated

420
00:19:45,039 --> 00:19:47,743
completely. We need not just one,

421
00:19:47,839 --> 00:19:50,391
but several humans in the loop. And I say several

422
00:19:50,463 --> 00:19:53,399
humans because we as people are also biased. We

423
00:19:53,407 --> 00:19:56,095
have human cognitive biases. So you cannot just

424
00:19:56,135 --> 00:19:58,783
have, um, one person's opinion. Sometimes

425
00:19:58,839 --> 00:20:01,327
for crucial things, you need to have five people's

426
00:20:01,351 --> 00:20:04,209
opinions. It's like taking the average of a model.

427
00:20:04,327 --> 00:20:07,109
So you get the results from the model, you run it through like five

428
00:20:07,157 --> 00:20:09,981
people's decision making and then you generate

429
00:20:10,013 --> 00:20:12,733
some results which are hopefully more

430
00:20:12,869 --> 00:20:14,597
equitable, more accurate.

431
00:20:14,741 --> 00:20:16,680
>> Anthony Weaver: Yeah, okay.

432
00:20:16,680 --> 00:20:19,225
Um, because I'm just thinking of,

433
00:20:20,045 --> 00:20:22,701
I'm glad that there are people like you who are actually

434
00:20:22,853 --> 00:20:25,741
stepping up and saying something in this world

435
00:20:25,853 --> 00:20:28,461
of tech. And I'm sure you had your

436
00:20:28,613 --> 00:20:31,501
set of boundaries to kind of push through,

437
00:20:31,580 --> 00:20:34,485
uh, what were the hardest challenges of actually getting

438
00:20:34,565 --> 00:20:35,945
into this industry?

439
00:20:36,960 --> 00:20:39,731
>> Sumedha Rai: Um, I think one of the things that

440
00:20:39,763 --> 00:20:42,571
is hard for people historically is

441
00:20:42,683 --> 00:20:45,563
to not come from a tech background that can sometimes be

442
00:20:45,579 --> 00:20:48,419
a blocker. Um, I actually

443
00:20:48,507 --> 00:20:51,415
started, um, in economics and then

444
00:20:51,715 --> 00:20:54,603
I did a master's in finance and I spent some time in investment

445
00:20:54,659 --> 00:20:57,635
banking and private equity. It had nothing to do with

446
00:20:57,675 --> 00:21:00,667
tech. And, uh, when I entered the

447
00:21:00,691 --> 00:21:03,469
field of tech, there were already people who were

448
00:21:03,557 --> 00:21:06,341
working, uh, in tech for the last like five

449
00:21:06,373 --> 00:21:09,157
years or something. So for them, programming was

450
00:21:09,221 --> 00:21:12,053
something that was second nature. And for me it

451
00:21:12,069 --> 00:21:14,725
was something new that I had to spend some time on. I taken,

452
00:21:14,845 --> 00:21:17,785
I'd taken lessons in Java and C

453
00:21:18,085 --> 00:21:20,877
during high school. Yeah, and

454
00:21:20,981 --> 00:21:23,621
you know, like, it builds, uh, your logical

455
00:21:23,693 --> 00:21:26,493
thinking, it builds Your reasoning, thinking, you know how if statements

456
00:21:26,549 --> 00:21:29,483
work but then um, you're now

457
00:21:29,579 --> 00:21:32,579
programming in a new language uh, which has a different syntax

458
00:21:32,627 --> 00:21:35,603
and it takes some time to get used to it. Whereas the people who

459
00:21:35,619 --> 00:21:38,419
have already been working in this. But then every

460
00:21:38,467 --> 00:21:41,347
person can have their strengths when they're entering a

461
00:21:41,371 --> 00:21:44,339
new field. So for instance for me uh, domain

462
00:21:44,387 --> 00:21:47,347
knowledge was a point of strength and statistics was a point

463
00:21:47,371 --> 00:21:50,227
of strength. So um, I was entering the field of

464
00:21:50,251 --> 00:21:53,107
data science and AI where the decision making also

465
00:21:53,131 --> 00:21:55,971
relies on uh, the statistical assumptions that we

466
00:21:56,003 --> 00:21:58,899
make. Some math driven uh, things that we have to take

467
00:21:58,947 --> 00:22:01,577
into account before we reach the programming stage.

468
00:22:01,721 --> 00:22:04,617
So um, I found my strengths over there and then I

469
00:22:04,641 --> 00:22:07,425
built up on this other side of like the tech knowledge

470
00:22:07,465 --> 00:22:10,041
and infrastructure knowledge and programming knowledge.

471
00:22:10,233 --> 00:22:13,105
So if you're a person who wants to

472
00:22:13,185 --> 00:22:15,905
get into tech but is not exactly from a tech

473
00:22:15,945 --> 00:22:18,761
background, it is not exactly a

474
00:22:18,793 --> 00:22:21,761
blocker. In today's world at least I see so

475
00:22:21,793 --> 00:22:24,449
many people entering it, taking some time to upsill

476
00:22:24,497 --> 00:22:27,361
themselves and we have reached a point where

477
00:22:27,393 --> 00:22:30,137
we're trying to democratize AI as much as possible. There are

478
00:22:30,201 --> 00:22:33,151
so many open source options available.

479
00:22:33,303 --> 00:22:36,183
When I started my journey, um, I was very

480
00:22:36,279 --> 00:22:38,759
clear on the fact that I have to do a formal

481
00:22:38,807 --> 00:22:41,647
master's program and I did it

482
00:22:41,711 --> 00:22:44,511
and it was very useful for me. But then I also saw

483
00:22:44,543 --> 00:22:47,383
people just going to boot camps. You

484
00:22:47,399 --> 00:22:50,263
know it's harder like compress that

485
00:22:50,439 --> 00:22:53,151
amount of knowledge in just like 16

486
00:22:53,223 --> 00:22:56,023
weeks or something. But people were using this option

487
00:22:56,159 --> 00:22:59,063
and then there is uh, like this whole segment of

488
00:22:59,079 --> 00:23:01,625
people who just go online and they pick up courses

489
00:23:01,815 --> 00:23:04,661
and over time they are doing very well

490
00:23:04,693 --> 00:23:07,629
as well. Like they, some people choose to specialize

491
00:23:07,677 --> 00:23:10,557
in a certain subsectionion of tech or

492
00:23:10,581 --> 00:23:13,120
AI and they get very, very good at it. Like um,

493
00:23:13,549 --> 00:23:16,213
visualizations. It's a very very

494
00:23:16,269 --> 00:23:19,269
important part of conveying your point. And

495
00:23:19,397 --> 00:23:22,381
I still think there is a lot of work that's needed from my

496
00:23:22,453 --> 00:23:25,365
side to find the perfect graph

497
00:23:25,445 --> 00:23:28,160
or the perfect um, um

498
00:23:28,205 --> 00:23:30,979
PowerPoint slide. That makes it,

499
00:23:31,037 --> 00:23:33,695
it so much better for a person who's from a non tech

500
00:23:33,735 --> 00:23:36,591
background to understand my point and people have mastered the

501
00:23:36,623 --> 00:23:39,407
skill and they are absolutely required. Like

502
00:23:39,431 --> 00:23:41,871
I'd love to get some tips from

503
00:23:41,903 --> 00:23:44,719
them. But uh, I think

504
00:23:44,767 --> 00:23:47,207
again my point is if you're

505
00:23:47,391 --> 00:23:50,343
interested in the field of AI but you don't come from a

506
00:23:50,359 --> 00:23:53,167
tech background, don't let that be a blocker.

507
00:23:53,311 --> 00:23:54,063
>> Anthony Weaver: Got you.

508
00:23:54,199 --> 00:23:56,951
So giveniving your experience growing up um, and studying

509
00:23:56,983 --> 00:23:59,743
in India, the UK and also the

510
00:23:59,799 --> 00:24:02,559
US like how does could these diverse cultural

511
00:24:02,647 --> 00:24:05,559
and educational settings like really shape your approach

512
00:24:05,607 --> 00:24:06,875
to problem solving?

513
00:24:07,300 --> 00:24:10,071
>> Sumedha Rai: Uh, I think collaboration is one of the

514
00:24:10,103 --> 00:24:12,879
biggest strengths that I have gained uh because

515
00:24:12,927 --> 00:24:15,519
I worked in teams that were

516
00:24:15,607 --> 00:24:18,295
culturally different, that had a different set of

517
00:24:18,335 --> 00:24:20,795
opinions and how they put it in front of you

518
00:24:21,175 --> 00:24:23,800
and they were also you know

519
00:24:23,800 --> 00:24:26,151
um, on different skill levels.

520
00:24:26,303 --> 00:24:29,045
So um, what I saw in

521
00:24:29,085 --> 00:24:32,053
one country was not necessarily how things

522
00:24:32,109 --> 00:24:35,077
were uh managed in another country both in

523
00:24:35,101 --> 00:24:37,910
terms of like communication, in terms of their um

524
00:24:38,309 --> 00:24:40,985
approach to problems. But what really helped me

525
00:24:41,365 --> 00:24:44,101
understand is that you can gain something out

526
00:24:44,133 --> 00:24:46,705
of every opinion, every perspective.

527
00:24:46,830 --> 00:24:49,581
Um um one of the things that we do in

528
00:24:49,613 --> 00:24:52,285
statistics sometimes or you know in modeling

529
00:24:52,325 --> 00:24:54,785
sometimes is pooling averages.

530
00:24:55,365 --> 00:24:58,013
And um, if the model is not performing

531
00:24:58,069 --> 00:25:01,061
great we get a lot of smaller

532
00:25:01,093 --> 00:25:03,813
models together and then get the average out of that.

533
00:25:03,949 --> 00:25:06,829
And to me uh, my cross country experience,

534
00:25:06,877 --> 00:25:09,605
my cross cultural experience was somewhat of a pooling

535
00:25:09,645 --> 00:25:12,341
average because I try to

536
00:25:12,373 --> 00:25:15,301
listen to everyone's opinion with the

537
00:25:15,333 --> 00:25:18,021
mindset that I'm going to pick out this one

538
00:25:18,093 --> 00:25:21,045
thing that's going to really resonate with me or I'm going to

539
00:25:21,085 --> 00:25:23,629
have this h. I didn't think of it like this

540
00:25:23,677 --> 00:25:26,615
moment and then I'm going to work on it and I'm going to understand

541
00:25:26,685 --> 00:25:29,379
and uh, how to go about it from a different perspective.

542
00:25:29,507 --> 00:25:32,203
And I actually I do gain a lot from conversations.

543
00:25:32,339 --> 00:25:35,043
It could be a person sitting in the UK and South Africa

544
00:25:35,099 --> 00:25:37,771
and India or my co worker in the US

545
00:25:37,963 --> 00:25:40,575
and talking about the same topic

546
00:25:40,875 --> 00:25:43,627
like how do you approach fraud detection in AI

547
00:25:43,771 --> 00:25:46,627
gives me so many different perspectives that my resulting model

548
00:25:46,651 --> 00:25:48,015
is actually very strong.

549
00:25:50,035 --> 00:25:52,939
>> Anthony Weaver: Nice. Uh because you

550
00:25:52,987 --> 00:25:55,575
also mentioned like um I'm thinking about from

551
00:25:56,015 --> 00:25:58,695
other things that you've talked about before but

552
00:25:58,775 --> 00:26:01,303
considering your work with like natural language

553
00:26:01,479 --> 00:26:04,311
processing or NLP's that you've talked

554
00:26:04,343 --> 00:26:07,195
about before and understanding like the user sentiments,

555
00:26:07,880 --> 00:26:10,703
um are they like the

556
00:26:10,719 --> 00:26:13,447
emotional challenges that the Sadwich generation have which

557
00:26:13,471 --> 00:26:15,975
is um the design of

558
00:26:16,015 --> 00:26:18,687
AI doesn't offer really practical

559
00:26:18,791 --> 00:26:21,649
assistance for them. So like how there

560
00:26:21,657 --> 00:26:24,609
are ways that they can actually utilize it

561
00:26:24,697 --> 00:26:27,505
in their day to day. So I would say they got

562
00:26:27,545 --> 00:26:30,257
kids and then they got parents to deal with. Is

563
00:26:30,281 --> 00:26:33,233
there any AI tools or suggestions

564
00:26:33,289 --> 00:26:34,445
that you have for them?

565
00:26:35,220 --> 00:26:38,005
>> Sumedha Rai: Um, I think since we're thinking of

566
00:26:38,320 --> 00:26:41,097
um a person who wants to us

567
00:26:41,121 --> 00:26:43,985
AI in their day to day lives I could come up with so many

568
00:26:44,065 --> 00:26:45,480
suggestions because um,

569
00:26:46,745 --> 00:26:49,741
we've really handed these uh, machine

570
00:26:49,773 --> 00:26:52,160
learning solutions to people's uh

571
00:26:52,160 --> 00:26:55,101
hands literally like fed them saying here's

572
00:26:55,133 --> 00:26:57,445
a $20 subscription much like

573
00:26:57,485 --> 00:27:00,440
net go interact with chat GB. But

574
00:27:00,440 --> 00:27:03,025
uh, since I'm on a finance podcast,

575
00:27:03,485 --> 00:27:05,933
let, um, me probably talk about some solutions

576
00:27:06,029 --> 00:27:08,805
that people should absolutely start

577
00:27:08,885 --> 00:27:11,465
using if they're not using it already.

578
00:27:11,900 --> 00:27:14,861
Um, I think money is an important topic in

579
00:27:14,893 --> 00:27:17,390
every family, for every individual. Um,

580
00:27:17,765 --> 00:27:20,725
so sometimes we do not know

581
00:27:20,765 --> 00:27:23,585
how to get started with money conversations.

582
00:27:24,085 --> 00:27:26,893
We don't know what to do with our money. And one of the worst

583
00:27:26,949 --> 00:27:29,773
mistakes that you could do with your money is to just let it

584
00:27:29,829 --> 00:27:32,613
sit in a bank account without doing anything

585
00:27:32,669 --> 00:27:35,645
with it. Um, so you absolutely need to start

586
00:27:35,685 --> 00:27:38,680
your journey to do something with it. Um,

587
00:27:38,680 --> 00:27:41,541
it could be that you want to save a lot more and then invested

588
00:27:41,613 --> 00:27:44,101
later. But you're, you know, to begin with, you're not

589
00:27:44,133 --> 00:27:46,995
saving at all and you're finding trouble

590
00:27:47,155 --> 00:27:49,963
understanding what am I doing with my money. Start with a

591
00:27:49,979 --> 00:27:52,899
budgeting app. Um, I've seen people who create

592
00:27:52,947 --> 00:27:55,771
these beautiful Excel sheets. Everything is

593
00:27:55,803 --> 00:27:58,803
color coded, everything is linked. But as you mentioned,

594
00:27:58,939 --> 00:28:01,835
a person with two kids and a full time job

595
00:28:01,875 --> 00:28:04,619
may not have the time to do it. So there are these

596
00:28:04,707 --> 00:28:07,615
apps that make budgeting really easy for you.

597
00:28:07,720 --> 00:28:10,651
Um, they look back at the transactions that you've

598
00:28:10,683 --> 00:28:13,243
made in the last, I don't know, two years, five years.

599
00:28:13,419 --> 00:28:16,027
And um, given a set of goals that you

600
00:28:16,051 --> 00:28:18,529
have, the future, they're gonna start nudging you in the right

601
00:28:18,577 --> 00:28:21,345
direction, saying, you know, okay, in the last

602
00:28:21,505 --> 00:28:24,440
24 months it looks like, um,

603
00:28:24,440 --> 00:28:27,393
you did not spend so much on travel, but suddenly I see a lot

604
00:28:27,409 --> 00:28:30,265
of Ubers right, happening. Are you

605
00:28:30,305 --> 00:28:33,273
just doing this because you're, you're not,

606
00:28:33,409 --> 00:28:36,401
you really needed it, or is this because of something that you

607
00:28:36,433 --> 00:28:39,401
could have avoided and you save instead? Now once

608
00:28:39,433 --> 00:28:41,600
you've started saving, you, um,

609
00:28:42,655 --> 00:28:45,607
have to start investing in something, uh, in somewhere because

610
00:28:45,631 --> 00:28:48,391
your money needs to make more money,

611
00:28:48,463 --> 00:28:51,407
so to say. So, um, what are you doing

612
00:28:51,431 --> 00:28:54,303
in your investment journey? Um, you, are

613
00:28:54,319 --> 00:28:57,191
you using, are you comfortable enough to say I already know where I

614
00:28:57,223 --> 00:29:00,087
want to put my stocks? In which case maybe a robo advisor could

615
00:29:00,111 --> 00:29:03,111
help. Like, you know, it a fraction of the cost and you could

616
00:29:03,143 --> 00:29:05,815
just do with some extra financial advice that a

617
00:29:05,895 --> 00:29:08,607
financial expert could give you. But maybe you couldn't afford the

618
00:29:08,671 --> 00:29:11,639
financial expt, so you could look at a robo advisor if you don't want

619
00:29:11,647 --> 00:29:14,353
to do that. There are investments that are more

620
00:29:14,409 --> 00:29:16,937
passive in nature. So, um, you

621
00:29:16,961 --> 00:29:19,921
could, uh, you pay a subscription cost or

622
00:29:19,993 --> 00:29:22,921
you could pay um, a dollar cost for every

623
00:29:22,953 --> 00:29:25,360
transaction that's made. But you start

624
00:29:25,360 --> 00:29:28,249
uh, tying up your investments to an index

625
00:29:28,297 --> 00:29:31,185
and you're passively investing it. And um, there are certain

626
00:29:31,225 --> 00:29:34,045
AI solutions that also like, tailor to your needs. Like,

627
00:29:34,100 --> 00:29:37,033
um, um, this is the ROI that

628
00:29:37,049 --> 00:29:39,921
I want to achieve in the next two years. These are my long

629
00:29:39,953 --> 00:29:42,873
term goals. I'm saving to buy a house. This

630
00:29:42,889 --> 00:29:45,420
is how much the house can cost me. Um,

631
00:29:45,615 --> 00:29:48,471
these are my risk preferences. So you give it a whole

632
00:29:48,503 --> 00:29:51,047
bunch of parameters and then it starts

633
00:29:51,111 --> 00:29:53,847
rebalancing your portfolio, given the market conditions,

634
00:29:53,991 --> 00:29:56,991
and starts putting the money in there. So those are also good solutions

635
00:29:57,063 --> 00:29:59,719
to have. Um, I've seen

636
00:29:59,767 --> 00:30:02,735
people who are very good at investing. They're like,

637
00:30:02,815 --> 00:30:05,775
look at the market every single day. I'm going through my own articles

638
00:30:05,815 --> 00:30:08,675
and I'm doing my own research for such people.

639
00:30:08,800 --> 00:30:11,743
Um, um, getting a

640
00:30:11,759 --> 00:30:14,757
summary of the financial news every single day could be helpful as

641
00:30:14,781 --> 00:30:17,389
well. Uh, I think the Bloomberg

642
00:30:17,437 --> 00:30:20,413
terminals have it where when you open the terminal, the first thing that

643
00:30:20,429 --> 00:30:22,997
you see on the screen, or like it's a part of the terminal somewhere,

644
00:30:23,021 --> 00:30:25,821
is five short bullet points about how the financial

645
00:30:25,853 --> 00:30:28,149
markets are looking. And for me, that's going to be really useful

646
00:30:28,197 --> 00:30:31,069
information because I might say, okay, this is

647
00:30:31,117 --> 00:30:33,757
something that I don't even invest in, so maybe it's not

648
00:30:33,781 --> 00:30:36,493
useful. But then I might find something to do with the Magnificent

649
00:30:36,549 --> 00:30:39,509
Seven and I will click into it and I'll start reading about it.

650
00:30:39,597 --> 00:30:42,195
So there are solutions, uh, for a

651
00:30:42,235 --> 00:30:44,875
person to look into their money

652
00:30:44,955 --> 00:30:47,907
more closely. And, um, I

653
00:30:47,931 --> 00:30:50,339
think people should absolutely start doing it

654
00:30:50,467 --> 00:30:53,275
today. Like, start getting close to

655
00:30:53,315 --> 00:30:56,131
where your money is going. How is it going? What can you

656
00:30:56,163 --> 00:30:59,123
do to make it go in a better place today if you're

657
00:30:59,139 --> 00:31:00,531
not already doing it?

658
00:31:00,723 --> 00:31:03,403
>> Anthony Weaver: And that brings up a point of now

659
00:31:03,539 --> 00:31:06,195
want to learn more about you, if you don't mind.

660
00:31:06,355 --> 00:31:06,390
>> Sumedha Rai: U.

661
00:31:06,390 --> 00:31:08,879
>> Anthony Weaver: Uh, it's like, did your

662
00:31:08,927 --> 00:31:11,711
parents, um, bring

663
00:31:11,743 --> 00:31:14,647
you into the financial realm or is this something that you just

664
00:31:14,671 --> 00:31:17,663
kind of like? You know what? I want to learn more about finances because we didn't talk about

665
00:31:17,679 --> 00:31:18,475
it in the house.

666
00:31:18,910 --> 00:31:20,715
>> Sumedha Rai: Uh, right now I think

667
00:31:22,415 --> 00:31:25,375
this is a really cool question. I would say

668
00:31:25,415 --> 00:31:27,270
yes and no. Okay, so,

669
00:31:27,270 --> 00:31:29,623
uh, my parents are actually

670
00:31:29,679 --> 00:31:32,480
doctors, so n, um,

671
00:31:33,515 --> 00:31:36,427
topics about investing were not exactly table

672
00:31:36,491 --> 00:31:38,590
conversations. Uh, not,

673
00:31:38,590 --> 00:31:41,555
um, we were not discussing this over dinner.

674
00:31:41,595 --> 00:31:44,403
I could tell you so much about, uh, Tylenol

675
00:31:44,459 --> 00:31:47,227
and Advilce and the thoughts

676
00:31:47,291 --> 00:31:49,150
that make up these medicines. Um,

677
00:31:49,835 --> 00:31:52,299
but, uh, we were not exactly

678
00:31:52,387 --> 00:31:55,347
discussing where to put the money or what to do with the

679
00:31:55,371 --> 00:31:58,339
next paycheck. But I do remember that

680
00:31:58,387 --> 00:32:01,299
my dad was very particular about doing his own

681
00:32:01,347 --> 00:32:04,327
taxes with his accountant So I remember he used to

682
00:32:04,351 --> 00:32:07,351
maintain this, uh, book which had all his

683
00:32:07,383 --> 00:32:10,191
transactions written down, so that when he's

684
00:32:10,223 --> 00:32:12,195
giving this information to the accountant,

685
00:32:12,975 --> 00:32:15,695
he's not a person who doesn't know

686
00:32:15,855 --> 00:32:18,703
what, where his money is going and how it's being treated or, like, how

687
00:32:18,719 --> 00:32:21,623
his taxes are getting filed. He knows exactly what

688
00:32:21,639 --> 00:32:24,135
he's giving to the accountant. Um, he knows his

689
00:32:24,175 --> 00:32:27,047
transactions. There are times when he would just like, call his accountant and say, oh,

690
00:32:27,071 --> 00:32:29,855
I see this over here, but I think I did this. Can you go to, like,

691
00:32:29,895 --> 00:32:32,875
page 14 and check this out? I have it in my note.

692
00:32:33,505 --> 00:32:36,481
And, um, that image stuck with

693
00:32:36,513 --> 00:32:39,249
me. And, um, I didn't

694
00:32:39,297 --> 00:32:42,153
start investing right away. Truth be told, I

695
00:32:42,209 --> 00:32:44,977
entered the investing game a little later. But

696
00:32:45,121 --> 00:32:47,889
now that I'm already in it, I do understand

697
00:32:47,977 --> 00:32:50,753
that if I'm managing my own money,

698
00:32:50,929 --> 00:32:53,857
I feel a lot more comfortable with it. Even if I'm

699
00:32:53,881 --> 00:32:56,777
asking someone else to do it with me. I would feel

700
00:32:56,801 --> 00:32:59,497
a lot more confident if I also had knowledge about

701
00:32:59,521 --> 00:33:02,207
it. And, um, this

702
00:33:02,311 --> 00:33:05,231
whole, um, uh, making sure that

703
00:33:05,263 --> 00:33:07,951
you know exactly what you spending habits

704
00:33:07,983 --> 00:33:10,475
is, is the key to it.

705
00:33:10,610 --> 00:33:13,447
Um, my parents used to discuss these things with

706
00:33:13,471 --> 00:33:15,967
each other. So, like, sometimes I would hear a

707
00:33:15,991 --> 00:33:18,935
conversation. So I know the conversations were happening

708
00:33:18,975 --> 00:33:21,847
in the house, but they were not exactly happening with us. So

709
00:33:21,871 --> 00:33:24,767
maybe this is what I would advise people,

710
00:33:24,951 --> 00:33:27,759
that there are certain conversations that you can also

711
00:33:27,847 --> 00:33:30,800
start having with your kids. Um,

712
00:33:30,800 --> 00:33:33,643
what does it mean to save? Um, what

713
00:33:33,659 --> 00:33:36,659
is actually the value of money? There's a

714
00:33:36,787 --> 00:33:39,620
really cool, uh, thing that I remember, uh,

715
00:33:39,620 --> 00:33:42,467
when I was talking to my nephew this one time, I asked him what did he

716
00:33:42,491 --> 00:33:45,075
want for his birthday? And he

717
00:33:45,115 --> 00:33:47,739
said, you could either get

718
00:33:47,787 --> 00:33:50,723
me a fidget spinner or you

719
00:33:50,739 --> 00:33:51,975
could get me an iPhone.

720
00:33:55,115 --> 00:33:58,091
And I remember thinking, oh, my

721
00:33:58,123 --> 00:34:01,085
God. I don't think you, that he has any idea of what,

722
00:34:01,125 --> 00:34:03,925
like, the cost of these two things are.

723
00:34:04,085 --> 00:34:04,549
>> Anthony Weaver: Right?

724
00:34:04,637 --> 00:34:07,400
>> Sumedha Rai: And this is just a funny example to say that,

725
00:34:07,400 --> 00:34:10,157
um, start teaching your kids

726
00:34:10,221 --> 00:34:13,173
the value of money. Like, what can work

727
00:34:13,229 --> 00:34:15,357
for them? What is

728
00:34:15,461 --> 00:34:18,413
$10? What can it buy? What is,

729
00:34:18,469 --> 00:34:21,285
you know, a thousand dollars? What's a good

730
00:34:21,325 --> 00:34:24,165
income to live off of, say, in New York City?

731
00:34:24,245 --> 00:34:27,157
And I'm not saying you need to have these conversations when the kids 3

732
00:34:27,181 --> 00:34:29,305
years old, but

733
00:34:29,905 --> 00:34:32,393
when they're in their teens or when they're in high

734
00:34:32,449 --> 00:34:35,361
school. Um, these are conversations that should

735
00:34:35,393 --> 00:34:38,149
be almost a part of our, um,

736
00:34:39,185 --> 00:34:40,365
daily talk.

737
00:34:40,659 --> 00:34:43,441
Uh, another thing that I would strongly recommend everyone

738
00:34:43,513 --> 00:34:46,205
to do is get close to your taxes.

739
00:34:46,599 --> 00:34:49,481
Um, in the last five years, I've just started doing

740
00:34:49,513 --> 00:34:51,937
taxes myself. It just makes

741
00:34:52,001 --> 00:34:54,873
me, uh, it forces me to look into

742
00:34:55,049 --> 00:34:57,937
what I did with my money retrospectively. Helps me

743
00:34:57,961 --> 00:35:00,829
plan a little bit. Uh, which is not to say that

744
00:35:00,937 --> 00:35:03,917
a lot of people have much more complex taxes.

745
00:35:03,981 --> 00:35:06,397
They have business entities that they need to take care

746
00:35:06,421 --> 00:35:09,085
of. But you could hand it to an

747
00:35:09,125 --> 00:35:12,069
accountant, but also exactly know how the tax

748
00:35:12,117 --> 00:35:14,869
was filed. U. You are going to learn so much

749
00:35:14,917 --> 00:35:17,757
more when you do your own taxes. And you might even get

750
00:35:17,781 --> 00:35:20,205
ideas about, like, what to do better next

751
00:35:20,245 --> 00:35:20,825
year.

752
00:35:21,245 --> 00:35:24,181
>> Anthony Weaver: So what tools are you using to file your taxes if you don't m mind, like, caus

753
00:35:24,213 --> 00:35:27,173
I was thinking, like, the big box ones, and I have

754
00:35:27,189 --> 00:35:30,061
a tax account that comes on, and she was like, well, you can

755
00:35:30,093 --> 00:35:33,093
go with them, but if you really want the money,

756
00:35:33,189 --> 00:35:35,305
you got to go to somebody now.

757
00:35:35,605 --> 00:35:38,357
>> Sumedha Rai: That's true. I'm not an expert on, like, tax

758
00:35:38,421 --> 00:35:41,381
structures and the best way to save money, so I use very simple

759
00:35:41,413 --> 00:35:43,941
tools. Right now, I'm. I so like, pick up

760
00:35:43,973 --> 00:35:46,837
TurboTax or Sprint Tax, but then it asks me

761
00:35:46,941 --> 00:35:49,265
so many questions to fill out

762
00:35:50,325 --> 00:35:53,181
a return that I feel like when it's

763
00:35:53,213 --> 00:35:56,157
asking me so many questions, I have to go back and look

764
00:35:56,181 --> 00:35:59,170
at the answers to those questions. And that informs me u.

765
00:35:59,170 --> 00:36:01,871
Uh, a lot more than. Than just

766
00:36:01,903 --> 00:36:04,735
like, giving everything, giving my W2 to the

767
00:36:04,775 --> 00:36:07,711
accountant and, um, and just asking the

768
00:36:07,743 --> 00:36:10,607
person to file the tax or e file it. Um, my

769
00:36:10,631 --> 00:36:13,615
point is, like, going to an accountant and getting

770
00:36:13,655 --> 00:36:16,455
the right help and getting the right ways to, uh, save money

771
00:36:16,535 --> 00:36:18,751
or filing the right returns and getting the right

772
00:36:18,783 --> 00:36:21,631
refunds, um, it's a very good thing

773
00:36:21,663 --> 00:36:24,647
to do. But also, do not make this a

774
00:36:24,671 --> 00:36:27,567
thing where you're completely uninformed. Because I've seen people just

775
00:36:27,591 --> 00:36:30,023
say, I'm not really sure my accountant does

776
00:36:30,039 --> 00:36:32,931
it. I'm not a big fan of

777
00:36:32,963 --> 00:36:35,175
that answer. Um, honestly.

778
00:36:35,995 --> 00:36:38,739
>> Anthony Weaver: Well, it shows up in your work ethic.

779
00:36:38,827 --> 00:36:41,779
And actually not. It goes back to the data set that you

780
00:36:41,787 --> 00:36:44,375
were providing, your AI model. It's like,

781
00:36:44,795 --> 00:36:47,691
is it doing what you wanted to do? And if it's

782
00:36:47,723 --> 00:36:50,539
not, you know why? Because you

783
00:36:50,627 --> 00:36:53,523
understand what the data set that you've given it, in a way, it

784
00:36:53,579 --> 00:36:56,491
should operate, and it's not. So it actually shows up in

785
00:36:56,523 --> 00:36:59,403
everything that you do, which I have picked up on just in

786
00:36:59,419 --> 00:37:02,275
our short conversation. Um, I'm sure you're like,

787
00:37:02,315 --> 00:37:04,915
you know, why is my food

788
00:37:04,995 --> 00:37:07,459
tasting this way? Do I need to add an extra

789
00:37:07,507 --> 00:37:08,051
spice?

790
00:37:08,203 --> 00:37:11,163
>> Sumedha Rai: Oh, my God, I'm such a big fan of research. Like, if

791
00:37:11,179 --> 00:37:14,035
I don't Understand it. I sometimes go crazy and I've been told

792
00:37:14,075 --> 00:37:16,859
this in the past that we don't exactly

793
00:37:16,907 --> 00:37:18,695
need 100% accuracy.

794
00:37:19,515 --> 00:37:22,243
20% will do it. And sometimes I

795
00:37:22,299 --> 00:37:25,291
deliver. But I see myself going back and saying, okay, yeah,

796
00:37:25,323 --> 00:37:26,855
but what happened to the 10%?

797
00:37:27,155 --> 00:37:28,455
>> Anthony Weaver: Yes, yes.

798
00:37:28,750 --> 00:37:28,760
>> Sumedha Rai: Um.

799
00:37:29,195 --> 00:37:31,987
>> Anthony Weaver: Cause it gets you to think about like these models that go

800
00:37:32,011 --> 00:37:34,859
out, um, far as like, oh yeah, you know, 29%

801
00:37:34,907 --> 00:37:37,739
of the people are doing this. So I'm like,

802
00:37:37,787 --> 00:37:40,627
well, what are the 80 mean or

803
00:37:40,651 --> 00:37:43,435
the 71% of the people doing that? The

804
00:37:43,475 --> 00:37:44,667
29% are doing?

805
00:37:44,731 --> 00:37:47,651
>> Sumedha Rai: Like, uh, you know, this is actually a recurring theme

806
00:37:47,683 --> 00:37:50,195
in uh, AI as well. The,

807
00:37:50,355 --> 00:37:53,155
the choice between explainable versus

808
00:37:53,275 --> 00:37:56,227
explainable AI versus black box models.

809
00:37:56,411 --> 00:37:59,231
And um, there are certain

810
00:37:59,383 --> 00:38:02,367
businesses that actually choose explainable

811
00:38:02,431 --> 00:38:05,391
AI over black box models, even though black box

812
00:38:05,463 --> 00:38:08,415
models are going to give them better metrics just because the

813
00:38:08,455 --> 00:38:11,063
explainability part is so critical to the

814
00:38:11,119 --> 00:38:13,879
business. Like, um, if, if

815
00:38:13,967 --> 00:38:16,727
a doctor was using an AI model and really wanted

816
00:38:16,751 --> 00:38:19,631
to understand the result, um, he or she needs to

817
00:38:19,663 --> 00:38:22,351
go back and see exactly how the result was generated.

818
00:38:22,503 --> 00:38:24,931
Cann I just go like, I don't know what happened.

819
00:38:25,003 --> 00:38:27,499
>> Anthony Weaver: This I just handed to the AI team.

820
00:38:27,547 --> 00:38:30,475
>> Sumedha Rai: Yeah. And um, believe you me,

821
00:38:30,515 --> 00:38:33,347
people are actually doing that in the development world where

822
00:38:33,371 --> 00:38:36,331
they say, they wash their hands off and say, I'm not

823
00:38:36,363 --> 00:38:39,040
really sure how this result, um, uh,

824
00:38:39,395 --> 00:38:42,387
you know, was generated because I fed this to

825
00:38:42,411 --> 00:38:45,211
the model and this is what it gave me. Now

826
00:38:45,323 --> 00:38:48,227
for certain things it might be okay. Like if I'm

827
00:38:48,251 --> 00:38:51,131
generating a summary of something and I get high

828
00:38:51,163 --> 00:38:53,851
level bullet points, I'm probably okay with a black box

829
00:38:53,883 --> 00:38:56,793
model as long as the result actually looks okay for

830
00:38:56,849 --> 00:38:59,761
something like, uh, you know, mark this as spam or

831
00:38:59,793 --> 00:39:02,681
not. I'm not too worried about, um, just marking

832
00:39:02,753 --> 00:39:05,729
something else's spam and saying the next time, just make sure this is

833
00:39:05,777 --> 00:39:08,737
also included. And I'm not really fussy about why did you

834
00:39:08,761 --> 00:39:11,441
mark something as spam or not spam? I mean maybe

835
00:39:11,473 --> 00:39:14,393
not, but for certain things, like as

836
00:39:14,409 --> 00:39:17,329
I was mentioning, credit decisioning, if a person

837
00:39:17,377 --> 00:39:20,105
is denied credit, um, I need to know

838
00:39:20,185 --> 00:39:21,650
exactly what happened, um,

839
00:39:22,815 --> 00:39:25,791
at what part in the model pipeline was

840
00:39:25,823 --> 00:39:28,395
this decision made to go from,

841
00:39:28,935 --> 00:39:31,679
yes, this is what I looked at and like this is

842
00:39:31,767 --> 00:39:34,343
how I came to the decision. I need to know these things.

843
00:39:34,519 --> 00:39:36,959
So yeah, explainability is very important to

844
00:39:36,967 --> 00:39:37,567
me.

845
00:39:37,751 --> 00:39:40,351
>> Anthony Weaver: That is awesome and I appreciate that

846
00:39:40,423 --> 00:39:43,095
you taking the level of detail to go

847
00:39:43,135 --> 00:39:45,831
through, look at all the fun stuff.

848
00:39:45,943 --> 00:39:48,439
That is not really many air quotes here with the fun

849
00:39:48,487 --> 00:39:49,115
stuff.

850
00:39:49,250 --> 00:39:52,133
Uh, because I'm thinking about like now

851
00:39:52,269 --> 00:39:55,173
as we get older, we move into like the third segment of the show, which

852
00:39:55,189 --> 00:39:57,997
is the features. Um, and

853
00:39:58,141 --> 00:40:00,485
I'm thinking about like when your parents are getting

854
00:40:00,525 --> 00:40:03,229
older, are you

855
00:40:03,277 --> 00:40:05,905
planning to have them live with you doing,

856
00:40:06,290 --> 00:40:08,549
um, what they call it, the assisted

857
00:40:08,597 --> 00:40:11,261
living. Have you thought about or had that

858
00:40:11,293 --> 00:40:12,825
conversation with your parents yet?

859
00:40:13,170 --> 00:40:16,025
>> Sumedha Rai: Um, I think yes. Um, we've

860
00:40:16,935 --> 00:40:19,703
almost started talking about it, I guess,

861
00:40:19,839 --> 00:40:22,783
but it's really a decision that's

862
00:40:22,959 --> 00:40:25,555
very, very personal to them. So

863
00:40:26,375 --> 00:40:29,235
I mean, I can have the conversations with them,

864
00:40:29,430 --> 00:40:32,423
um, but really, at the end of

865
00:40:32,439 --> 00:40:35,031
the day, it's exactly what they want that's gonna

866
00:40:35,063 --> 00:40:37,735
happen. I know

867
00:40:37,895 --> 00:40:40,799
and I know this decision is a very important one. And for

868
00:40:40,847 --> 00:40:43,617
different people based on different situations, it' the

869
00:40:43,681 --> 00:40:46,465
result might look very different. But at least in my

870
00:40:46,505 --> 00:40:49,281
case, um, it's really up to my

871
00:40:49,313 --> 00:40:51,670
parents. Um, they uh,

872
00:40:51,785 --> 00:40:54,521
were good with their finances, I

873
00:40:54,553 --> 00:40:57,073
think, and that'that's great. So

874
00:40:57,169 --> 00:40:59,513
they are able to make that choice

875
00:40:59,609 --> 00:41:02,361
independently. And, um, I will

876
00:41:02,393 --> 00:41:04,297
probably have very little say in this.

877
00:41:04,401 --> 00:41:07,225
>> Anthony Weaver: Okay. I actually think what it be

878
00:41:07,305 --> 00:41:10,057
possible that people can actually just put in their parent

879
00:41:10,121 --> 00:41:12,813
information into an AI model and be like,

880
00:41:12,869 --> 00:41:15,861
hey, what is the best solution in this

881
00:41:15,893 --> 00:41:18,837
situation? Because we can't figure it out. Just dump it

882
00:41:18,861 --> 00:41:21,425
all in and just say pick one.

883
00:41:21,530 --> 00:41:24,190
>> Sumedha Rai: Uh, wow, uh, wow. Uh,

884
00:41:24,190 --> 00:41:25,985
very, very sensitive question.

885
00:41:27,390 --> 00:41:29,813
Uh, to answer that question, I think

886
00:41:29,989 --> 00:41:32,385
yes, there could be a solution,

887
00:41:32,925 --> 00:41:35,309
but also coming back to the same

888
00:41:35,437 --> 00:41:38,361
conversation and um, first of all,

889
00:41:38,393 --> 00:41:41,049
you understand why that prediction was made and

890
00:41:41,097 --> 00:41:43,881
secondly, I mean, make sure that someone has a

891
00:41:43,913 --> 00:41:45,445
choice to say yes or no.

892
00:41:47,945 --> 00:41:50,345
And I'm sure like people are gonn to do that,

893
00:41:50,505 --> 00:41:53,217
but were something like

894
00:41:53,281 --> 00:41:56,113
this to get created, um, the

895
00:41:56,129 --> 00:41:59,097
end result should not be. I think the

896
00:41:59,121 --> 00:42:01,737
a said so, so this is what

897
00:42:01,761 --> 00:42:04,420
we'renna do. Uh,

898
00:42:04,420 --> 00:42:06,857
it should, um, also be like,

899
00:42:06,921 --> 00:42:09,841
oh, uh, this looks like a cool result. Maybe we

900
00:42:09,873 --> 00:42:11,729
can talk about it, I guess.

901
00:42:11,777 --> 00:42:14,737
>> Anthony Weaver: Yeah, I think that would be cool. Um, at least

902
00:42:14,761 --> 00:42:17,205
to get the conversation starteduse a non bias.

903
00:42:17,625 --> 00:42:20,625
Yeah, not saye person, but a non biased

904
00:42:20,665 --> 00:42:23,465
entity, uh, is providing a solution

905
00:42:23,545 --> 00:42:26,537
based on the data that is s provided. I think that'be pretty

906
00:42:26,561 --> 00:42:27,033
cool.

907
00:42:27,169 --> 00:42:29,961
>> Sumedha Rai: Yeah. Again, like, was it really non biased though?

908
00:42:30,033 --> 00:42:33,027
So understand what the

909
00:42:33,051 --> 00:42:35,947
data was trained on. You need to go

910
00:42:35,971 --> 00:42:38,787
to the very start of the solution and say,

911
00:42:38,851 --> 00:42:41,707
are you sure it's not biased? And yeah, that's when you

912
00:42:41,731 --> 00:42:43,331
rely on the non biased result.

913
00:42:43,403 --> 00:42:46,355
>> Anthony Weaver: So to say, uh, yes, yes, yes, full

914
00:42:46,395 --> 00:42:48,055
circle. I love it that you.

915
00:42:50,475 --> 00:42:53,379
So, um, right now do you want to focus on

916
00:42:53,427 --> 00:42:56,283
You. So what areas um,

917
00:42:56,339 --> 00:42:59,065
are you focusing on improving in your

918
00:42:59,105 --> 00:43:00,765
life or even your career?

919
00:43:01,665 --> 00:43:04,417
>> Sumedha Rai: Right. I think having the right conversations

920
00:43:04,521 --> 00:43:07,377
for me is very important to me at this point of time

921
00:43:07,401 --> 00:43:10,337
in my life. Um, I feel like

922
00:43:10,361 --> 00:43:12,721
I get 24 hours in a day

923
00:43:12,873 --> 00:43:15,685
and um, a lot of that is spent

924
00:43:16,145 --> 00:43:19,017
on working um, I don't know,

925
00:43:19,121 --> 00:43:22,065
like preparing for a speech at a conference

926
00:43:22,145 --> 00:43:24,977
or, or writing an article. I love doing these things.

927
00:43:25,041 --> 00:43:27,849
I inna put more information

928
00:43:27,937 --> 00:43:30,817
out there. Good, um, information out there

929
00:43:30,881 --> 00:43:33,673
if I can. Um, and so there

930
00:43:33,689 --> 00:43:36,329
isn't a lot of time that's honestly

931
00:43:36,417 --> 00:43:39,410
spent talking with people, uh,

932
00:43:39,585 --> 00:43:42,585
and having a uh, very long conversation.

933
00:43:42,705 --> 00:43:45,593
So when I do get the chance, like if I go at

934
00:43:45,609 --> 00:43:48,601
a conference or if I am m talking with a

935
00:43:48,633 --> 00:43:51,481
friend, I have started making this

936
00:43:51,553 --> 00:43:54,417
conscious effort to also bring up topics that

937
00:43:54,441 --> 00:43:56,965
are important for me to grow as a person.

938
00:43:57,410 --> 00:44:00,007
Um, and I really encourage everyone to do it. I

939
00:44:00,031 --> 00:44:02,783
mean um, that is not to say that you should not have

940
00:44:02,839 --> 00:44:05,695
conversations about uh, things that are not related

941
00:44:05,735 --> 00:44:08,655
to this list of important topics. It's um, I

942
00:44:08,735 --> 00:44:11,543
can totally understand gets very stressful. Sometimes you just need to

943
00:44:11,559 --> 00:44:14,475
vend. Sometimes you need to have conversations about,

944
00:44:14,815 --> 00:44:17,355
I don't know, um, shoes

945
00:44:18,935 --> 00:44:21,655
and uh, um. Those are also

946
00:44:21,695 --> 00:44:24,615
important conversations for your brain. But also start making

947
00:44:24,655 --> 00:44:27,327
a conscious effort to start having think the right

948
00:44:27,391 --> 00:44:30,359
conversations now. Um, the word

949
00:44:30,407 --> 00:44:33,175
right I wantn stress on it because the word

950
00:44:33,215 --> 00:44:35,975
right might um, be different for

951
00:44:36,015 --> 00:44:38,943
different people based on where they are in their lives.

952
00:44:39,119 --> 00:44:41,823
But if you have a list of these topics that you think are

953
00:44:41,879 --> 00:44:44,743
important, um, start talking to people

954
00:44:44,799 --> 00:44:47,631
about it. If they're in the industry or if they're in that

955
00:44:47,663 --> 00:44:50,635
domain, you're going to get a lot of good perspectives

956
00:44:50,935 --> 00:44:53,823
and it's going to help you build on top of it. If you're

957
00:44:53,839 --> 00:44:56,831
a person who likes to research about topics, just

958
00:44:56,863 --> 00:44:59,837
getting those conversations started, it is going to be very good for you.

959
00:44:59,861 --> 00:45:02,105
So for instance, when I go for a conference,

960
00:45:02,580 --> 00:45:05,425
um, and if I'm speaking at a conference,

961
00:45:05,720 --> 00:45:08,565
um, I'm very alert right before the speaking topic.

962
00:45:08,605 --> 00:45:11,445
I'm very very alert. So I

963
00:45:11,485 --> 00:45:14,173
would actually uh, try to get a slot

964
00:45:14,309 --> 00:45:17,029
that is later in the day which

965
00:45:17,077 --> 00:45:19,717
forces me to listen to all the conversations that are

966
00:45:19,741 --> 00:45:22,733
happening beforehand. And I would consciously make

967
00:45:22,749 --> 00:45:25,621
a decision to just go very early, listen to everything

968
00:45:25,773 --> 00:45:28,661
and really absorb. And as I mentioned before,

969
00:45:28,693 --> 00:45:31,477
there are times and I when I tell myself

970
00:45:31,541 --> 00:45:34,325
h. I did not think of it from this perspective. I

971
00:45:34,365 --> 00:45:37,061
already had an opinion. But this is also an extra

972
00:45:37,093 --> 00:45:39,650
opinion that could really, you know,

973
00:45:39,650 --> 00:45:42,253
um, um. I could have a different take

974
00:45:42,389 --> 00:45:45,333
on, on things if I looked at it from this point of

975
00:45:45,349 --> 00:45:48,141
view. And those things are important for me.

976
00:45:48,213 --> 00:45:50,845
So at this, at this time in my

977
00:45:50,885 --> 00:45:53,345
career, I'm looking to have

978
00:45:53,965 --> 00:45:55,050
good, useful, uh,

979
00:45:56,093 --> 00:45:58,661
conversations with a, uh, lot of

980
00:45:58,693 --> 00:46:01,330
people in my, um, in my,

981
00:46:01,330 --> 00:46:04,253
um, industry.

982
00:46:04,389 --> 00:46:07,293
In industry, it could be in my life, it could be a part of

983
00:46:07,309 --> 00:46:10,277
my friend circle. Um, I think when I

984
00:46:10,301 --> 00:46:13,093
hesitated with the word like industry,

985
00:46:13,269 --> 00:46:16,245
it's because sometimes the conversations are

986
00:46:16,325 --> 00:46:18,705
just about AI and how awesome it is.

987
00:46:19,325 --> 00:46:22,253
And then I met someone recently

988
00:46:22,309 --> 00:46:25,189
in journalism and she said,

989
00:46:25,357 --> 00:46:27,837
you know, I absolutely do not understand it at

990
00:46:27,861 --> 00:46:30,590
all. And my simple ask, um,

991
00:46:31,053 --> 00:46:33,997
for her was, okay, like, we don't have time

992
00:46:34,021 --> 00:46:36,957
right now, but let's set up some time, have a cup of coffee and just

993
00:46:37,021 --> 00:46:39,813
discuss what you don't understand about AI and I don't

994
00:46:39,869 --> 00:46:42,629
understand about journalism. And then let's inform each

995
00:46:42,677 --> 00:46:45,549
other about how it could be maybe a collaborative, collaborative

996
00:46:45,597 --> 00:46:48,573
effort. Uh, uh, some time back, I was speaking

997
00:46:48,629 --> 00:46:51,583
to a person who works in economics, and

998
00:46:51,639 --> 00:46:54,623
he said, oh my God, there are so many conversations around AI.

999
00:46:54,759 --> 00:46:57,663
Sometimes you don't need it. That is true

1000
00:46:57,839 --> 00:47:00,583
causality. I need causal inference. I need

1001
00:47:00,599 --> 00:47:03,351
to understand what thing caused the other

1002
00:47:03,383 --> 00:47:05,895
thing. I don't need predictions for my

1003
00:47:05,975 --> 00:47:08,695
problem. And I was like, yeah, that's, that's great.

1004
00:47:08,775 --> 00:47:11,687
I understand. I'm working in AI, but

1005
00:47:11,831 --> 00:47:14,479
not every cause has to be

1006
00:47:14,527 --> 00:47:17,527
championed by AI. There are certain things that might

1007
00:47:17,551 --> 00:47:20,455
not require it. So I'm really trying to have

1008
00:47:20,495 --> 00:47:23,461
these conversations with so many different people,

1009
00:47:23,653 --> 00:47:26,305
whether it's at a conference, whether it's at

1010
00:47:26,765 --> 00:47:29,437
a meetup, whether it's with people in my

1011
00:47:29,461 --> 00:47:32,445
own company. But I'm consciously making

1012
00:47:32,485 --> 00:47:35,237
an effort to carve out some time to have

1013
00:47:35,261 --> 00:47:38,221
these important conversations. And another thing

1014
00:47:38,253 --> 00:47:41,021
that I'm trying to do is to put my

1015
00:47:41,053 --> 00:47:43,933
efforts in the right direction. Um, that's

1016
00:47:43,989 --> 00:47:46,397
because in the industry that I'm working

1017
00:47:46,421 --> 00:47:49,351
in, there is so much noise. Like

1018
00:47:49,383 --> 00:47:52,183
the word AI is an umbrella term for

1019
00:47:52,319 --> 00:47:55,191
so many different things about it.

1020
00:47:55,383 --> 00:47:58,015
Yeah, I really need to do my research to

1021
00:47:58,095 --> 00:48:00,791
understand in the next three years, what am I going to

1022
00:48:00,823 --> 00:48:03,383
concentrate on? Am I going to start

1023
00:48:03,439 --> 00:48:06,335
learning a new language? Am I going to start

1024
00:48:06,375 --> 00:48:09,159
learning a new concept? Am I going to start looking at it

1025
00:48:09,207 --> 00:48:12,191
from a project point of view, a research point of view,

1026
00:48:12,223 --> 00:48:15,099
a business point of view, what am I going to do in

1027
00:48:15,107 --> 00:48:17,787
the next three years? And I'm trying to chart a course for

1028
00:48:17,811 --> 00:48:20,295
myself because this,

1029
00:48:20,360 --> 00:48:22,827
um, tech market is changing overnight

1030
00:48:22,891 --> 00:48:25,883
sometimes. So I need to make sure that I'm, you

1031
00:48:25,899 --> 00:48:28,499
know, I'm up to date with the topics.

1032
00:48:28,587 --> 00:48:31,435
I'm in a way ahead of the curve a little bit.

1033
00:48:31,555 --> 00:48:34,323
So I'm putting in a lot of time and effort to research

1034
00:48:34,419 --> 00:48:36,415
my path moving forward.

1035
00:48:36,740 --> 00:48:39,259
Um, conversations, uh, are a part of.

1036
00:48:39,267 --> 00:48:42,253
>> Anthony Weaver: It, um, because you got me thinking

1037
00:48:42,309 --> 00:48:45,160
about conversations just in general. Um,

1038
00:48:45,160 --> 00:48:48,060
and these are the things that I talk about, um,

1039
00:48:48,060 --> 00:48:50,941
with other podcasters. It's like the future of

1040
00:48:50,973 --> 00:48:53,749
podcasting. Um, what tools are people

1041
00:48:53,837 --> 00:48:56,065
using, how they streamlining their services

1042
00:48:56,365 --> 00:48:59,157
and so forth. Like, we can talk hours. Like, I could

1043
00:48:59,181 --> 00:49:02,173
talk your head off about this stuff. Um, and I'm sure you

1044
00:49:02,269 --> 00:49:05,205
can talk my head off about AI and like, how you guys are

1045
00:49:05,245 --> 00:49:08,237
looking into pretty much the same topics but

1046
00:49:08,261 --> 00:49:11,087
just at a different take of, of like, how are you streamlining your

1047
00:49:11,151 --> 00:49:14,127
process? How are you streamlining your code? How are you looking

1048
00:49:14,151 --> 00:49:17,143
in power consumption, uh, with these models? Because you

1049
00:49:17,159 --> 00:49:20,135
hear about, I think it says it's almost like an ocean's worth of

1050
00:49:20,175 --> 00:49:22,615
cooling that you need just to run some

1051
00:49:22,735 --> 00:49:23,391
programs.

1052
00:49:23,543 --> 00:49:26,119
>> Sumedha Rai: Energy corprint is very high right now. Yeah,

1053
00:49:26,207 --> 00:49:26,875
yeah.

1054
00:49:27,280 --> 00:49:30,255
>> Anthony Weaver: Um, and so it's just like how we

1055
00:49:30,295 --> 00:49:33,191
all have the same general topics.

1056
00:49:33,263 --> 00:49:36,057
It is just how do we approach it and

1057
00:49:36,201 --> 00:49:39,073
the mindset about how of it all, like,

1058
00:49:39,209 --> 00:49:42,201
then again goes back to do you even really need it?

1059
00:49:42,313 --> 00:49:45,129
Because sometimes the simplest solution, like how I think it was what

1060
00:49:45,177 --> 00:49:48,113
NASA when they went up in space and it was like, we gota find a pen

1061
00:49:48,169 --> 00:49:49,529
that can do anti gravity.

1062
00:49:49,617 --> 00:49:51,657
>> Sumedha Rai: My God, use a pen'a.

1063
00:49:51,681 --> 00:49:52,177
>> Anthony Weaver: Pist.

1064
00:49:52,321 --> 00:49:55,193
>> Sumedha Rai: Right, yeah, I've

1065
00:49:55,209 --> 00:49:58,049
heard about that. Uh, yeah, there were issues

1066
00:49:58,097 --> 00:50:00,921
with that as well. The lead can break and

1067
00:50:00,953 --> 00:50:03,681
stuff. But I understand, like, coming back and

1068
00:50:03,713 --> 00:50:06,551
saying simple can help, doesn't need to be

1069
00:50:06,583 --> 00:50:08,455
overco complicated. Don't overkill.

1070
00:50:08,535 --> 00:50:11,460
>> Anthony Weaver: Yeah, I think that might be the simplest,

1071
00:50:11,460 --> 00:50:14,035
uh, way to sum up this episode.

1072
00:50:14,695 --> 00:50:16,475
>> Sumedha Rai: Yeah, of course, yeah.

1073
00:50:17,430 --> 00:50:20,167
Um, I say this to, uh, startups

1074
00:50:20,191 --> 00:50:22,991
and businesses looking to integrate AI. Think about three

1075
00:50:23,023 --> 00:50:25,671
things for sure. Need time and

1076
00:50:25,703 --> 00:50:28,663
money. Do you need it? Do you

1077
00:50:28,679 --> 00:50:31,555
have the time to, um, create

1078
00:50:31,595 --> 00:50:34,491
it? Do you have the money to afford it? And I think the

1079
00:50:34,523 --> 00:50:36,955
need part is very important. Like, answer that question

1080
00:50:37,035 --> 00:50:40,030
first. Can it be done in a simpler way? Uh,

1081
00:50:40,030 --> 00:50:42,867
do you need to throw this into a chat bot or

1082
00:50:42,891 --> 00:50:45,843
a generative AI solution and pay for each of

1083
00:50:45,859 --> 00:50:48,763
those API calls? Or can you do it

1084
00:50:48,779 --> 00:50:50,175
in a simpler way? Yes.

1085
00:50:50,595 --> 00:50:51,379
>> Anthony Weaver: Love it.

1086
00:50:51,507 --> 00:50:54,459
U. Um, before we get to the final four, is, ah, there anything that you want

1087
00:50:54,467 --> 00:50:57,301
to leave the audience with before we dive into the final

1088
00:50:57,333 --> 00:50:57,861
four?

1089
00:50:58,013 --> 00:51:00,605
>> Sumedha Rai: I think this became like a recurring theme in this

1090
00:51:00,645 --> 00:51:03,645
podcast for me, but my mantra is research

1091
00:51:03,685 --> 00:51:06,505
it and prepare for it. So

1092
00:51:07,125 --> 00:51:09,557
specifically for this

1093
00:51:09,701 --> 00:51:12,461
technology that has, uh, so much going

1094
00:51:12,493 --> 00:51:15,141
on for it. There's so much talk about

1095
00:51:15,173 --> 00:51:17,789
this technology. Research the

1096
00:51:17,877 --> 00:51:20,605
technology, research how to use it, research how

1097
00:51:20,685 --> 00:51:23,621
to, um, how other people are using it, research how it's being

1098
00:51:23,653 --> 00:51:26,413
used in your industry, and then sort of like start

1099
00:51:26,469 --> 00:51:29,355
preparing for what comes next. Um, if you feel

1100
00:51:29,395 --> 00:51:32,211
that you need to upckill yourself, start taking those steps

1101
00:51:32,243 --> 00:51:33,015
right now.

1102
00:51:33,595 --> 00:51:36,331
>> Anthony Weaver: Perfect. All right, you ready for the final four?

1103
00:51:36,443 --> 00:51:38,323
>> Sumedha Rai: Yes, of course. Looking forward to it.

1104
00:51:38,419 --> 00:51:39,175
>> Anthony Weaver: Awesome.

1105
00:51:46,995 --> 00:51:49,907
Number one, what does wealth mean to

1106
00:51:49,931 --> 00:51:50,495
you?

1107
00:51:51,420 --> 00:51:54,040
>> Sumedha Rai: Um, uh, a, ah,

1108
00:51:54,221 --> 00:51:56,773
really, really cool question. And

1109
00:51:56,909 --> 00:51:59,785
I actually thought about it and I think,

1110
00:52:00,300 --> 00:52:02,665
um, I would say that

1111
00:52:03,365 --> 00:52:06,117
wealth for me is having the right

1112
00:52:06,221 --> 00:52:09,197
skills and the competence to

1113
00:52:09,221 --> 00:52:12,109
be able to generate the tradable currency

1114
00:52:12,157 --> 00:52:15,029
in the future if I need to. And

1115
00:52:15,117 --> 00:52:18,061
I say this because your skill

1116
00:52:18,093 --> 00:52:21,033
set is very important. Something could happen

1117
00:52:21,089 --> 00:52:23,961
tomorrow. You might not have your job, you might have

1118
00:52:23,993 --> 00:52:26,817
to use up your savings. Um, a lot of

1119
00:52:26,841 --> 00:52:29,841
things could happen to this, this fiat money

1120
00:52:29,993 --> 00:52:32,849
that you keep in your bank accounts or to this commodity

1121
00:52:32,897 --> 00:52:35,737
money or to like, uh, a land you on or

1122
00:52:35,801 --> 00:52:38,553
something. But do

1123
00:52:38,609 --> 00:52:41,305
you actually have the skills to generate more

1124
00:52:41,345 --> 00:52:44,025
tradable currency that we need, um, to

1125
00:52:44,065 --> 00:52:46,961
survive, to like, buy goods and

1126
00:52:47,033 --> 00:52:49,945
services for us? Do you have the skills? Do you have the competence

1127
00:52:49,985 --> 00:52:52,863
to do it? Do you feel confident, couldn't enough to do it? And if

1128
00:52:52,879 --> 00:52:55,727
you feel like, okay, I could

1129
00:52:55,871 --> 00:52:58,863
lose my job and be unemployed for the next six months,

1130
00:52:58,919 --> 00:53:01,823
but I'm pretty sure with my current skill set, I would be

1131
00:53:01,839 --> 00:53:04,775
able to get, um, on the right track to start

1132
00:53:04,815 --> 00:53:07,710
generating this wealth again. Um,

1133
00:53:07,710 --> 00:53:10,159
your skills, in my opinion, are then your real

1134
00:53:10,207 --> 00:53:13,175
wealth. And um, the

1135
00:53:13,215 --> 00:53:16,151
second more philosophical take on it is

1136
00:53:16,183 --> 00:53:19,087
that the definition of enough can be

1137
00:53:19,111 --> 00:53:20,835
very different for different people.

1138
00:53:21,435 --> 00:53:24,195
So do you feel

1139
00:53:24,315 --> 00:53:27,307
like you have enough and does that give

1140
00:53:27,331 --> 00:53:29,135
you a sense of contentment?

1141
00:53:29,795 --> 00:53:32,507
So that sense of contentment for me

1142
00:53:32,571 --> 00:53:35,515
is tied to the idea of wealth. A

1143
00:53:35,555 --> 00:53:38,523
person with two kids might have a different

1144
00:53:38,579 --> 00:53:41,427
sense of what wealth means than

1145
00:53:41,451 --> 00:53:44,415
a person who is maybe, you know,

1146
00:53:44,560 --> 00:53:47,139
um, does not have kids or has like dependent

1147
00:53:47,187 --> 00:53:49,883
parents or is working in a different industry

1148
00:53:50,059 --> 00:53:52,957
that, that you know, does not generate enough

1149
00:53:53,021 --> 00:53:56,013
numbers to get to a certain, um, salary

1150
00:53:56,069 --> 00:53:58,917
bracket. But at the same time, given

1151
00:53:58,981 --> 00:54:01,309
your current situations, given your current

1152
00:54:01,357 --> 00:54:03,701
geography, given how you've been

1153
00:54:03,773 --> 00:54:06,701
working, do you feel content with where

1154
00:54:06,733 --> 00:54:09,305
you are? Do you feel content with the number,

1155
00:54:09,690 --> 00:54:11,905
um, that you have in your bank account?

1156
00:54:12,725 --> 00:54:15,453
If yes, yes, you do have wealth,

1157
00:54:15,549 --> 00:54:18,157
and then you can plan for how to sustain this

1158
00:54:18,301 --> 00:54:21,285
sense of contentment. But if you feel

1159
00:54:21,325 --> 00:54:23,789
like I m. Don't think this is enough,

1160
00:54:23,957 --> 00:54:26,893
then ask yourself, what is your definition

1161
00:54:26,949 --> 00:54:28,985
of enough? And work towards it.

1162
00:54:29,605 --> 00:54:32,165
>> Anthony Weaver: Right. This is a sad

1163
00:54:32,205 --> 00:54:34,625
question. I'm just thinking of, like,

1164
00:54:35,085 --> 00:54:37,869
when you say enough, because I know you like to,

1165
00:54:38,037 --> 00:54:40,757
you know, like gu. Um, um, to do your research.

1166
00:54:40,941 --> 00:54:43,733
When is enough research for you to say, yes,

1167
00:54:43,789 --> 00:54:46,261
I'm moving forward and I believe this to be the

1168
00:54:46,293 --> 00:54:46,865
case.

1169
00:54:47,175 --> 00:54:49,435
>> Sumedha Rai: Know at some point I just get tired of the screen.

1170
00:54:52,010 --> 00:54:54,927
Uh, no, let me, uh. Yeah, that's

1171
00:54:54,951 --> 00:54:57,870
a good question. That's a good question. Uh,

1172
00:54:57,870 --> 00:55:00,647
I once told someone, you did not go to the seventh

1173
00:55:00,711 --> 00:55:02,395
page of Google Page.

1174
00:55:05,655 --> 00:55:08,543
I think enough for me is to start

1175
00:55:08,599 --> 00:55:11,230
feeling confident that I can,

1176
00:55:11,230 --> 00:55:13,807
um, start the Endeav.

1177
00:55:13,991 --> 00:55:16,439
So if I'm creating a

1178
00:55:16,487 --> 00:55:19,407
model for fraud detection and I know

1179
00:55:19,471 --> 00:55:22,295
nothing about it, enough for me

1180
00:55:22,415 --> 00:55:25,295
is to start looking into what space

1181
00:55:25,335 --> 00:55:28,115
I'm working on. What do the features mean to me?

1182
00:55:28,575 --> 00:55:31,543
What exactly are the rules that govern

1183
00:55:31,599 --> 00:55:34,351
the model that I'm looking at? And then at a

1184
00:55:34,383 --> 00:55:37,167
point where I feel, okay, let me just start coding this

1185
00:55:37,191 --> 00:55:39,875
out, I feel like at that point the research is.

1186
00:55:40,245 --> 00:55:43,093
And then I obviously, like, there are certain

1187
00:55:43,149 --> 00:55:45,917
questions that come up, um, when I'm

1188
00:55:45,941 --> 00:55:48,309
coding, then I come back to it again.

1189
00:55:48,477 --> 00:55:51,341
But the first part of research

1190
00:55:51,373 --> 00:55:54,125
is always longer than the successive

1191
00:55:54,165 --> 00:55:56,725
parts after that. It's like a marginal addition on top of your

1192
00:55:56,765 --> 00:55:59,221
knowledge. So, yeah, I think

1193
00:55:59,293 --> 00:56:02,277
typically enough for me means I'm ready

1194
00:56:02,301 --> 00:56:03,237
to get started.

1195
00:56:03,381 --> 00:56:04,029
>> Anthony Weaver: Okay.

1196
00:56:04,157 --> 00:56:06,917
>> Sumedha Rai: After that, it's going to be, um, an incremental value

1197
00:56:07,021 --> 00:56:09,345
to every data point that you give me.

1198
00:56:09,645 --> 00:56:12,293
>> Anthony Weaver: I like that. All right, number

1199
00:56:12,349 --> 00:56:14,925
two, what was your worst money

1200
00:56:14,965 --> 00:56:15,705
mistake?

1201
00:56:16,540 --> 00:56:19,125
>> Sumedha Rai: Uh, I think I said before I let my money sit in the

1202
00:56:19,165 --> 00:56:21,741
bank account, a, uh, savings bank account, generating

1203
00:56:21,773 --> 00:56:24,773
no, uh, interest for some time. And

1204
00:56:24,869 --> 00:56:27,693
I, I could have started sooner,

1205
00:56:27,869 --> 00:56:30,749
and I did start, but if there

1206
00:56:30,757 --> 00:56:33,717
is anyone out there who's still not doing it, get started

1207
00:56:33,781 --> 00:56:36,617
today. You need to make your money work for you

1208
00:56:36,641 --> 00:56:37,405
as well.

1209
00:56:38,105 --> 00:56:39,005
>> Anthony Weaver: Love it.

1210
00:56:39,465 --> 00:56:42,185
Number three, is there a book that inspire

1211
00:56:42,265 --> 00:56:44,445
your journey or change your perspective?

1212
00:56:45,385 --> 00:56:47,640
>> Sumedha Rai: My. Ah, God, this is a very tricky one, honestly.

1213
00:56:47,640 --> 00:56:50,337
Uh, there are so many different books that give

1214
00:56:50,361 --> 00:56:53,025
you so many different takes on

1215
00:56:53,065 --> 00:56:54,780
things. So, um,

1216
00:56:55,625 --> 00:56:58,433
sometimes it really depends upon what I

1217
00:56:58,449 --> 00:57:00,605
was looking for at that point in my life.

1218
00:57:01,140 --> 00:57:03,865
Um, and the book might change,

1219
00:57:03,945 --> 00:57:06,740
but then there's this book, um,

1220
00:57:06,740 --> 00:57:09,697
called the Alchemist. A lot of people have read it, of

1221
00:57:09,721 --> 00:57:12,673
course, and I think the reason why I'm

1222
00:57:12,689 --> 00:57:14,810
mentioning it is that it brings out a very, um,

1223
00:57:15,625 --> 00:57:18,020
very famous Kind of, um,

1224
00:57:18,385 --> 00:57:21,185
way to approach things in life. Start

1225
00:57:21,225 --> 00:57:24,073
learning from the journey. Like, start learning from

1226
00:57:24,089 --> 00:57:26,913
your experiences. Of course, like, having an end goal is very

1227
00:57:26,969 --> 00:57:29,969
important, but while you're chasing that end

1228
00:57:30,017 --> 00:57:32,871
goal, uh, don't forget that

1229
00:57:33,023 --> 00:57:35,935
every small thing, every small thing that goes

1230
00:57:35,975 --> 00:57:38,711
wrong is actually like, not really going wrong. You're

1231
00:57:38,743 --> 00:57:41,711
learning something from, um, it. Because

1232
00:57:41,903 --> 00:57:44,599
depending upon our goals, it could take

1233
00:57:44,727 --> 00:57:46,955
days, months, or years to get there.

1234
00:57:47,335 --> 00:57:50,127
And if we don't start learning from

1235
00:57:50,151 --> 00:57:53,071
our experiences, it's gonna feel like it's a

1236
00:57:53,103 --> 00:57:55,795
very long journey to get where we're trying to get.

1237
00:57:56,255 --> 00:57:58,647
But at the same time, if we have this mindset

1238
00:57:58,711 --> 00:58:01,139
saying, let me learn from every small

1239
00:58:01,187 --> 00:58:04,187
thing. Journal it if you want to, because when you look back at

1240
00:58:04,211 --> 00:58:07,059
it, it'll feel like a positive reinforcement and it's going

1241
00:58:07,067 --> 00:58:09,819
to make you, um, work in a more positive

1242
00:58:09,867 --> 00:58:12,507
way towards your new goal. But

1243
00:58:12,691 --> 00:58:15,410
start learning from every little thing. Um,

1244
00:58:15,410 --> 00:58:18,375
and I think that book really

1245
00:58:18,755 --> 00:58:21,259
talked about it in a very nice

1246
00:58:21,307 --> 00:58:23,867
philosophical way. So I actually ended up reading it

1247
00:58:23,891 --> 00:58:26,455
twice at different times in my life.

1248
00:58:26,790 --> 00:58:29,653
Um, another thing that I'm reading right now is Atomic Habits.

1249
00:58:29,829 --> 00:58:30,589
>> Anthony Weaver: Love that book.

1250
00:58:30,637 --> 00:58:33,413
>> Sumedha Rai: Yes, it's a great book. Um, we

1251
00:58:33,429 --> 00:58:36,373
are all, um, a slave to our habits. Sometimes we want

1252
00:58:36,389 --> 00:58:39,237
to break the old ones, the bad ones, and

1253
00:58:39,261 --> 00:58:42,093
we want to start with the new ones. And again, it has,

1254
00:58:42,229 --> 00:58:45,229
it has a very similar theme saying, you know, learn from the little

1255
00:58:45,277 --> 00:58:48,197
ones as well. But, um, if you guys haven't

1256
00:58:48,221 --> 00:58:51,021
checked out that book, you, you should think about reading it.

1257
00:58:51,133 --> 00:58:53,525
>> Anthony Weaver: Definitely. U, uh, yeah,

1258
00:58:54,825 --> 00:58:56,805
I'll leave that one. We'll talk offline.

1259
00:58:57,665 --> 00:58:58,445
>> Sumedha Rai: Yeah.

1260
00:58:59,425 --> 00:59:02,329
>> Anthony Weaver: U, uh, number four, what is

1261
00:59:02,377 --> 00:59:04,445
your favorite dish to make?

1262
00:59:05,065 --> 00:59:08,001
>> Sumedha Rai: Wow. Uh, I like

1263
00:59:08,073 --> 00:59:10,925
making a vegetarian lasagna. So,

1264
00:59:11,250 --> 00:59:13,849
um, I'm vegetarian, I love Italian

1265
00:59:13,897 --> 00:59:16,689
food. And, um, I

1266
00:59:16,777 --> 00:59:19,617
started experimenting with it, I think two or three years

1267
00:59:19,641 --> 00:59:21,905
ago. I burnt a lot of them over B.

1268
00:59:23,605 --> 00:59:26,581
But I've come to a point where I know how to

1269
00:59:26,613 --> 00:59:29,100
layer them properly and, uh,

1270
00:59:29,445 --> 00:59:32,309
to use different kinds of cheese. I know which one

1271
00:59:32,357 --> 00:59:35,213
works and I can customize it based

1272
00:59:35,269 --> 00:59:38,181
on dietary requirements at this point. And

1273
00:59:38,213 --> 00:59:40,973
honestly, this is sometimes my catchphrase

1274
00:59:41,069 --> 00:59:44,010
for someone who I'm inviting over, um,

1275
00:59:44,010 --> 00:59:46,549
as a guest. I'd be like, yeah, let's play some board

1276
00:59:46,597 --> 00:59:49,475
game. And you know what? I do a mean lasagna.

1277
00:59:49,635 --> 00:59:52,163
And, and just leave it at

1278
00:59:52,179 --> 00:59:52,735
that.

1279
00:59:54,475 --> 00:59:57,467
>> Anthony Weaver: Check them. Are you using, um, are

1280
00:59:57,531 --> 01:00:00,515
using like tofu? Like, what are you using as your, your meat

1281
01:00:00,555 --> 01:00:03,379
based, quote unquote meat bas. Uh, because I heard some people

1282
01:00:03,427 --> 01:00:06,163
use like lentils, um, or like some Type of

1283
01:00:06,179 --> 01:00:07,095
toefu base.

1284
01:00:07,435 --> 01:00:10,419
>> Sumedha Rai: I like to mix vegetables, uh, and

1285
01:00:10,467 --> 01:00:13,059
create like a medley. So I chop them up real

1286
01:00:13,147 --> 01:00:15,875
small. And so there's squash,

1287
01:00:15,995 --> 01:00:18,799
zucchini, um, um, spinach, carrot,

1288
01:00:18,827 --> 01:00:21,767
carro, um, mushrooms. And I like

1289
01:00:21,871 --> 01:00:24,647
to chop them up real, real nice. So then

1290
01:00:24,671 --> 01:00:27,631
it almost feels like, um, it's. It's

1291
01:00:27,703 --> 01:00:30,703
one base, but it's actually a medley of these bases. And

1292
01:00:30,719 --> 01:00:33,719
it. The flavor is really nice in your mouth.

1293
01:00:33,887 --> 01:00:36,783
>> Anthony Weaver: Yes. Oh, man, this is. I need to see

1294
01:00:36,799 --> 01:00:39,479
a picture. I. To follow your social. To see.

1295
01:00:39,647 --> 01:00:42,463
>> Sumedha Rai: Yeah, I'SEND one to you or, or

1296
01:00:42,519 --> 01:00:45,463
the next time that I'm inviting you to a party again,

1297
01:00:45,479 --> 01:00:48,471
I'm going to use this. I'm gonna say, hey, uh, let's play some board

1298
01:00:48,503 --> 01:00:50,785
games. Let's talk finance. But also, I make a mean

1299
01:00:50,855 --> 01:00:51,745
lasagna.

1300
01:00:52,125 --> 01:00:54,185
>> Anthony Weaver: Okay. I'll be sure to be hungry.

1301
01:00:56,765 --> 01:00:56,920
>> Sumedha Rai: All.

1302
01:00:56,920 --> 01:00:57,221
>> Anthony Weaver: Ah, right.

1303
01:00:57,253 --> 01:01:00,213
The very last question of the show, which is where could people

1304
01:01:00,269 --> 01:01:02,105
find out more about you?

1305
01:01:02,460 --> 01:01:05,160
>> Sumedha Rai: Um, I am on, um,

1306
01:01:05,205 --> 01:01:07,469
LinkedIn. I have a personal

1307
01:01:07,597 --> 01:01:10,189
website. Um, the website has a link

1308
01:01:10,237 --> 01:01:12,869
to shoot me a personal message as well, if you want

1309
01:01:12,917 --> 01:01:15,805
to. I love having conversations about

1310
01:01:15,845 --> 01:01:18,559
AI Whether you're a person working in it, it.

1311
01:01:18,647 --> 01:01:21,343
Whether you're a person looking to find out more about it,

1312
01:01:21,399 --> 01:01:24,039
just shoot me a message now. We'll make sure that I get back to

1313
01:01:24,047 --> 01:01:26,871
you. I want to enable as many conversations as

1314
01:01:26,943 --> 01:01:29,719
possible. Sometimes one on one is not always possible, so

1315
01:01:29,727 --> 01:01:32,650
I do a group conversation. Um,

1316
01:01:32,650 --> 01:01:35,035
but yeah, I'll make sure I get back to you.

1317
01:01:35,090 --> 01:01:38,075
Um, so hit me up, please.

1318
01:01:38,535 --> 01:01:41,055
>> Anthony Weaver: Awesome. Well, thank you so much,

1319
01:01:41,175 --> 01:01:44,119
Samitha. I greatly appreciate all of the

1320
01:01:44,167 --> 01:01:47,007
information that you provided to us. Um, especially about the

1321
01:01:47,031 --> 01:01:49,929
AI stuff. Like, I'm thinking I'm doing AI, but

1322
01:01:49,977 --> 01:01:52,945
then people like, well, as long, large, uh, language

1323
01:01:52,985 --> 01:01:55,921
models, and I'm like, okay, well, whatever is AI in

1324
01:01:55,993 --> 01:01:58,770
industry. So, uh,

1325
01:01:59,305 --> 01:02:02,129
you know, I greatly appreciate everything you

1326
01:02:02,177 --> 01:02:05,113
provided us. And one of the things, everybody, if you're listening, if

1327
01:02:05,129 --> 01:02:07,897
you made it to the end, I just want to let you know that you have what it

1328
01:02:07,921 --> 01:02:10,841
takes to go to that next level. You know, you listen

1329
01:02:10,873 --> 01:02:13,825
to Samita'story um, you saw

1330
01:02:13,865 --> 01:02:16,769
where her parents came from. You saw that she took a whole different

1331
01:02:16,817 --> 01:02:19,817
path than what they did. You have the ability to

1332
01:02:19,841 --> 01:02:22,833
change your life, your trajectory. All you have to do is just put in that

1333
01:02:22,849 --> 01:02:25,401
effort and take time out of your day to invest in

1334
01:02:25,433 --> 01:02:28,313
yourself. I wish you all the best. We out.

1335
01:02:28,489 --> 01:02:28,705
Peace.