March 31, 2023
#79 The Intersection of Data Science, Analytics, and AI: Insights from Mohammad Arshad
In this episode, we speak with Mohammad Arshad, a data science, analytics, and AI expert with extensive experience in academia and industry. Mohammad shares his insights on the different layers of data, the business outcomes that data science and analytics can solve, and the most successful implementations of AI solutions. He also discusses the importance of data collection practices, the verticals that should focus on adopting AI, and the future of AI and data science.
Find more about Mohammad here:
https://www.linkedin.com/in/mdarshad/
1
00:00:06,100 --> 00:00:08,800
Hello, and welcome to a new
episode of the city of show with
2
00:00:08,800 --> 00:00:09,900
Matt.
Matt, my name is Matt.
3
00:00:09,900 --> 00:00:14,000
Matt and in each episode, our
exploring latest trends inside
4
00:00:14,000 --> 00:00:17,200
strategies from different
topics, like cybersecurity,
5
00:00:17,200 --> 00:00:21,200
digital transformation, emerging
Tech and also startups and
6
00:00:21,200 --> 00:00:24,400
Entrepreneurship.
And as, you know, sometimes I
7
00:00:24,400 --> 00:00:28,500
have as guest thought leaders,
innovators and entrepreneurs,
8
00:00:28,800 --> 00:00:32,700
who would like to hear from,
Them more about their areas of
9
00:00:32,700 --> 00:00:37,300
expertise to tell us about all
the Innovations and all the
10
00:00:37,300 --> 00:00:39,100
latest news of tech and
business.
11
00:00:39,400 --> 00:00:43,000
And today, I am pleased to have
with me Muhammad our chat, who
12
00:00:43,000 --> 00:00:47,900
is an expert in data science,
and artificial intelligence.
13
00:00:47,900 --> 00:00:50,300
Thank you so much.
Mohammed for joining me today.
14
00:00:50,700 --> 00:00:54,000
If you can, just introduce
yourself to us and what you do.
15
00:00:55,000 --> 00:00:57,800
Thank you very much for the
wonderful introduction.
16
00:00:58,100 --> 00:01:00,400
I'm Muhammad ershad.
I have been a day.
17
00:01:00,500 --> 00:01:04,500
A scientist for 17 years of my
life incorporate.
18
00:01:04,500 --> 00:01:09,300
I was working for companies like
Dell have let Packard Accenture
19
00:01:09,300 --> 00:01:14,200
module for team in Dubai.
And recently I started my own
20
00:01:14,200 --> 00:01:19,000
startup decoding data science
that is into Consulting and
21
00:01:19,000 --> 00:01:24,800
Academy and I had a passion of
you can say giving mentorship
22
00:01:24,800 --> 00:01:32,400
and P creating teams data teams
because there is A lot of Gap in
23
00:01:32,400 --> 00:01:36,000
the requirement for companies
and the education system.
24
00:01:36,000 --> 00:01:39,500
So this and this from that
particular need, this startup
25
00:01:39,700 --> 00:01:42,200
emerged.
Where in I do lot of you can
26
00:01:42,200 --> 00:01:45,900
say, awareness sessions.
What are the main Foundation
27
00:01:45,900 --> 00:01:50,200
skills that are required and all
these things for people to learn
28
00:01:50,200 --> 00:01:53,400
Ai and not get scared of it?
That's great.
29
00:01:53,700 --> 00:01:58,000
So, the first question, I want
to ask him how much because That
30
00:01:58,000 --> 00:02:01,600
will lead us to more questions
but let's start from the basics,
31
00:02:01,600 --> 00:02:05,400
right?
So we hear a lot of these words
32
00:02:05,500 --> 00:02:10,800
but sometimes people mix them up
so we hear about data as data,
33
00:02:10,800 --> 00:02:14,500
then data science data analytics
and then you know AI.
34
00:02:14,600 --> 00:02:19,600
So can you little bit destruct
this to us like how it all stars
35
00:02:19,600 --> 00:02:23,200
and then you know what are the
layers that come on top of it?
36
00:02:24,700 --> 00:02:27,800
Yes, sir.
So to be honest, my mat.
37
00:02:27,800 --> 00:02:31,800
So there is a this particular
definition about artificial
38
00:02:31,800 --> 00:02:35,700
intelligence data science
machine learning becomes very
39
00:02:35,700 --> 00:02:41,200
complicated and there is a lot
of you can say, confusion or
40
00:02:41,200 --> 00:02:45,800
there in the market.
So, to be honest, how I talk to
41
00:02:45,800 --> 00:02:48,900
the business, specifically, to
the business, or any people who
42
00:02:48,900 --> 00:02:54,600
are new to this and who want to
know is, Whatever Buzz words
43
00:02:54,600 --> 00:03:00,400
that are there just the root
word of everything is data.
44
00:03:00,900 --> 00:03:06,700
So how we can focus on the data
and generate insights from the
45
00:03:06,700 --> 00:03:12,400
data, that is the concept of AI.
Now, it depends upon lot of
46
00:03:12,400 --> 00:03:16,700
conditions, the business
outcomes that you want and what
47
00:03:16,700 --> 00:03:20,800
is the data that is available?
After that only, we can
48
00:03:21,000 --> 00:03:26,000
understand which part.
Of the AI ecosystem will be used
49
00:03:26,000 --> 00:03:30,600
in your to help solve your
business problem or to be used
50
00:03:30,600 --> 00:03:35,400
in your specific case.
So if we can just think it has a
51
00:03:35,500 --> 00:03:37,900
tool kit, right?
There are a lot of tool kits
52
00:03:37,900 --> 00:03:43,300
available in the AI ecosystem
and you as a business user, do
53
00:03:43,300 --> 00:03:45,100
not need to know so much
details.
54
00:03:45,500 --> 00:03:51,200
You just need a consultant or an
expert's help to decode your
55
00:03:51,200 --> 00:03:55,000
problem.
And I suggest you which is the
56
00:03:55,000 --> 00:03:59,800
best tools it for you to use
because most of the cases, most
57
00:03:59,800 --> 00:04:03,700
of the businesses they don't
need to use AI to be honest,
58
00:04:03,700 --> 00:04:05,700
right?
They can do with simple plain
59
00:04:05,700 --> 00:04:10,600
reporting, but now we don't want
to explain this and we don't
60
00:04:10,600 --> 00:04:12,800
want to create that confusion to
them.
61
00:04:13,300 --> 00:04:18,300
We keep everything in the AI
ecosystem, even the reporting
62
00:04:18,300 --> 00:04:21,399
part.
Now if I can talk in more
63
00:04:21,399 --> 00:04:24,700
details about this, so there are
There is a road map for any
64
00:04:24,700 --> 00:04:29,300
company to start with AI.
And the first is making sure
65
00:04:29,300 --> 00:04:33,400
that the data is getting
collected properly, the, the
66
00:04:34,000 --> 00:04:38,200
pipelines are fixed, and there
is a way where people or the
67
00:04:38,200 --> 00:04:42,500
company is collecting data, that
is the first step second step is
68
00:04:42,500 --> 00:04:45,800
on that data.
Basic reporting you need to
69
00:04:45,800 --> 00:04:49,700
start doing and creating
insights from the reports, the
70
00:04:49,700 --> 00:04:51,900
data analysis part from the
reports.
71
00:04:52,400 --> 00:04:55,000
After all these is Fix up
because most of the business
72
00:04:55,000 --> 00:04:59,400
questions that are, there can be
answered by just doing data
73
00:04:59,400 --> 00:05:02,100
analysis, on historical data of
the company.
74
00:05:02,400 --> 00:05:07,400
That is there after that, if it
is required, according to the
75
00:05:07,400 --> 00:05:12,200
business quotient, we can use.
You can say any machine learning
76
00:05:12,200 --> 00:05:15,200
or AI, whatever tool kit that is
there.
77
00:05:15,200 --> 00:05:20,400
So you don't need to business.
Don't need to do much lost focus
78
00:05:20,400 --> 00:05:23,300
and get confused by this,
because the more the details
79
00:05:23,400 --> 00:05:27,000
Tails, we talk, right?
As a data expert, the more
80
00:05:27,000 --> 00:05:31,900
people get afraid and they get
away from this, which is not the
81
00:05:31,900 --> 00:05:35,600
case here.
Thank you very much for this
82
00:05:35,600 --> 00:05:36,600
nation.
Mohammed.
83
00:05:36,600 --> 00:05:41,600
Now, you talked about business
outcomes and I love this
84
00:05:41,600 --> 00:05:44,200
approach, right?
So because at the end of the day
85
00:05:44,400 --> 00:05:48,100
and I loved also, the way you
mentioned, solving business
86
00:05:48,100 --> 00:05:52,300
cases.
Now, what are the major
87
00:05:52,300 --> 00:05:56,600
problems?
You think, you know, data
88
00:05:56,600 --> 00:06:03,600
science, and data analysis, AI
are today focusing to solve for
89
00:06:04,100 --> 00:06:08,000
organizations.
Yeah, that's right.
90
00:06:08,000 --> 00:06:10,500
Now, to be honest, the main
challenges is, I was telling you
91
00:06:10,500 --> 00:06:15,400
the first part making sure that
the data is getting collected in
92
00:06:15,400 --> 00:06:20,700
a proper way.
And what all systems are there
93
00:06:20,700 --> 00:06:27,500
in place for data to get create
generated and you can say saved
94
00:06:27,800 --> 00:06:32,000
because this is very important
for any company to try to
95
00:06:32,000 --> 00:06:37,500
capture, whatever data it is
possible because See if the data
96
00:06:37,500 --> 00:06:41,500
culture and the data collection
system is not in place, people
97
00:06:41,500 --> 00:06:48,100
are still using just Standalone
excels and doing analysis, stand
98
00:06:48,100 --> 00:06:50,900
alone, and there is no unique
data.
99
00:06:50,900 --> 00:06:54,100
Mart in the company, where in
everything is getting stored at
100
00:06:54,100 --> 00:06:58,700
a daily level, then the, there
is a big problem.
101
00:06:58,700 --> 00:07:02,000
You cannot do any analysis, you
cannot do any reporting on that.
102
00:07:02,000 --> 00:07:05,500
So this is a major challenge for
many companies, even if there
103
00:07:05,500 --> 00:07:09,400
are Same place they have systems
in place but they are not
104
00:07:09,400 --> 00:07:12,500
storing the data even if they
are storing the data, it is not
105
00:07:12,500 --> 00:07:16,000
at one particular space and that
one particular level like
106
00:07:16,000 --> 00:07:21,300
suppose customer data if we want
to have all the customer data,
107
00:07:21,400 --> 00:07:25,100
all the attributes of a customer
at one place, it is very tough
108
00:07:25,100 --> 00:07:27,800
to get, they will be marketing.
Team will have some data as
109
00:07:27,800 --> 00:07:29,500
separate Place.
Finance, team will have some
110
00:07:29,500 --> 00:07:32,300
data separate plays, HR will
have data separate Place,
111
00:07:32,700 --> 00:07:35,800
transaction data, will be
separate employee demographics,
112
00:07:35,900 --> 00:07:38,600
graphic data will be separate.
So those different different
113
00:07:38,600 --> 00:07:40,500
things.
We need to get it at one place
114
00:07:40,500 --> 00:07:45,700
and build a customer view, store
level view, product level View
115
00:07:45,900 --> 00:07:49,800
and that is, the thing is, if
people are not able to that is
116
00:07:49,800 --> 00:07:54,200
one of the main challenges that
I feel currently is to have, you
117
00:07:54,200 --> 00:07:58,200
know, there is a lot of data,
but there's a lot of noise, how
118
00:07:58,200 --> 00:08:03,500
to get the clean and the
meaningful data and store it in
119
00:08:03,500 --> 00:08:07,100
a data Mart.
So that you can Value out of it,
120
00:08:07,100 --> 00:08:10,100
which is the one of the main
challenges and one of the main
121
00:08:10,100 --> 00:08:13,900
pain points that companies and
businesses are facing.
122
00:08:14,600 --> 00:08:20,400
Yeah, that's very insightful.
Now, this bring to me, you know,
123
00:08:20,400 --> 00:08:24,300
this, this question because you
talked about how this confusion
124
00:08:24,300 --> 00:08:26,700
can happen.
So what do you think?
125
00:08:26,700 --> 00:08:31,900
You know, like are the most
successful implementations of
126
00:08:31,900 --> 00:08:35,400
this AI solution?
Like how can you give us some
127
00:08:35,400 --> 00:08:37,799
examples?
Apple's of without even
128
00:08:37,799 --> 00:08:39,400
mentioning like, maybe customer
names.
129
00:08:39,400 --> 00:08:42,500
But at least, you know how the
outcomes.
130
00:08:42,500 --> 00:08:46,900
We're after implementing the
right AI framework at a company.
131
00:08:49,000 --> 00:08:52,800
So you are asking for my
examples or the examples that
132
00:08:52,800 --> 00:08:56,300
I've seen industry, you seen in
the industry are so.
133
00:08:56,300 --> 00:09:00,500
So the best, the best thing, the
best case and the many one of us
134
00:09:00,500 --> 00:09:03,400
will relate with.
It is the recommendation engine,
135
00:09:03,400 --> 00:09:05,800
that is a recommendation
algorithms that are there,
136
00:09:05,800 --> 00:09:08,900
right?
So those algorithms they use lot
137
00:09:08,900 --> 00:09:12,300
of data, right?
Like, suppose I give you example
138
00:09:12,300 --> 00:09:15,500
of Amazon.
Yeah, the products that you are
139
00:09:15,500 --> 00:09:21,100
buying, so the products that you
buy The recommendation algorithm
140
00:09:21,100 --> 00:09:24,900
does not only work on product to
product match.
141
00:09:25,400 --> 00:09:29,900
Like it also works on the type
of customer Persona that you
142
00:09:29,900 --> 00:09:34,600
have, right?
Because you if Suppose there is
143
00:09:34,600 --> 00:09:38,500
one General product, if you buy
suppose you buy a printer you
144
00:09:38,500 --> 00:09:42,800
will also buy ketteridge.
The that is a recommendation
145
00:09:42,800 --> 00:09:46,500
that they will but this this
algorithm goes much further.
146
00:09:47,300 --> 00:09:49,400
Right.
It will be, will see the type of
147
00:09:49,400 --> 00:09:56,900
cursed customer Persona is
buying pattern, which brand he
148
00:09:56,900 --> 00:10:00,000
likes.
What is the most probable
149
00:10:00,000 --> 00:10:02,800
products that he will buy along
with this product?
150
00:10:03,100 --> 00:10:08,400
It can be historical baskets of
the same personas that are
151
00:10:08,400 --> 00:10:11,600
there.
So those are the very useful use
152
00:10:11,600 --> 00:10:15,900
cases that the companies are
using and specifically the
153
00:10:15,900 --> 00:10:18,900
e-commerce companies.
He's to give recommendation
154
00:10:18,900 --> 00:10:22,000
engines.
This is, this is one of the best
155
00:10:22,000 --> 00:10:28,800
use cases of AI + ml in action,
and this really causes lot of
156
00:10:28,800 --> 00:10:31,600
upsells right up says, cross
sales.
157
00:10:32,000 --> 00:10:36,400
And companies are really
improving this model every day.
158
00:10:36,800 --> 00:10:42,000
If they are getting insights
from the customer, whether he is
159
00:10:42,000 --> 00:10:45,000
buying the recommended products
or not buying the recommended
160
00:10:45,000 --> 00:10:49,100
products, and that is the Back
Loop goes into the algorithm and
161
00:10:49,100 --> 00:10:50,900
it make becomes better and
better.
162
00:10:52,400 --> 00:10:54,600
Very, very insightful.
I would say.
163
00:10:54,700 --> 00:10:58,700
Now, you mentioned something
about, you know, and do I see
164
00:10:58,700 --> 00:11:02,900
that you highlighted a lot
about, you know, collecting the
165
00:11:02,900 --> 00:11:05,500
data, right?
So you said, like this is, this
166
00:11:05,500 --> 00:11:08,400
is the number one thing that any
company would do.
167
00:11:08,600 --> 00:11:11,300
Now, mentioning this, what do
you think?
168
00:11:11,300 --> 00:11:16,400
Are, you know, the things that
companies should care about when
169
00:11:16,400 --> 00:11:20,000
collecting the data?
I mean, whether from a privacy
170
00:11:20,000 --> 00:11:24,100
perspective, whether it's like,
from Um, the, you know, data I
171
00:11:24,100 --> 00:11:26,900
would say, health itself, how
clean it is.
172
00:11:27,300 --> 00:11:31,300
So what are the major out say
points?
173
00:11:31,600 --> 00:11:35,600
That really they should look at
when they start collecting the
174
00:11:35,600 --> 00:11:39,700
data?
So see for collecting data
175
00:11:39,700 --> 00:11:43,100
specifically.
There are some specific laws of
176
00:11:43,100 --> 00:11:45,400
a particular country, right,
life for yui.
177
00:11:46,200 --> 00:11:53,200
The if you are collecting data
of yui preferred customers local
178
00:11:53,200 --> 00:11:57,000
or the people who are in yui,
the data should reside in UAE
179
00:11:57,000 --> 00:11:59,600
itself.
So, this is a rule that is
180
00:11:59,600 --> 00:12:02,800
there.
And so we know one thing is, you
181
00:12:02,800 --> 00:12:07,100
need to know the rule of the
land, the data rules, there are
182
00:12:07,700 --> 00:12:13,000
The in place and second thing is
regarding in yui, there is
183
00:12:13,400 --> 00:12:19,300
specifically there is no, you
can say role that we cannot use
184
00:12:19,900 --> 00:12:25,600
gender and race in models,
right?
185
00:12:25,600 --> 00:12:30,100
But in some countries in us and
gdpr on you Europe, there are
186
00:12:30,100 --> 00:12:35,800
some rules that you cannot use
those type of variables in the
187
00:12:35,800 --> 00:12:41,400
model because it becomes Diverse
it is against the diversity.
188
00:12:41,400 --> 00:12:45,100
Lodge logic.
But here there is no, such rule
189
00:12:45,100 --> 00:12:48,900
for that in the model.
But there is Rule regarding the
190
00:12:48,900 --> 00:12:53,600
sharing of personal information
that is there.
191
00:12:53,600 --> 00:12:59,800
So the rights of the data, they,
when they are giving it to do
192
00:12:59,800 --> 00:13:05,100
analysis, it should be.
There should be a masking logic
193
00:13:05,100 --> 00:13:09,500
that or the names.
And email IDs and contact
194
00:13:09,500 --> 00:13:12,600
numbers.
When you are sharing with there
195
00:13:12,600 --> 00:13:15,500
should be not, you should be
randomized.
196
00:13:16,000 --> 00:13:18,500
And there should be a separate
dictionary of all the person
197
00:13:18,500 --> 00:13:23,000
informations and the team.
Only very specific teams should
198
00:13:23,000 --> 00:13:28,000
have access to the personal
information only for targeting
199
00:13:28,000 --> 00:13:31,300
purposes and all those things.
But the general the other team
200
00:13:31,300 --> 00:13:35,500
members What doing the analysis
and other departments?
201
00:13:35,900 --> 00:13:39,600
The technology department should
not have access to the real
202
00:13:40,000 --> 00:13:44,300
information information and it
can only be linked towards a
203
00:13:44,300 --> 00:13:46,500
random ideas of.
These are some of the basic
204
00:13:47,800 --> 00:13:52,700
principles for protecting the
personal information and but the
205
00:13:52,700 --> 00:13:55,900
real usage of models.
The information on models for
206
00:13:55,900 --> 00:14:00,000
gender is, there is nothing like
there is no rule specifically in
207
00:14:00,000 --> 00:14:02,800
this region.
Okay, I see.
208
00:14:02,800 --> 00:14:06,200
So smelly, they need to care
about, you know, protecting the
209
00:14:06,200 --> 00:14:09,500
data, having the right access to
it, to the people who should
210
00:14:09,500 --> 00:14:11,700
have.
And of course, follow the
211
00:14:11,700 --> 00:14:15,300
regulation of whatever land this
algorithm will work.
212
00:14:15,500 --> 00:14:18,300
Now you mentioned something,
which triggered a, you know,
213
00:14:18,300 --> 00:14:21,600
something in my mind and it's,
you know, I like it actually
214
00:14:21,600 --> 00:14:24,900
because you said AI systems,
they are not for everyone,
215
00:14:24,900 --> 00:14:28,600
right?
So what are really the companies
216
00:14:28,600 --> 00:14:33,600
that they should focus on?
Having a I like, which verticals
217
00:14:33,600 --> 00:14:35,600
do you see?
This is more relevant in.
218
00:14:37,100 --> 00:14:43,100
So the see the red areas like
see first is the retail retail
219
00:14:43,400 --> 00:14:48,400
wherein you have lot of product
information, a lot of custom
220
00:14:48,400 --> 00:14:54,300
information that is one place
where you can do a lot of data
221
00:14:54,300 --> 00:14:57,200
analysis.
Second is marketing.
222
00:14:58,700 --> 00:15:02,200
Zoe all the digital marketing.
That happens, Performance
223
00:15:02,200 --> 00:15:07,000
Marketing, that happens.
They they have lot of data, and
224
00:15:07,000 --> 00:15:11,600
you can do lot of.
You can say, Roi calculations
225
00:15:11,700 --> 00:15:16,900
return of, in of Investments,
the customer lifetime value, the
226
00:15:16,900 --> 00:15:22,200
effectiveness of your media
spends, there are some models
227
00:15:22,300 --> 00:15:26,500
specific in the marketing phase
space from the historical data.
228
00:15:26,700 --> 00:15:30,700
You can make Make models.
Like, if you spend one dollar on
229
00:15:30,700 --> 00:15:36,500
marketing, how much will be the
on which channel you will spend,
230
00:15:37,300 --> 00:15:40,100
how much will be the increase in
sales?
231
00:15:40,100 --> 00:15:42,300
So there are, this is the
marketing mix models that are
232
00:15:42,300 --> 00:15:44,700
there.
This really helps for marketers
233
00:15:44,700 --> 00:15:48,500
to know which is the most
effective channel to put the
234
00:15:48,500 --> 00:15:51,400
money dollars in that particular
channel.
235
00:15:51,400 --> 00:15:54,400
So marketing it is being used
Healthcare.
236
00:15:54,500 --> 00:16:00,100
It is very much used because
there a lot of to be honest
237
00:16:00,100 --> 00:16:04,100
Healthcare is the oldest the
Farmer's Life, Sciences and
238
00:16:04,100 --> 00:16:08,800
Health Care is one of the oldest
verticals that have been using
239
00:16:08,800 --> 00:16:11,400
artificial intelligence much
before.
240
00:16:11,400 --> 00:16:14,900
It was the hype has happened,
they have been using for a long
241
00:16:14,900 --> 00:16:19,100
time because all the drugs that
we use, they have been tested
242
00:16:19,200 --> 00:16:24,100
and statistically only they have
been You can say, they have
243
00:16:24,100 --> 00:16:28,300
passed the test significance
level test to for it to be
244
00:16:28,300 --> 00:16:30,600
rolled out to the masses.
So it is there.
245
00:16:30,600 --> 00:16:33,200
Those are the fields, to be
honest everywhere where there is
246
00:16:33,200 --> 00:16:36,600
data.
There is there generated you
247
00:16:36,600 --> 00:16:39,300
start using?
As I told you, AI is just not
248
00:16:39,300 --> 00:16:43,300
machine learning a, I can be
also be simple data analysis,
249
00:16:43,300 --> 00:16:46,000
right?
Because that is the level you
250
00:16:46,000 --> 00:16:48,800
start at once.
You are doing data analysis,
251
00:16:48,800 --> 00:16:52,100
once you are doing all those
things, you have a good history
252
00:16:52,100 --> 00:16:56,400
of Atta a one year of data, then
you can go ahead with doing.
253
00:16:56,800 --> 00:17:03,300
You can say Predictive
Analytics, but to be honest,
254
00:17:03,300 --> 00:17:08,700
whenever I don't say, when I say
AI is not for everyone, it is
255
00:17:10,400 --> 00:17:13,099
the modeling part.
The modeling part is not for
256
00:17:13,099 --> 00:17:18,000
everyone.
AI is, as I told you earlier on
257
00:17:18,200 --> 00:17:24,099
solving, the problem with data
it can be a simple data
258
00:17:24,099 --> 00:17:27,300
analysis.
Also it is, it may not be
259
00:17:27,300 --> 00:17:31,300
According to some, it may not be
defined as AI, but you are using
260
00:17:31,300 --> 00:17:35,700
data historical data to do some
analysis and tell to the
261
00:17:35,700 --> 00:17:38,600
management or the business that
this reform the data.
262
00:17:38,600 --> 00:17:43,800
This is the story and this can
But right, so this is the way I
263
00:17:43,800 --> 00:17:47,800
don't really confuse.
When I say I is not used for
264
00:17:48,100 --> 00:17:50,500
many of the people.
It is the modeling part.
265
00:17:50,500 --> 00:17:54,800
The advanced machine learning
part, but data analysis, anyone
266
00:17:54,800 --> 00:17:58,300
who has even 30 days of data,
they can also do some data
267
00:17:58,300 --> 00:18:01,400
analysis and they can derive
some value out of it.
268
00:18:02,200 --> 00:18:05,000
Yeah, I think this is a clear
the things more.
269
00:18:05,200 --> 00:18:09,300
Thank you for you know, this
clarification and yeah, like I
270
00:18:09,308 --> 00:18:11,000
agree with you.
You like, you know, even
271
00:18:11,000 --> 00:18:13,600
whenever you have some
historical data, this is kind of
272
00:18:13,800 --> 00:18:16,300
analysis you can conduct.
And I think this is the one of
273
00:18:16,300 --> 00:18:20,300
the oldest we had, you
mentioned, you know, the example
274
00:18:20,300 --> 00:18:23,300
of these algorithms that do
suggestions for you, similar to
275
00:18:23,300 --> 00:18:27,400
what you see at Amazon or, you
know, like even by social media,
276
00:18:27,400 --> 00:18:29,400
they use it.
So when they suggest for you,
277
00:18:29,400 --> 00:18:31,300
some people to connect with and
so on.
278
00:18:31,500 --> 00:18:33,800
So yeah, thank you for this
clarification.
279
00:18:34,000 --> 00:18:38,800
Now, moving forward.
What are the major threads that
280
00:18:38,800 --> 00:18:40,600
you are seeing in this?
This field.
281
00:18:40,600 --> 00:18:44,200
Now, we are seeing like
generative AI, a lot of hype
282
00:18:44,200 --> 00:18:48,100
about GPT and all this stuff.
So I know that everyone is
283
00:18:48,100 --> 00:18:49,900
talking about it but other than
Chad G.
284
00:18:49,900 --> 00:18:53,900
PT and degenerative AI, what are
like the other things that now
285
00:18:53,900 --> 00:18:57,800
are trending?
I would say in this field, So
286
00:18:57,800 --> 00:19:02,600
the thing is Muhammad.
So one thing is that is really
287
00:19:02,600 --> 00:19:08,000
going to happen and he's already
started is the way we search the
288
00:19:08,000 --> 00:19:12,700
way we have been searching.
So, since internet has come in
289
00:19:12,700 --> 00:19:16,200
90s, right?
We had used Yahoo index page,
290
00:19:16,200 --> 00:19:21,300
index base search and now Google
also built upon it.
291
00:19:21,800 --> 00:19:25,600
They also via we're searching
for a keyword and we were
292
00:19:25,600 --> 00:19:29,600
getting 10 or 15 or we are
getting out of pages and it was
293
00:19:29,600 --> 00:19:32,400
getting sorted according to the
ranking and then we have to
294
00:19:32,400 --> 00:19:35,800
search, which are the values
that we are getting that has
295
00:19:35,800 --> 00:19:39,500
been a search for almost 20 25,
25 years, right?
296
00:19:39,500 --> 00:19:41,400
This is the way we have been
searching internet.
297
00:19:42,000 --> 00:19:45,200
Now the survey we are going to
search is going to change.
298
00:19:46,000 --> 00:19:50,700
We now we are going to ask tools
like Taj uppity already being
299
00:19:50,700 --> 00:19:54,700
has started the being AI, very
new type in what you want.
300
00:19:55,700 --> 00:19:58,700
Like us because most of the
searches are like when you're
301
00:19:58,700 --> 00:20:02,500
searching other than you, you
are trying to ask a question to
302
00:20:02,600 --> 00:20:05,600
Google and Google is going to
give you the pages.
303
00:20:06,100 --> 00:20:08,100
But now what is these tools
doing?
304
00:20:08,100 --> 00:20:10,400
It is going to give you one
answer.
305
00:20:10,600 --> 00:20:14,400
It is like going to be more of a
conversational way.
306
00:20:14,400 --> 00:20:16,600
I wear in.
This is going to your prom that
307
00:20:16,600 --> 00:20:22,500
you're going to give it is going
to search for the Values that
308
00:20:22,500 --> 00:20:25,600
are most similar to your prompt
and going to give you a
309
00:20:25,600 --> 00:20:29,500
one-on-one output and this is
going to be.
310
00:20:29,600 --> 00:20:32,100
You may like it.
You may not like it, but this is
311
00:20:32,100 --> 00:20:36,300
the way because most of the
searches are not too technical,
312
00:20:36,300 --> 00:20:38,800
right?
If you see most of the searches
313
00:20:38,800 --> 00:20:41,700
that and eight ninety percent of
the searches are just for
314
00:20:41,700 --> 00:20:46,200
information.
And if for that, these steps
315
00:20:46,200 --> 00:20:48,600
type of tools really work.
Well, and this is going to
316
00:20:48,600 --> 00:20:52,100
change the way we interact.
ER act, it is going to change.
317
00:20:52,100 --> 00:20:56,300
Like this is the, the prompts
that are there, like, even Arts.
318
00:20:56,700 --> 00:21:00,400
You give prompts an artist
getting generated you give from
319
00:21:00,400 --> 00:21:03,900
say, videos getting generated.
So now the content that is going
320
00:21:03,900 --> 00:21:06,200
to come in, the market is going
to be very much.
321
00:21:07,000 --> 00:21:10,800
You can say generated by Ai and
it is maybe it is a concern as
322
00:21:10,800 --> 00:21:12,000
well.
And it is also.
323
00:21:13,500 --> 00:21:15,500
You can say it is going to
change the way it is.
324
00:21:15,500 --> 00:21:19,400
Just going to be a change.
In the way these things we are
325
00:21:19,400 --> 00:21:23,500
going to.
Interact with the internet and
326
00:21:23,500 --> 00:21:26,900
it is this is going to grow more
and more and to be honest more
327
00:21:26,900 --> 00:21:29,900
than the positives there are a
lot of negatives are as well.
328
00:21:29,900 --> 00:21:33,800
There are a lot of negatives.
These things are going to
329
00:21:33,800 --> 00:21:38,800
increase the garbage that is
there already on the internet
330
00:21:38,800 --> 00:21:41,400
because AI is learning from
garbage?
331
00:21:41,400 --> 00:21:44,600
That is already there.
When you train your model on
332
00:21:44,600 --> 00:21:47,700
garbage, you will get garbage.
And again, the same garbage
333
00:21:47,700 --> 00:21:51,200
again going back.
So you're In the cycle is
334
00:21:51,300 --> 00:21:53,100
completely now.
What is happening all the
335
00:21:53,100 --> 00:21:56,000
content that is in Internet is
just getting repurposed
336
00:21:56,000 --> 00:21:57,900
repurposed and the quality is
going down.
337
00:21:58,300 --> 00:21:59,900
So those are the things that are
happening.
338
00:21:59,900 --> 00:22:05,200
Ethics is a big problem, ethics
discrimination bias. - all those
339
00:22:05,200 --> 00:22:06,800
things, you will get more and
more.
340
00:22:07,100 --> 00:22:10,700
And there are a lot of
challenges and awareness is very
341
00:22:10,700 --> 00:22:15,700
important nowadays upon People
Like Us on the usage and how you
342
00:22:15,700 --> 00:22:19,000
need to control these things and
how you need to be a better
343
00:22:19,000 --> 00:22:21,200
contributor, which I can.
Talk more about later.
344
00:22:22,200 --> 00:22:26,600
Yeah, that's informative and
just to add one thing and I
345
00:22:26,600 --> 00:22:30,800
always repeat this on my
podcast, the guys what you are
346
00:22:30,800 --> 00:22:34,500
seeing is based on data, which
was generated at some point at
347
00:22:34,500 --> 00:22:38,500
some stage by us.
And yes, do you know what the
348
00:22:38,500 --> 00:22:41,100
algorithm is doing is just
bringing back this.
349
00:22:41,100 --> 00:22:44,400
So thank you, Mom at 4-4 like
focusing on this.
350
00:22:44,700 --> 00:22:48,400
It's not inventing the wheel.
Yes, it looks to you that.
351
00:22:48,500 --> 00:22:52,700
It's a text written by By
someone, that's it might look
352
00:22:52,700 --> 00:22:56,000
unique but it didn't learn by
itself, right?
353
00:22:56,000 --> 00:23:00,600
So it was all content, which was
available to the model, and this
354
00:23:00,600 --> 00:23:02,500
is why it looks so real, I
believe, right?
355
00:23:02,500 --> 00:23:06,800
So, but and you mentioned about
like, you know, this, the
356
00:23:06,800 --> 00:23:09,300
future.
I'm really interested to to, to
357
00:23:09,300 --> 00:23:12,700
know, or like at least to know
your point of view about the
358
00:23:12,700 --> 00:23:16,600
future that you just mentioned
like how are you seeing things
359
00:23:16,600 --> 00:23:20,500
moving forward with that?
Right now.
360
00:23:20,500 --> 00:23:27,100
See, I was just talking today
morning that you calculators
361
00:23:27,100 --> 00:23:29,300
when it was invented.
There was lot of protests and
362
00:23:29,300 --> 00:23:33,400
lot of hype and lot of people
who are against calculator
363
00:23:33,400 --> 00:23:38,500
usage, initially for people who
want to learn as, and the
364
00:23:38,500 --> 00:23:42,800
similar thing is now happening
with Chad GPT like many, you can
365
00:23:42,800 --> 00:23:45,300
say professors.
And many teachers are protesting
366
00:23:45,300 --> 00:23:49,200
the use of this because what is
this going to?
367
00:23:49,400 --> 00:23:54,000
It is going to hamper the
learning because these tools are
368
00:23:54,100 --> 00:23:57,600
making us more and more lazy.
As we can see, Mammoth the
369
00:23:57,600 --> 00:24:02,500
future as you, the technology is
there to help us, but what it is
370
00:24:02,500 --> 00:24:06,600
actually doing is it is making
people lazy, right?
371
00:24:06,600 --> 00:24:11,200
Everything at your fingertips.
It is humans are not designed to
372
00:24:11,300 --> 00:24:15,400
be so lazy, but these tools are
making us lazy.
373
00:24:15,500 --> 00:24:20,200
And so what is happening, only
the, the learning Capability of
374
00:24:20,200 --> 00:24:24,200
the masses is going to reduce
automatically so they are always
375
00:24:24,200 --> 00:24:29,200
they are the consumers, 95% of
the people are just consumers
376
00:24:29,600 --> 00:24:33,100
and they are just not giving any
value.
377
00:24:33,800 --> 00:24:37,900
They are just using their just
the people, the rest, five
378
00:24:37,900 --> 00:24:42,200
percent of the people who are
actually learning from this and
379
00:24:42,200 --> 00:24:44,600
understanding this.
And they're actually the numbers
380
00:24:44,600 --> 00:24:47,900
are now reducing more and more
because of this the usage of
381
00:24:47,900 --> 00:24:53,100
this and the debate about this.
The creativity is going down and
382
00:24:53,100 --> 00:24:56,400
it is the thing is you see
machines are getting more and
383
00:24:56,400 --> 00:25:00,100
more powerful.
Now because of this there are
384
00:25:00,300 --> 00:25:02,000
there is going to be lot of job
losses.
385
00:25:02,000 --> 00:25:06,300
People are not talking about it
so much, but a job that was
386
00:25:06,300 --> 00:25:09,100
being done by two people.
Now can be done by one people,
387
00:25:09,100 --> 00:25:12,000
one person if he is using this,
these AI tools.
388
00:25:12,500 --> 00:25:17,800
So people the jobs are going to
reduce many of the you can say
389
00:25:17,800 --> 00:25:21,100
manual, autonomous jobs, Manual
jobs that are there, they can
390
00:25:21,100 --> 00:25:25,200
get automated.
So there is going to be a huge
391
00:25:25,200 --> 00:25:27,700
shift.
And one more thing is there.
392
00:25:27,700 --> 00:25:33,300
Now, now companies have to
understand how they can use,
393
00:25:33,400 --> 00:25:35,300
they cannot blindly use these
tools.
394
00:25:35,700 --> 00:25:40,100
They need some knowledge.
They need some need to draw a
395
00:25:40,100 --> 00:25:44,600
line and it is to be honest.
It is a very debatable topic.
396
00:25:44,800 --> 00:25:49,900
These there is a news
yesterday's news that Lot of
397
00:25:49,900 --> 00:25:52,300
thought leaders have written to
open a.
398
00:25:52,300 --> 00:25:58,500
I even including Elon Musk to do
a pause on the development of
399
00:25:58,500 --> 00:26:02,300
GPT, 4.5, or five after the
release of four.
400
00:26:02,600 --> 00:26:07,400
They weren't a pause because
they weren't a lot of, you can
401
00:26:07,400 --> 00:26:13,600
say, SOB pure Sops to be put in
place about the ethics about the
402
00:26:13,600 --> 00:26:19,300
standards right now, which is
not there to make it more and
403
00:26:19,400 --> 00:26:21,200
more clean.
Because there is a lot of
404
00:26:21,200 --> 00:26:24,000
garbage that is there in open AI
as well.
405
00:26:24,600 --> 00:26:27,600
And they need to do lot of
procedures according to the
406
00:26:27,608 --> 00:26:31,300
situation and make sure that
they are not giving harmful
407
00:26:31,300 --> 00:26:34,600
information.
Because to be honest, even kids
408
00:26:34,600 --> 00:26:39,600
are using taji PT, right?
So the text that it comes out
409
00:26:40,200 --> 00:26:45,900
has to be controlled in a way
because it is going to as we
410
00:26:45,900 --> 00:26:50,400
know, young minds, if they are
Given such information.
411
00:26:50,400 --> 00:26:52,900
It is not good for them as well.
So these are the things that
412
00:26:52,900 --> 00:26:54,300
they have asked them to do a
pause.
413
00:26:54,700 --> 00:26:58,400
And there is a lot of debate
online as well about the how,
414
00:26:58,400 --> 00:27:02,500
how to be ready with for this
particular wave because it is
415
00:27:02,700 --> 00:27:05,600
going to change a lot of things.
Yeah.
416
00:27:07,500 --> 00:27:11,900
Like for the first time maybe on
the podcast I will say something
417
00:27:11,900 --> 00:27:17,100
before I close while I
understand, you know, the
418
00:27:17,100 --> 00:27:20,800
concerns.
I've seen the news also this
419
00:27:20,800 --> 00:27:25,600
morning, where people like Ellen
mosque at Steve Wozniak.
420
00:27:25,600 --> 00:27:32,400
The Apple co-founder they signed
this petition to, to open a, i
421
00:27:32,400 --> 00:27:37,700
the company behind the chat GPT
my out, save my opinion here.
422
00:27:39,000 --> 00:27:44,600
As I can understand the
frustration, I can understand
423
00:27:45,200 --> 00:27:47,900
how people are afraid about all
this.
424
00:27:50,600 --> 00:27:56,500
But I think like, I'm not able
to understand the need for pose,
425
00:27:56,500 --> 00:28:00,100
right?
Because I always try to link to
426
00:28:00,100 --> 00:28:02,500
previous historical moment,
right?
427
00:28:02,500 --> 00:28:07,000
So when the internet came up,
you know all the regulations
428
00:28:07,000 --> 00:28:09,600
came after the internet became
more widespread, right?
429
00:28:09,600 --> 00:28:12,800
So because people wanted to
understand what's this internet
430
00:28:12,800 --> 00:28:17,200
is and then try to put the legal
stuff again.
431
00:28:17,200 --> 00:28:21,900
I'm not saying that.
I agree that, you know, open AI
432
00:28:21,900 --> 00:28:25,800
is the best thing or chat GPT.
Of course not and I agreed with
433
00:28:25,800 --> 00:28:28,200
you about.
It has garbage, I agree with
434
00:28:28,200 --> 00:28:34,400
that but the, the reality is and
I've seen some articles about
435
00:28:34,400 --> 00:28:37,800
this.
Any company today, not opening
436
00:28:37,800 --> 00:28:41,100
our people think that open area
is the only player, but it's not
437
00:28:41,400 --> 00:28:45,600
anyone today can collect the
data and this is part of what a
438
00:28:45,600 --> 00:28:48,700
data, scientists, they do.
I think Mom Would you not agree
439
00:28:48,700 --> 00:28:50,600
with me?
So there's something a process
440
00:28:50,600 --> 00:28:53,700
called Data Mining, right?
So, you have to find the data,
441
00:28:53,700 --> 00:28:58,700
you have to bring the data and
it's, it's a small world now.
442
00:28:58,700 --> 00:29:03,400
Like anyone and I mean by
anyone, yes, anyone if they have
443
00:29:03,400 --> 00:29:08,000
the tools, they can get the data
and they can, I'm expecting,
444
00:29:08,000 --> 00:29:11,400
maybe they are not getting the
attention, but I'm expecting to
445
00:29:11,400 --> 00:29:14,300
see another, you know, chat GPT
with the same.
446
00:29:14,300 --> 00:29:17,600
I mean efficiency and all those
and very near future.
447
00:29:17,600 --> 00:29:20,300
So if open, I will not do it,
someone else gonna do it.
448
00:29:20,300 --> 00:29:25,500
So this is my opinion, the other
thing, yes, a lot of use cases
449
00:29:25,500 --> 00:29:29,700
can be and that is a question
mark about it.
450
00:29:31,500 --> 00:29:34,800
But what I have seen, you know,
from both let's say Microsoft
451
00:29:34,800 --> 00:29:38,700
and Google with, you know, the
integration that they are
452
00:29:38,700 --> 00:29:43,300
planning to do with the
day-to-day tools like Microsoft
453
00:29:43,300 --> 00:29:47,400
Office, or on the Google
workspace on Google side.
454
00:29:48,200 --> 00:29:52,600
I think, you know, any any topic
that can touch the business from
455
00:29:52,600 --> 00:29:57,500
the following or humans.
If it can reduce the time to do
456
00:29:57,500 --> 00:30:02,000
it, if it can reduce the cost to
do it, if it can make them more
457
00:30:02,000 --> 00:30:05,200
productive, I can understand the
fears.
458
00:30:05,300 --> 00:30:08,900
But you know, like why the same
hype didn't come when the phone
459
00:30:08,900 --> 00:30:10,200
was there?
Because the phone?
460
00:30:10,200 --> 00:30:12,800
Also, if you think about the
smartphone, I mean the
461
00:30:12,800 --> 00:30:17,600
smartphone moment.
It also made people kind of, you
462
00:30:17,600 --> 00:30:19,600
know, like the They get
attracted to it.
463
00:30:19,800 --> 00:30:22,900
Same for the social media.
I mean, whether it's Facebook,
464
00:30:22,900 --> 00:30:25,200
Twitter, whatever it is
Tick-Tock, right?
465
00:30:26,700 --> 00:30:29,600
I see what you are trying to
say.
466
00:30:29,700 --> 00:30:32,100
And I agree with you, by the
way, like I'm not fully
467
00:30:32,200 --> 00:30:37,500
disagreeing, but just to put my
two cents, I understand the
468
00:30:37,500 --> 00:30:40,200
concern.
But I would say it's too late to
469
00:30:40,200 --> 00:30:44,900
stop it and if not open a, if
openly I stopped today releasing
470
00:30:45,100 --> 00:30:47,900
GPT five or whatever, believe
me.
471
00:30:48,500 --> 00:30:52,100
Someone else can release that or
something even better.
472
00:30:52,100 --> 00:30:54,100
I think it's too late to stop
the train.
473
00:30:54,100 --> 00:30:59,300
It's at left the station.
I say it left the station, but
474
00:30:59,300 --> 00:31:04,000
of course, the effort should be
about how to force these
475
00:31:04,000 --> 00:31:07,800
companies like open a.i. to
clean the data to remove the
476
00:31:07,800 --> 00:31:10,600
garbage as you called it.
A hundred percent agree that
477
00:31:10,600 --> 00:31:14,800
it's there before we close any
final thoughts from your side
478
00:31:14,800 --> 00:31:16,700
Muhammad because we've come to
an end.
479
00:31:17,800 --> 00:31:18,800
You know, I know.
Yeah.
480
00:31:18,800 --> 00:31:22,500
So this to be honest I can talk
hours and hours about this topic
481
00:31:23,100 --> 00:31:26,900
and we can surely, I will just
summarize.
482
00:31:26,900 --> 00:31:30,400
So yeah, there are a lot of
challenges because was just as a
483
00:31:30,400 --> 00:31:37,000
takeaway, is somebody try to be
open about this change.
484
00:31:37,000 --> 00:31:44,000
So you need to prevent to even
to learn about the sea is Like a
485
00:31:44,000 --> 00:31:47,400
Knife, right?
It's a, the you can use it.
486
00:31:47,600 --> 00:31:50,700
For the wrong things.
Also, you can use it for cooking
487
00:31:50,700 --> 00:31:53,300
and cooking as well.
So you but you need to learn
488
00:31:53,400 --> 00:31:56,500
about knife.
So that is you need to learn the
489
00:31:56,500 --> 00:31:59,400
good things, the bad thing, so
how to use it when to use it.
490
00:31:59,700 --> 00:32:02,700
So think it like a tool and that
is the thing that I have tell
491
00:32:02,700 --> 00:32:05,900
everyone there is going to be
debate, there is going to be
492
00:32:05,900 --> 00:32:08,900
some misunderstanding, but at
the end of the day, it is just a
493
00:32:08,900 --> 00:32:13,600
tool to help humans.
So that is the my final thoughts
494
00:32:13,600 --> 00:32:16,000
about this and it was very nice
talking to you.
495
00:32:17,100 --> 00:32:18,500
Thank you.
Oh, thank you very much for your
496
00:32:18,500 --> 00:32:20,900
time today.
And again, I have to agree with
497
00:32:20,900 --> 00:32:24,000
the last sentence guys, train
yourself.
498
00:32:24,000 --> 00:32:26,200
Understand what is this tool
about?
499
00:32:26,200 --> 00:32:29,800
It's not what you see on Tik-Tok
or Instagram whatever because I
500
00:32:29,800 --> 00:32:33,100
start to see weird things there.
So there is a science, there is
501
00:32:33,100 --> 00:32:36,400
a science behind it, understand
the science and believe me the
502
00:32:36,400 --> 00:32:37,700
moment you understand the
science.
503
00:32:37,700 --> 00:32:40,400
You will see the benefit of it.
Thank you very much Muhammad for
504
00:32:40,400 --> 00:32:44,800
joining me today as usual, I'm
saying this to my audience.
505
00:32:44,800 --> 00:32:47,800
If you have any questions, if
you have any feedback about This
506
00:32:47,800 --> 00:32:51,200
episode or about the show.
If you want to be on the show,
507
00:32:51,200 --> 00:32:55,400
same as Muhammad, he was my
guest today, please feel free to
508
00:32:55,400 --> 00:32:57,800
contact me.
Whether by email LinkedIn or
509
00:32:57,800 --> 00:33:01,100
Twitter, I would be more than
happy to talk to you and don't
510
00:33:01,100 --> 00:33:03,400
forget to subscribe.
If you are watching this on the
511
00:33:03,400 --> 00:33:06,500
YouTube channel or if you are
listening to this on your
512
00:33:06,500 --> 00:33:09,900
favorite podcast platform, thank
you very much.
513
00:33:10,000 --> 00:33:12,000
And until we meet in another
episode,
00:00:06,100 --> 00:00:08,800
Hello, and welcome to a new
episode of the city of show with
2
00:00:08,800 --> 00:00:09,900
Matt.
Matt, my name is Matt.
3
00:00:09,900 --> 00:00:14,000
Matt and in each episode, our
exploring latest trends inside
4
00:00:14,000 --> 00:00:17,200
strategies from different
topics, like cybersecurity,
5
00:00:17,200 --> 00:00:21,200
digital transformation, emerging
Tech and also startups and
6
00:00:21,200 --> 00:00:24,400
Entrepreneurship.
And as, you know, sometimes I
7
00:00:24,400 --> 00:00:28,500
have as guest thought leaders,
innovators and entrepreneurs,
8
00:00:28,800 --> 00:00:32,700
who would like to hear from,
Them more about their areas of
9
00:00:32,700 --> 00:00:37,300
expertise to tell us about all
the Innovations and all the
10
00:00:37,300 --> 00:00:39,100
latest news of tech and
business.
11
00:00:39,400 --> 00:00:43,000
And today, I am pleased to have
with me Muhammad our chat, who
12
00:00:43,000 --> 00:00:47,900
is an expert in data science,
and artificial intelligence.
13
00:00:47,900 --> 00:00:50,300
Thank you so much.
Mohammed for joining me today.
14
00:00:50,700 --> 00:00:54,000
If you can, just introduce
yourself to us and what you do.
15
00:00:55,000 --> 00:00:57,800
Thank you very much for the
wonderful introduction.
16
00:00:58,100 --> 00:01:00,400
I'm Muhammad ershad.
I have been a day.
17
00:01:00,500 --> 00:01:04,500
A scientist for 17 years of my
life incorporate.
18
00:01:04,500 --> 00:01:09,300
I was working for companies like
Dell have let Packard Accenture
19
00:01:09,300 --> 00:01:14,200
module for team in Dubai.
And recently I started my own
20
00:01:14,200 --> 00:01:19,000
startup decoding data science
that is into Consulting and
21
00:01:19,000 --> 00:01:24,800
Academy and I had a passion of
you can say giving mentorship
22
00:01:24,800 --> 00:01:32,400
and P creating teams data teams
because there is A lot of Gap in
23
00:01:32,400 --> 00:01:36,000
the requirement for companies
and the education system.
24
00:01:36,000 --> 00:01:39,500
So this and this from that
particular need, this startup
25
00:01:39,700 --> 00:01:42,200
emerged.
Where in I do lot of you can
26
00:01:42,200 --> 00:01:45,900
say, awareness sessions.
What are the main Foundation
27
00:01:45,900 --> 00:01:50,200
skills that are required and all
these things for people to learn
28
00:01:50,200 --> 00:01:53,400
Ai and not get scared of it?
That's great.
29
00:01:53,700 --> 00:01:58,000
So, the first question, I want
to ask him how much because That
30
00:01:58,000 --> 00:02:01,600
will lead us to more questions
but let's start from the basics,
31
00:02:01,600 --> 00:02:05,400
right?
So we hear a lot of these words
32
00:02:05,500 --> 00:02:10,800
but sometimes people mix them up
so we hear about data as data,
33
00:02:10,800 --> 00:02:14,500
then data science data analytics
and then you know AI.
34
00:02:14,600 --> 00:02:19,600
So can you little bit destruct
this to us like how it all stars
35
00:02:19,600 --> 00:02:23,200
and then you know what are the
layers that come on top of it?
36
00:02:24,700 --> 00:02:27,800
Yes, sir.
So to be honest, my mat.
37
00:02:27,800 --> 00:02:31,800
So there is a this particular
definition about artificial
38
00:02:31,800 --> 00:02:35,700
intelligence data science
machine learning becomes very
39
00:02:35,700 --> 00:02:41,200
complicated and there is a lot
of you can say, confusion or
40
00:02:41,200 --> 00:02:45,800
there in the market.
So, to be honest, how I talk to
41
00:02:45,800 --> 00:02:48,900
the business, specifically, to
the business, or any people who
42
00:02:48,900 --> 00:02:54,600
are new to this and who want to
know is, Whatever Buzz words
43
00:02:54,600 --> 00:03:00,400
that are there just the root
word of everything is data.
44
00:03:00,900 --> 00:03:06,700
So how we can focus on the data
and generate insights from the
45
00:03:06,700 --> 00:03:12,400
data, that is the concept of AI.
Now, it depends upon lot of
46
00:03:12,400 --> 00:03:16,700
conditions, the business
outcomes that you want and what
47
00:03:16,700 --> 00:03:20,800
is the data that is available?
After that only, we can
48
00:03:21,000 --> 00:03:26,000
understand which part.
Of the AI ecosystem will be used
49
00:03:26,000 --> 00:03:30,600
in your to help solve your
business problem or to be used
50
00:03:30,600 --> 00:03:35,400
in your specific case.
So if we can just think it has a
51
00:03:35,500 --> 00:03:37,900
tool kit, right?
There are a lot of tool kits
52
00:03:37,900 --> 00:03:43,300
available in the AI ecosystem
and you as a business user, do
53
00:03:43,300 --> 00:03:45,100
not need to know so much
details.
54
00:03:45,500 --> 00:03:51,200
You just need a consultant or an
expert's help to decode your
55
00:03:51,200 --> 00:03:55,000
problem.
And I suggest you which is the
56
00:03:55,000 --> 00:03:59,800
best tools it for you to use
because most of the cases, most
57
00:03:59,800 --> 00:04:03,700
of the businesses they don't
need to use AI to be honest,
58
00:04:03,700 --> 00:04:05,700
right?
They can do with simple plain
59
00:04:05,700 --> 00:04:10,600
reporting, but now we don't want
to explain this and we don't
60
00:04:10,600 --> 00:04:12,800
want to create that confusion to
them.
61
00:04:13,300 --> 00:04:18,300
We keep everything in the AI
ecosystem, even the reporting
62
00:04:18,300 --> 00:04:21,399
part.
Now if I can talk in more
63
00:04:21,399 --> 00:04:24,700
details about this, so there are
There is a road map for any
64
00:04:24,700 --> 00:04:29,300
company to start with AI.
And the first is making sure
65
00:04:29,300 --> 00:04:33,400
that the data is getting
collected properly, the, the
66
00:04:34,000 --> 00:04:38,200
pipelines are fixed, and there
is a way where people or the
67
00:04:38,200 --> 00:04:42,500
company is collecting data, that
is the first step second step is
68
00:04:42,500 --> 00:04:45,800
on that data.
Basic reporting you need to
69
00:04:45,800 --> 00:04:49,700
start doing and creating
insights from the reports, the
70
00:04:49,700 --> 00:04:51,900
data analysis part from the
reports.
71
00:04:52,400 --> 00:04:55,000
After all these is Fix up
because most of the business
72
00:04:55,000 --> 00:04:59,400
questions that are, there can be
answered by just doing data
73
00:04:59,400 --> 00:05:02,100
analysis, on historical data of
the company.
74
00:05:02,400 --> 00:05:07,400
That is there after that, if it
is required, according to the
75
00:05:07,400 --> 00:05:12,200
business quotient, we can use.
You can say any machine learning
76
00:05:12,200 --> 00:05:15,200
or AI, whatever tool kit that is
there.
77
00:05:15,200 --> 00:05:20,400
So you don't need to business.
Don't need to do much lost focus
78
00:05:20,400 --> 00:05:23,300
and get confused by this,
because the more the details
79
00:05:23,400 --> 00:05:27,000
Tails, we talk, right?
As a data expert, the more
80
00:05:27,000 --> 00:05:31,900
people get afraid and they get
away from this, which is not the
81
00:05:31,900 --> 00:05:35,600
case here.
Thank you very much for this
82
00:05:35,600 --> 00:05:36,600
nation.
Mohammed.
83
00:05:36,600 --> 00:05:41,600
Now, you talked about business
outcomes and I love this
84
00:05:41,600 --> 00:05:44,200
approach, right?
So because at the end of the day
85
00:05:44,400 --> 00:05:48,100
and I loved also, the way you
mentioned, solving business
86
00:05:48,100 --> 00:05:52,300
cases.
Now, what are the major
87
00:05:52,300 --> 00:05:56,600
problems?
You think, you know, data
88
00:05:56,600 --> 00:06:03,600
science, and data analysis, AI
are today focusing to solve for
89
00:06:04,100 --> 00:06:08,000
organizations.
Yeah, that's right.
90
00:06:08,000 --> 00:06:10,500
Now, to be honest, the main
challenges is, I was telling you
91
00:06:10,500 --> 00:06:15,400
the first part making sure that
the data is getting collected in
92
00:06:15,400 --> 00:06:20,700
a proper way.
And what all systems are there
93
00:06:20,700 --> 00:06:27,500
in place for data to get create
generated and you can say saved
94
00:06:27,800 --> 00:06:32,000
because this is very important
for any company to try to
95
00:06:32,000 --> 00:06:37,500
capture, whatever data it is
possible because See if the data
96
00:06:37,500 --> 00:06:41,500
culture and the data collection
system is not in place, people
97
00:06:41,500 --> 00:06:48,100
are still using just Standalone
excels and doing analysis, stand
98
00:06:48,100 --> 00:06:50,900
alone, and there is no unique
data.
99
00:06:50,900 --> 00:06:54,100
Mart in the company, where in
everything is getting stored at
100
00:06:54,100 --> 00:06:58,700
a daily level, then the, there
is a big problem.
101
00:06:58,700 --> 00:07:02,000
You cannot do any analysis, you
cannot do any reporting on that.
102
00:07:02,000 --> 00:07:05,500
So this is a major challenge for
many companies, even if there
103
00:07:05,500 --> 00:07:09,400
are Same place they have systems
in place but they are not
104
00:07:09,400 --> 00:07:12,500
storing the data even if they
are storing the data, it is not
105
00:07:12,500 --> 00:07:16,000
at one particular space and that
one particular level like
106
00:07:16,000 --> 00:07:21,300
suppose customer data if we want
to have all the customer data,
107
00:07:21,400 --> 00:07:25,100
all the attributes of a customer
at one place, it is very tough
108
00:07:25,100 --> 00:07:27,800
to get, they will be marketing.
Team will have some data as
109
00:07:27,800 --> 00:07:29,500
separate Place.
Finance, team will have some
110
00:07:29,500 --> 00:07:32,300
data separate plays, HR will
have data separate Place,
111
00:07:32,700 --> 00:07:35,800
transaction data, will be
separate employee demographics,
112
00:07:35,900 --> 00:07:38,600
graphic data will be separate.
So those different different
113
00:07:38,600 --> 00:07:40,500
things.
We need to get it at one place
114
00:07:40,500 --> 00:07:45,700
and build a customer view, store
level view, product level View
115
00:07:45,900 --> 00:07:49,800
and that is, the thing is, if
people are not able to that is
116
00:07:49,800 --> 00:07:54,200
one of the main challenges that
I feel currently is to have, you
117
00:07:54,200 --> 00:07:58,200
know, there is a lot of data,
but there's a lot of noise, how
118
00:07:58,200 --> 00:08:03,500
to get the clean and the
meaningful data and store it in
119
00:08:03,500 --> 00:08:07,100
a data Mart.
So that you can Value out of it,
120
00:08:07,100 --> 00:08:10,100
which is the one of the main
challenges and one of the main
121
00:08:10,100 --> 00:08:13,900
pain points that companies and
businesses are facing.
122
00:08:14,600 --> 00:08:20,400
Yeah, that's very insightful.
Now, this bring to me, you know,
123
00:08:20,400 --> 00:08:24,300
this, this question because you
talked about how this confusion
124
00:08:24,300 --> 00:08:26,700
can happen.
So what do you think?
125
00:08:26,700 --> 00:08:31,900
You know, like are the most
successful implementations of
126
00:08:31,900 --> 00:08:35,400
this AI solution?
Like how can you give us some
127
00:08:35,400 --> 00:08:37,799
examples?
Apple's of without even
128
00:08:37,799 --> 00:08:39,400
mentioning like, maybe customer
names.
129
00:08:39,400 --> 00:08:42,500
But at least, you know how the
outcomes.
130
00:08:42,500 --> 00:08:46,900
We're after implementing the
right AI framework at a company.
131
00:08:49,000 --> 00:08:52,800
So you are asking for my
examples or the examples that
132
00:08:52,800 --> 00:08:56,300
I've seen industry, you seen in
the industry are so.
133
00:08:56,300 --> 00:09:00,500
So the best, the best thing, the
best case and the many one of us
134
00:09:00,500 --> 00:09:03,400
will relate with.
It is the recommendation engine,
135
00:09:03,400 --> 00:09:05,800
that is a recommendation
algorithms that are there,
136
00:09:05,800 --> 00:09:08,900
right?
So those algorithms they use lot
137
00:09:08,900 --> 00:09:12,300
of data, right?
Like, suppose I give you example
138
00:09:12,300 --> 00:09:15,500
of Amazon.
Yeah, the products that you are
139
00:09:15,500 --> 00:09:21,100
buying, so the products that you
buy The recommendation algorithm
140
00:09:21,100 --> 00:09:24,900
does not only work on product to
product match.
141
00:09:25,400 --> 00:09:29,900
Like it also works on the type
of customer Persona that you
142
00:09:29,900 --> 00:09:34,600
have, right?
Because you if Suppose there is
143
00:09:34,600 --> 00:09:38,500
one General product, if you buy
suppose you buy a printer you
144
00:09:38,500 --> 00:09:42,800
will also buy ketteridge.
The that is a recommendation
145
00:09:42,800 --> 00:09:46,500
that they will but this this
algorithm goes much further.
146
00:09:47,300 --> 00:09:49,400
Right.
It will be, will see the type of
147
00:09:49,400 --> 00:09:56,900
cursed customer Persona is
buying pattern, which brand he
148
00:09:56,900 --> 00:10:00,000
likes.
What is the most probable
149
00:10:00,000 --> 00:10:02,800
products that he will buy along
with this product?
150
00:10:03,100 --> 00:10:08,400
It can be historical baskets of
the same personas that are
151
00:10:08,400 --> 00:10:11,600
there.
So those are the very useful use
152
00:10:11,600 --> 00:10:15,900
cases that the companies are
using and specifically the
153
00:10:15,900 --> 00:10:18,900
e-commerce companies.
He's to give recommendation
154
00:10:18,900 --> 00:10:22,000
engines.
This is, this is one of the best
155
00:10:22,000 --> 00:10:28,800
use cases of AI + ml in action,
and this really causes lot of
156
00:10:28,800 --> 00:10:31,600
upsells right up says, cross
sales.
157
00:10:32,000 --> 00:10:36,400
And companies are really
improving this model every day.
158
00:10:36,800 --> 00:10:42,000
If they are getting insights
from the customer, whether he is
159
00:10:42,000 --> 00:10:45,000
buying the recommended products
or not buying the recommended
160
00:10:45,000 --> 00:10:49,100
products, and that is the Back
Loop goes into the algorithm and
161
00:10:49,100 --> 00:10:50,900
it make becomes better and
better.
162
00:10:52,400 --> 00:10:54,600
Very, very insightful.
I would say.
163
00:10:54,700 --> 00:10:58,700
Now, you mentioned something
about, you know, and do I see
164
00:10:58,700 --> 00:11:02,900
that you highlighted a lot
about, you know, collecting the
165
00:11:02,900 --> 00:11:05,500
data, right?
So you said, like this is, this
166
00:11:05,500 --> 00:11:08,400
is the number one thing that any
company would do.
167
00:11:08,600 --> 00:11:11,300
Now, mentioning this, what do
you think?
168
00:11:11,300 --> 00:11:16,400
Are, you know, the things that
companies should care about when
169
00:11:16,400 --> 00:11:20,000
collecting the data?
I mean, whether from a privacy
170
00:11:20,000 --> 00:11:24,100
perspective, whether it's like,
from Um, the, you know, data I
171
00:11:24,100 --> 00:11:26,900
would say, health itself, how
clean it is.
172
00:11:27,300 --> 00:11:31,300
So what are the major out say
points?
173
00:11:31,600 --> 00:11:35,600
That really they should look at
when they start collecting the
174
00:11:35,600 --> 00:11:39,700
data?
So see for collecting data
175
00:11:39,700 --> 00:11:43,100
specifically.
There are some specific laws of
176
00:11:43,100 --> 00:11:45,400
a particular country, right,
life for yui.
177
00:11:46,200 --> 00:11:53,200
The if you are collecting data
of yui preferred customers local
178
00:11:53,200 --> 00:11:57,000
or the people who are in yui,
the data should reside in UAE
179
00:11:57,000 --> 00:11:59,600
itself.
So, this is a rule that is
180
00:11:59,600 --> 00:12:02,800
there.
And so we know one thing is, you
181
00:12:02,800 --> 00:12:07,100
need to know the rule of the
land, the data rules, there are
182
00:12:07,700 --> 00:12:13,000
The in place and second thing is
regarding in yui, there is
183
00:12:13,400 --> 00:12:19,300
specifically there is no, you
can say role that we cannot use
184
00:12:19,900 --> 00:12:25,600
gender and race in models,
right?
185
00:12:25,600 --> 00:12:30,100
But in some countries in us and
gdpr on you Europe, there are
186
00:12:30,100 --> 00:12:35,800
some rules that you cannot use
those type of variables in the
187
00:12:35,800 --> 00:12:41,400
model because it becomes Diverse
it is against the diversity.
188
00:12:41,400 --> 00:12:45,100
Lodge logic.
But here there is no, such rule
189
00:12:45,100 --> 00:12:48,900
for that in the model.
But there is Rule regarding the
190
00:12:48,900 --> 00:12:53,600
sharing of personal information
that is there.
191
00:12:53,600 --> 00:12:59,800
So the rights of the data, they,
when they are giving it to do
192
00:12:59,800 --> 00:13:05,100
analysis, it should be.
There should be a masking logic
193
00:13:05,100 --> 00:13:09,500
that or the names.
And email IDs and contact
194
00:13:09,500 --> 00:13:12,600
numbers.
When you are sharing with there
195
00:13:12,600 --> 00:13:15,500
should be not, you should be
randomized.
196
00:13:16,000 --> 00:13:18,500
And there should be a separate
dictionary of all the person
197
00:13:18,500 --> 00:13:23,000
informations and the team.
Only very specific teams should
198
00:13:23,000 --> 00:13:28,000
have access to the personal
information only for targeting
199
00:13:28,000 --> 00:13:31,300
purposes and all those things.
But the general the other team
200
00:13:31,300 --> 00:13:35,500
members What doing the analysis
and other departments?
201
00:13:35,900 --> 00:13:39,600
The technology department should
not have access to the real
202
00:13:40,000 --> 00:13:44,300
information information and it
can only be linked towards a
203
00:13:44,300 --> 00:13:46,500
random ideas of.
These are some of the basic
204
00:13:47,800 --> 00:13:52,700
principles for protecting the
personal information and but the
205
00:13:52,700 --> 00:13:55,900
real usage of models.
The information on models for
206
00:13:55,900 --> 00:14:00,000
gender is, there is nothing like
there is no rule specifically in
207
00:14:00,000 --> 00:14:02,800
this region.
Okay, I see.
208
00:14:02,800 --> 00:14:06,200
So smelly, they need to care
about, you know, protecting the
209
00:14:06,200 --> 00:14:09,500
data, having the right access to
it, to the people who should
210
00:14:09,500 --> 00:14:11,700
have.
And of course, follow the
211
00:14:11,700 --> 00:14:15,300
regulation of whatever land this
algorithm will work.
212
00:14:15,500 --> 00:14:18,300
Now you mentioned something,
which triggered a, you know,
213
00:14:18,300 --> 00:14:21,600
something in my mind and it's,
you know, I like it actually
214
00:14:21,600 --> 00:14:24,900
because you said AI systems,
they are not for everyone,
215
00:14:24,900 --> 00:14:28,600
right?
So what are really the companies
216
00:14:28,600 --> 00:14:33,600
that they should focus on?
Having a I like, which verticals
217
00:14:33,600 --> 00:14:35,600
do you see?
This is more relevant in.
218
00:14:37,100 --> 00:14:43,100
So the see the red areas like
see first is the retail retail
219
00:14:43,400 --> 00:14:48,400
wherein you have lot of product
information, a lot of custom
220
00:14:48,400 --> 00:14:54,300
information that is one place
where you can do a lot of data
221
00:14:54,300 --> 00:14:57,200
analysis.
Second is marketing.
222
00:14:58,700 --> 00:15:02,200
Zoe all the digital marketing.
That happens, Performance
223
00:15:02,200 --> 00:15:07,000
Marketing, that happens.
They they have lot of data, and
224
00:15:07,000 --> 00:15:11,600
you can do lot of.
You can say, Roi calculations
225
00:15:11,700 --> 00:15:16,900
return of, in of Investments,
the customer lifetime value, the
226
00:15:16,900 --> 00:15:22,200
effectiveness of your media
spends, there are some models
227
00:15:22,300 --> 00:15:26,500
specific in the marketing phase
space from the historical data.
228
00:15:26,700 --> 00:15:30,700
You can make Make models.
Like, if you spend one dollar on
229
00:15:30,700 --> 00:15:36,500
marketing, how much will be the
on which channel you will spend,
230
00:15:37,300 --> 00:15:40,100
how much will be the increase in
sales?
231
00:15:40,100 --> 00:15:42,300
So there are, this is the
marketing mix models that are
232
00:15:42,300 --> 00:15:44,700
there.
This really helps for marketers
233
00:15:44,700 --> 00:15:48,500
to know which is the most
effective channel to put the
234
00:15:48,500 --> 00:15:51,400
money dollars in that particular
channel.
235
00:15:51,400 --> 00:15:54,400
So marketing it is being used
Healthcare.
236
00:15:54,500 --> 00:16:00,100
It is very much used because
there a lot of to be honest
237
00:16:00,100 --> 00:16:04,100
Healthcare is the oldest the
Farmer's Life, Sciences and
238
00:16:04,100 --> 00:16:08,800
Health Care is one of the oldest
verticals that have been using
239
00:16:08,800 --> 00:16:11,400
artificial intelligence much
before.
240
00:16:11,400 --> 00:16:14,900
It was the hype has happened,
they have been using for a long
241
00:16:14,900 --> 00:16:19,100
time because all the drugs that
we use, they have been tested
242
00:16:19,200 --> 00:16:24,100
and statistically only they have
been You can say, they have
243
00:16:24,100 --> 00:16:28,300
passed the test significance
level test to for it to be
244
00:16:28,300 --> 00:16:30,600
rolled out to the masses.
So it is there.
245
00:16:30,600 --> 00:16:33,200
Those are the fields, to be
honest everywhere where there is
246
00:16:33,200 --> 00:16:36,600
data.
There is there generated you
247
00:16:36,600 --> 00:16:39,300
start using?
As I told you, AI is just not
248
00:16:39,300 --> 00:16:43,300
machine learning a, I can be
also be simple data analysis,
249
00:16:43,300 --> 00:16:46,000
right?
Because that is the level you
250
00:16:46,000 --> 00:16:48,800
start at once.
You are doing data analysis,
251
00:16:48,800 --> 00:16:52,100
once you are doing all those
things, you have a good history
252
00:16:52,100 --> 00:16:56,400
of Atta a one year of data, then
you can go ahead with doing.
253
00:16:56,800 --> 00:17:03,300
You can say Predictive
Analytics, but to be honest,
254
00:17:03,300 --> 00:17:08,700
whenever I don't say, when I say
AI is not for everyone, it is
255
00:17:10,400 --> 00:17:13,099
the modeling part.
The modeling part is not for
256
00:17:13,099 --> 00:17:18,000
everyone.
AI is, as I told you earlier on
257
00:17:18,200 --> 00:17:24,099
solving, the problem with data
it can be a simple data
258
00:17:24,099 --> 00:17:27,300
analysis.
Also it is, it may not be
259
00:17:27,300 --> 00:17:31,300
According to some, it may not be
defined as AI, but you are using
260
00:17:31,300 --> 00:17:35,700
data historical data to do some
analysis and tell to the
261
00:17:35,700 --> 00:17:38,600
management or the business that
this reform the data.
262
00:17:38,600 --> 00:17:43,800
This is the story and this can
But right, so this is the way I
263
00:17:43,800 --> 00:17:47,800
don't really confuse.
When I say I is not used for
264
00:17:48,100 --> 00:17:50,500
many of the people.
It is the modeling part.
265
00:17:50,500 --> 00:17:54,800
The advanced machine learning
part, but data analysis, anyone
266
00:17:54,800 --> 00:17:58,300
who has even 30 days of data,
they can also do some data
267
00:17:58,300 --> 00:18:01,400
analysis and they can derive
some value out of it.
268
00:18:02,200 --> 00:18:05,000
Yeah, I think this is a clear
the things more.
269
00:18:05,200 --> 00:18:09,300
Thank you for you know, this
clarification and yeah, like I
270
00:18:09,308 --> 00:18:11,000
agree with you.
You like, you know, even
271
00:18:11,000 --> 00:18:13,600
whenever you have some
historical data, this is kind of
272
00:18:13,800 --> 00:18:16,300
analysis you can conduct.
And I think this is the one of
273
00:18:16,300 --> 00:18:20,300
the oldest we had, you
mentioned, you know, the example
274
00:18:20,300 --> 00:18:23,300
of these algorithms that do
suggestions for you, similar to
275
00:18:23,300 --> 00:18:27,400
what you see at Amazon or, you
know, like even by social media,
276
00:18:27,400 --> 00:18:29,400
they use it.
So when they suggest for you,
277
00:18:29,400 --> 00:18:31,300
some people to connect with and
so on.
278
00:18:31,500 --> 00:18:33,800
So yeah, thank you for this
clarification.
279
00:18:34,000 --> 00:18:38,800
Now, moving forward.
What are the major threads that
280
00:18:38,800 --> 00:18:40,600
you are seeing in this?
This field.
281
00:18:40,600 --> 00:18:44,200
Now, we are seeing like
generative AI, a lot of hype
282
00:18:44,200 --> 00:18:48,100
about GPT and all this stuff.
So I know that everyone is
283
00:18:48,100 --> 00:18:49,900
talking about it but other than
Chad G.
284
00:18:49,900 --> 00:18:53,900
PT and degenerative AI, what are
like the other things that now
285
00:18:53,900 --> 00:18:57,800
are trending?
I would say in this field, So
286
00:18:57,800 --> 00:19:02,600
the thing is Muhammad.
So one thing is that is really
287
00:19:02,600 --> 00:19:08,000
going to happen and he's already
started is the way we search the
288
00:19:08,000 --> 00:19:12,700
way we have been searching.
So, since internet has come in
289
00:19:12,700 --> 00:19:16,200
90s, right?
We had used Yahoo index page,
290
00:19:16,200 --> 00:19:21,300
index base search and now Google
also built upon it.
291
00:19:21,800 --> 00:19:25,600
They also via we're searching
for a keyword and we were
292
00:19:25,600 --> 00:19:29,600
getting 10 or 15 or we are
getting out of pages and it was
293
00:19:29,600 --> 00:19:32,400
getting sorted according to the
ranking and then we have to
294
00:19:32,400 --> 00:19:35,800
search, which are the values
that we are getting that has
295
00:19:35,800 --> 00:19:39,500
been a search for almost 20 25,
25 years, right?
296
00:19:39,500 --> 00:19:41,400
This is the way we have been
searching internet.
297
00:19:42,000 --> 00:19:45,200
Now the survey we are going to
search is going to change.
298
00:19:46,000 --> 00:19:50,700
We now we are going to ask tools
like Taj uppity already being
299
00:19:50,700 --> 00:19:54,700
has started the being AI, very
new type in what you want.
300
00:19:55,700 --> 00:19:58,700
Like us because most of the
searches are like when you're
301
00:19:58,700 --> 00:20:02,500
searching other than you, you
are trying to ask a question to
302
00:20:02,600 --> 00:20:05,600
Google and Google is going to
give you the pages.
303
00:20:06,100 --> 00:20:08,100
But now what is these tools
doing?
304
00:20:08,100 --> 00:20:10,400
It is going to give you one
answer.
305
00:20:10,600 --> 00:20:14,400
It is like going to be more of a
conversational way.
306
00:20:14,400 --> 00:20:16,600
I wear in.
This is going to your prom that
307
00:20:16,600 --> 00:20:22,500
you're going to give it is going
to search for the Values that
308
00:20:22,500 --> 00:20:25,600
are most similar to your prompt
and going to give you a
309
00:20:25,600 --> 00:20:29,500
one-on-one output and this is
going to be.
310
00:20:29,600 --> 00:20:32,100
You may like it.
You may not like it, but this is
311
00:20:32,100 --> 00:20:36,300
the way because most of the
searches are not too technical,
312
00:20:36,300 --> 00:20:38,800
right?
If you see most of the searches
313
00:20:38,800 --> 00:20:41,700
that and eight ninety percent of
the searches are just for
314
00:20:41,700 --> 00:20:46,200
information.
And if for that, these steps
315
00:20:46,200 --> 00:20:48,600
type of tools really work.
Well, and this is going to
316
00:20:48,600 --> 00:20:52,100
change the way we interact.
ER act, it is going to change.
317
00:20:52,100 --> 00:20:56,300
Like this is the, the prompts
that are there, like, even Arts.
318
00:20:56,700 --> 00:21:00,400
You give prompts an artist
getting generated you give from
319
00:21:00,400 --> 00:21:03,900
say, videos getting generated.
So now the content that is going
320
00:21:03,900 --> 00:21:06,200
to come in, the market is going
to be very much.
321
00:21:07,000 --> 00:21:10,800
You can say generated by Ai and
it is maybe it is a concern as
322
00:21:10,800 --> 00:21:12,000
well.
And it is also.
323
00:21:13,500 --> 00:21:15,500
You can say it is going to
change the way it is.
324
00:21:15,500 --> 00:21:19,400
Just going to be a change.
In the way these things we are
325
00:21:19,400 --> 00:21:23,500
going to.
Interact with the internet and
326
00:21:23,500 --> 00:21:26,900
it is this is going to grow more
and more and to be honest more
327
00:21:26,900 --> 00:21:29,900
than the positives there are a
lot of negatives are as well.
328
00:21:29,900 --> 00:21:33,800
There are a lot of negatives.
These things are going to
329
00:21:33,800 --> 00:21:38,800
increase the garbage that is
there already on the internet
330
00:21:38,800 --> 00:21:41,400
because AI is learning from
garbage?
331
00:21:41,400 --> 00:21:44,600
That is already there.
When you train your model on
332
00:21:44,600 --> 00:21:47,700
garbage, you will get garbage.
And again, the same garbage
333
00:21:47,700 --> 00:21:51,200
again going back.
So you're In the cycle is
334
00:21:51,300 --> 00:21:53,100
completely now.
What is happening all the
335
00:21:53,100 --> 00:21:56,000
content that is in Internet is
just getting repurposed
336
00:21:56,000 --> 00:21:57,900
repurposed and the quality is
going down.
337
00:21:58,300 --> 00:21:59,900
So those are the things that are
happening.
338
00:21:59,900 --> 00:22:05,200
Ethics is a big problem, ethics
discrimination bias. - all those
339
00:22:05,200 --> 00:22:06,800
things, you will get more and
more.
340
00:22:07,100 --> 00:22:10,700
And there are a lot of
challenges and awareness is very
341
00:22:10,700 --> 00:22:15,700
important nowadays upon People
Like Us on the usage and how you
342
00:22:15,700 --> 00:22:19,000
need to control these things and
how you need to be a better
343
00:22:19,000 --> 00:22:21,200
contributor, which I can.
Talk more about later.
344
00:22:22,200 --> 00:22:26,600
Yeah, that's informative and
just to add one thing and I
345
00:22:26,600 --> 00:22:30,800
always repeat this on my
podcast, the guys what you are
346
00:22:30,800 --> 00:22:34,500
seeing is based on data, which
was generated at some point at
347
00:22:34,500 --> 00:22:38,500
some stage by us.
And yes, do you know what the
348
00:22:38,500 --> 00:22:41,100
algorithm is doing is just
bringing back this.
349
00:22:41,100 --> 00:22:44,400
So thank you, Mom at 4-4 like
focusing on this.
350
00:22:44,700 --> 00:22:48,400
It's not inventing the wheel.
Yes, it looks to you that.
351
00:22:48,500 --> 00:22:52,700
It's a text written by By
someone, that's it might look
352
00:22:52,700 --> 00:22:56,000
unique but it didn't learn by
itself, right?
353
00:22:56,000 --> 00:23:00,600
So it was all content, which was
available to the model, and this
354
00:23:00,600 --> 00:23:02,500
is why it looks so real, I
believe, right?
355
00:23:02,500 --> 00:23:06,800
So, but and you mentioned about
like, you know, this, the
356
00:23:06,800 --> 00:23:09,300
future.
I'm really interested to to, to
357
00:23:09,300 --> 00:23:12,700
know, or like at least to know
your point of view about the
358
00:23:12,700 --> 00:23:16,600
future that you just mentioned
like how are you seeing things
359
00:23:16,600 --> 00:23:20,500
moving forward with that?
Right now.
360
00:23:20,500 --> 00:23:27,100
See, I was just talking today
morning that you calculators
361
00:23:27,100 --> 00:23:29,300
when it was invented.
There was lot of protests and
362
00:23:29,300 --> 00:23:33,400
lot of hype and lot of people
who are against calculator
363
00:23:33,400 --> 00:23:38,500
usage, initially for people who
want to learn as, and the
364
00:23:38,500 --> 00:23:42,800
similar thing is now happening
with Chad GPT like many, you can
365
00:23:42,800 --> 00:23:45,300
say professors.
And many teachers are protesting
366
00:23:45,300 --> 00:23:49,200
the use of this because what is
this going to?
367
00:23:49,400 --> 00:23:54,000
It is going to hamper the
learning because these tools are
368
00:23:54,100 --> 00:23:57,600
making us more and more lazy.
As we can see, Mammoth the
369
00:23:57,600 --> 00:24:02,500
future as you, the technology is
there to help us, but what it is
370
00:24:02,500 --> 00:24:06,600
actually doing is it is making
people lazy, right?
371
00:24:06,600 --> 00:24:11,200
Everything at your fingertips.
It is humans are not designed to
372
00:24:11,300 --> 00:24:15,400
be so lazy, but these tools are
making us lazy.
373
00:24:15,500 --> 00:24:20,200
And so what is happening, only
the, the learning Capability of
374
00:24:20,200 --> 00:24:24,200
the masses is going to reduce
automatically so they are always
375
00:24:24,200 --> 00:24:29,200
they are the consumers, 95% of
the people are just consumers
376
00:24:29,600 --> 00:24:33,100
and they are just not giving any
value.
377
00:24:33,800 --> 00:24:37,900
They are just using their just
the people, the rest, five
378
00:24:37,900 --> 00:24:42,200
percent of the people who are
actually learning from this and
379
00:24:42,200 --> 00:24:44,600
understanding this.
And they're actually the numbers
380
00:24:44,600 --> 00:24:47,900
are now reducing more and more
because of this the usage of
381
00:24:47,900 --> 00:24:53,100
this and the debate about this.
The creativity is going down and
382
00:24:53,100 --> 00:24:56,400
it is the thing is you see
machines are getting more and
383
00:24:56,400 --> 00:25:00,100
more powerful.
Now because of this there are
384
00:25:00,300 --> 00:25:02,000
there is going to be lot of job
losses.
385
00:25:02,000 --> 00:25:06,300
People are not talking about it
so much, but a job that was
386
00:25:06,300 --> 00:25:09,100
being done by two people.
Now can be done by one people,
387
00:25:09,100 --> 00:25:12,000
one person if he is using this,
these AI tools.
388
00:25:12,500 --> 00:25:17,800
So people the jobs are going to
reduce many of the you can say
389
00:25:17,800 --> 00:25:21,100
manual, autonomous jobs, Manual
jobs that are there, they can
390
00:25:21,100 --> 00:25:25,200
get automated.
So there is going to be a huge
391
00:25:25,200 --> 00:25:27,700
shift.
And one more thing is there.
392
00:25:27,700 --> 00:25:33,300
Now, now companies have to
understand how they can use,
393
00:25:33,400 --> 00:25:35,300
they cannot blindly use these
tools.
394
00:25:35,700 --> 00:25:40,100
They need some knowledge.
They need some need to draw a
395
00:25:40,100 --> 00:25:44,600
line and it is to be honest.
It is a very debatable topic.
396
00:25:44,800 --> 00:25:49,900
These there is a news
yesterday's news that Lot of
397
00:25:49,900 --> 00:25:52,300
thought leaders have written to
open a.
398
00:25:52,300 --> 00:25:58,500
I even including Elon Musk to do
a pause on the development of
399
00:25:58,500 --> 00:26:02,300
GPT, 4.5, or five after the
release of four.
400
00:26:02,600 --> 00:26:07,400
They weren't a pause because
they weren't a lot of, you can
401
00:26:07,400 --> 00:26:13,600
say, SOB pure Sops to be put in
place about the ethics about the
402
00:26:13,600 --> 00:26:19,300
standards right now, which is
not there to make it more and
403
00:26:19,400 --> 00:26:21,200
more clean.
Because there is a lot of
404
00:26:21,200 --> 00:26:24,000
garbage that is there in open AI
as well.
405
00:26:24,600 --> 00:26:27,600
And they need to do lot of
procedures according to the
406
00:26:27,608 --> 00:26:31,300
situation and make sure that
they are not giving harmful
407
00:26:31,300 --> 00:26:34,600
information.
Because to be honest, even kids
408
00:26:34,600 --> 00:26:39,600
are using taji PT, right?
So the text that it comes out
409
00:26:40,200 --> 00:26:45,900
has to be controlled in a way
because it is going to as we
410
00:26:45,900 --> 00:26:50,400
know, young minds, if they are
Given such information.
411
00:26:50,400 --> 00:26:52,900
It is not good for them as well.
So these are the things that
412
00:26:52,900 --> 00:26:54,300
they have asked them to do a
pause.
413
00:26:54,700 --> 00:26:58,400
And there is a lot of debate
online as well about the how,
414
00:26:58,400 --> 00:27:02,500
how to be ready with for this
particular wave because it is
415
00:27:02,700 --> 00:27:05,600
going to change a lot of things.
Yeah.
416
00:27:07,500 --> 00:27:11,900
Like for the first time maybe on
the podcast I will say something
417
00:27:11,900 --> 00:27:17,100
before I close while I
understand, you know, the
418
00:27:17,100 --> 00:27:20,800
concerns.
I've seen the news also this
419
00:27:20,800 --> 00:27:25,600
morning, where people like Ellen
mosque at Steve Wozniak.
420
00:27:25,600 --> 00:27:32,400
The Apple co-founder they signed
this petition to, to open a, i
421
00:27:32,400 --> 00:27:37,700
the company behind the chat GPT
my out, save my opinion here.
422
00:27:39,000 --> 00:27:44,600
As I can understand the
frustration, I can understand
423
00:27:45,200 --> 00:27:47,900
how people are afraid about all
this.
424
00:27:50,600 --> 00:27:56,500
But I think like, I'm not able
to understand the need for pose,
425
00:27:56,500 --> 00:28:00,100
right?
Because I always try to link to
426
00:28:00,100 --> 00:28:02,500
previous historical moment,
right?
427
00:28:02,500 --> 00:28:07,000
So when the internet came up,
you know all the regulations
428
00:28:07,000 --> 00:28:09,600
came after the internet became
more widespread, right?
429
00:28:09,600 --> 00:28:12,800
So because people wanted to
understand what's this internet
430
00:28:12,800 --> 00:28:17,200
is and then try to put the legal
stuff again.
431
00:28:17,200 --> 00:28:21,900
I'm not saying that.
I agree that, you know, open AI
432
00:28:21,900 --> 00:28:25,800
is the best thing or chat GPT.
Of course not and I agreed with
433
00:28:25,800 --> 00:28:28,200
you about.
It has garbage, I agree with
434
00:28:28,200 --> 00:28:34,400
that but the, the reality is and
I've seen some articles about
435
00:28:34,400 --> 00:28:37,800
this.
Any company today, not opening
436
00:28:37,800 --> 00:28:41,100
our people think that open area
is the only player, but it's not
437
00:28:41,400 --> 00:28:45,600
anyone today can collect the
data and this is part of what a
438
00:28:45,600 --> 00:28:48,700
data, scientists, they do.
I think Mom Would you not agree
439
00:28:48,700 --> 00:28:50,600
with me?
So there's something a process
440
00:28:50,600 --> 00:28:53,700
called Data Mining, right?
So, you have to find the data,
441
00:28:53,700 --> 00:28:58,700
you have to bring the data and
it's, it's a small world now.
442
00:28:58,700 --> 00:29:03,400
Like anyone and I mean by
anyone, yes, anyone if they have
443
00:29:03,400 --> 00:29:08,000
the tools, they can get the data
and they can, I'm expecting,
444
00:29:08,000 --> 00:29:11,400
maybe they are not getting the
attention, but I'm expecting to
445
00:29:11,400 --> 00:29:14,300
see another, you know, chat GPT
with the same.
446
00:29:14,300 --> 00:29:17,600
I mean efficiency and all those
and very near future.
447
00:29:17,600 --> 00:29:20,300
So if open, I will not do it,
someone else gonna do it.
448
00:29:20,300 --> 00:29:25,500
So this is my opinion, the other
thing, yes, a lot of use cases
449
00:29:25,500 --> 00:29:29,700
can be and that is a question
mark about it.
450
00:29:31,500 --> 00:29:34,800
But what I have seen, you know,
from both let's say Microsoft
451
00:29:34,800 --> 00:29:38,700
and Google with, you know, the
integration that they are
452
00:29:38,700 --> 00:29:43,300
planning to do with the
day-to-day tools like Microsoft
453
00:29:43,300 --> 00:29:47,400
Office, or on the Google
workspace on Google side.
454
00:29:48,200 --> 00:29:52,600
I think, you know, any any topic
that can touch the business from
455
00:29:52,600 --> 00:29:57,500
the following or humans.
If it can reduce the time to do
456
00:29:57,500 --> 00:30:02,000
it, if it can reduce the cost to
do it, if it can make them more
457
00:30:02,000 --> 00:30:05,200
productive, I can understand the
fears.
458
00:30:05,300 --> 00:30:08,900
But you know, like why the same
hype didn't come when the phone
459
00:30:08,900 --> 00:30:10,200
was there?
Because the phone?
460
00:30:10,200 --> 00:30:12,800
Also, if you think about the
smartphone, I mean the
461
00:30:12,800 --> 00:30:17,600
smartphone moment.
It also made people kind of, you
462
00:30:17,600 --> 00:30:19,600
know, like the They get
attracted to it.
463
00:30:19,800 --> 00:30:22,900
Same for the social media.
I mean, whether it's Facebook,
464
00:30:22,900 --> 00:30:25,200
Twitter, whatever it is
Tick-Tock, right?
465
00:30:26,700 --> 00:30:29,600
I see what you are trying to
say.
466
00:30:29,700 --> 00:30:32,100
And I agree with you, by the
way, like I'm not fully
467
00:30:32,200 --> 00:30:37,500
disagreeing, but just to put my
two cents, I understand the
468
00:30:37,500 --> 00:30:40,200
concern.
But I would say it's too late to
469
00:30:40,200 --> 00:30:44,900
stop it and if not open a, if
openly I stopped today releasing
470
00:30:45,100 --> 00:30:47,900
GPT five or whatever, believe
me.
471
00:30:48,500 --> 00:30:52,100
Someone else can release that or
something even better.
472
00:30:52,100 --> 00:30:54,100
I think it's too late to stop
the train.
473
00:30:54,100 --> 00:30:59,300
It's at left the station.
I say it left the station, but
474
00:30:59,300 --> 00:31:04,000
of course, the effort should be
about how to force these
475
00:31:04,000 --> 00:31:07,800
companies like open a.i. to
clean the data to remove the
476
00:31:07,800 --> 00:31:10,600
garbage as you called it.
A hundred percent agree that
477
00:31:10,600 --> 00:31:14,800
it's there before we close any
final thoughts from your side
478
00:31:14,800 --> 00:31:16,700
Muhammad because we've come to
an end.
479
00:31:17,800 --> 00:31:18,800
You know, I know.
Yeah.
480
00:31:18,800 --> 00:31:22,500
So this to be honest I can talk
hours and hours about this topic
481
00:31:23,100 --> 00:31:26,900
and we can surely, I will just
summarize.
482
00:31:26,900 --> 00:31:30,400
So yeah, there are a lot of
challenges because was just as a
483
00:31:30,400 --> 00:31:37,000
takeaway, is somebody try to be
open about this change.
484
00:31:37,000 --> 00:31:44,000
So you need to prevent to even
to learn about the sea is Like a
485
00:31:44,000 --> 00:31:47,400
Knife, right?
It's a, the you can use it.
486
00:31:47,600 --> 00:31:50,700
For the wrong things.
Also, you can use it for cooking
487
00:31:50,700 --> 00:31:53,300
and cooking as well.
So you but you need to learn
488
00:31:53,400 --> 00:31:56,500
about knife.
So that is you need to learn the
489
00:31:56,500 --> 00:31:59,400
good things, the bad thing, so
how to use it when to use it.
490
00:31:59,700 --> 00:32:02,700
So think it like a tool and that
is the thing that I have tell
491
00:32:02,700 --> 00:32:05,900
everyone there is going to be
debate, there is going to be
492
00:32:05,900 --> 00:32:08,900
some misunderstanding, but at
the end of the day, it is just a
493
00:32:08,900 --> 00:32:13,600
tool to help humans.
So that is the my final thoughts
494
00:32:13,600 --> 00:32:16,000
about this and it was very nice
talking to you.
495
00:32:17,100 --> 00:32:18,500
Thank you.
Oh, thank you very much for your
496
00:32:18,500 --> 00:32:20,900
time today.
And again, I have to agree with
497
00:32:20,900 --> 00:32:24,000
the last sentence guys, train
yourself.
498
00:32:24,000 --> 00:32:26,200
Understand what is this tool
about?
499
00:32:26,200 --> 00:32:29,800
It's not what you see on Tik-Tok
or Instagram whatever because I
500
00:32:29,800 --> 00:32:33,100
start to see weird things there.
So there is a science, there is
501
00:32:33,100 --> 00:32:36,400
a science behind it, understand
the science and believe me the
502
00:32:36,400 --> 00:32:37,700
moment you understand the
science.
503
00:32:37,700 --> 00:32:40,400
You will see the benefit of it.
Thank you very much Muhammad for
504
00:32:40,400 --> 00:32:44,800
joining me today as usual, I'm
saying this to my audience.
505
00:32:44,800 --> 00:32:47,800
If you have any questions, if
you have any feedback about This
506
00:32:47,800 --> 00:32:51,200
episode or about the show.
If you want to be on the show,
507
00:32:51,200 --> 00:32:55,400
same as Muhammad, he was my
guest today, please feel free to
508
00:32:55,400 --> 00:32:57,800
contact me.
Whether by email LinkedIn or
509
00:32:57,800 --> 00:33:01,100
Twitter, I would be more than
happy to talk to you and don't
510
00:33:01,100 --> 00:33:03,400
forget to subscribe.
If you are watching this on the
511
00:33:03,400 --> 00:33:06,500
YouTube channel or if you are
listening to this on your
512
00:33:06,500 --> 00:33:09,900
favorite podcast platform, thank
you very much.
513
00:33:10,000 --> 00:33:12,000
And until we meet in another
episode,
