Listen to the latest episode of Data Career Podcast: Helping You Land a Data Analyst Job FAST

Avery: 00:00:00

Will AI replace me as a data analyst.

2

: 00:00:04

That is a question I've been

getting a lot of recently.

3

: 00:00:07

And in today's episode,

I wanted to dig into.

4

: 00:00:10

If I actually think that's

going to be the case.

5

: 00:00:12

Once again, guys, welcome back to the data

career podcast, the number one podcast

6

: 00:00:16

for your landing, your first data job.

7

: 00:00:19

I'm your host, Avery Smith.

8

: 00:00:21

And I run a company called data

career jumpstart where I help

9

: 00:00:23

people land their first data job.

10

: 00:00:26

So let's talk about this whole AI,

maybe taking the place of data analysts.

11

: 00:00:31

Is it going to happen?

12

: 00:00:32

Has it already happened?

13

: 00:00:34

Let's get into some thoughts.

14

: 00:00:46

All right.

15

: 00:00:47

So AI is not coming.

16

: 00:00:48

It is here.

17

: 00:00:49

You guys with revolutionary new products

like ChatgPT GPT4 MidJourney It's

18

: 00:00:56

very clear that AI is all over the

place, but what does that mean to you

19

: 00:01:00

as someone who is in the data world

or trying to land their first data

20

: 00:01:03

job, are you going to be replaced?

21

: 00:01:06

For a career you're trying so hard to work

into and you can't even break into yet.

22

: 00:01:09

Is it going to be obsolete?

23

: 00:01:11

I actually got a message on LinkedIn this

week that asked something very similar.

24

: 00:01:16

They said that they've been hearing more

and more about AI and ChatGPT 4 I came

25

: 00:01:20

across this newsletter the other day that

said that data analyst might be replaced.

26

: 00:01:24

And I begin to wonder about future job

security for junior data analysts like me.

27

: 00:01:29

Simply because this is moving so quickly.

28

: 00:01:31

I just wonder what your

thoughts were on this.

29

: 00:01:33

So I thought I'd make this podcast

episode to kind of talk through

30

: 00:01:35

what I think about ChatGPT Uh, LLMs.

31

: 00:01:41

AI, everything in general.

32

: 00:01:43

Is it coming for you?

33

: 00:01:44

My answer is no.

34

: 00:01:46

My answer is no.

35

: 00:01:47

And the reason is, is AI is just a tool.

36

: 00:01:52

You know, we really think

of AI is really scary.

37

: 00:01:54

Like these robots that could

potentially take over the world.

38

: 00:01:57

And, you know, maybe it will

be that eventually, but right

39

: 00:01:59

now all AI is, is a hammer.

40

: 00:02:01

It still needs the human

to actually do something.

41

: 00:02:05

Now there are really cool

things like auto GBT.

42

: 00:02:08

If you've never heard of that before.

43

: 00:02:09

Pause the podcast, you can look it up.

44

: 00:02:11

I'm not the most well versed on it.

45

: 00:02:13

So I'm not going to try to explain it in

depth in this episode, but basically it's

46

: 00:02:17

like where you have multiple AI agents.

47

: 00:02:20

So imagine chat GPT talking

to another chat CPT, which is

48

: 00:02:24

talking to another chat GPT.

49

: 00:02:26

And it can basically, instead

of having a human talk to

50

: 00:02:29

GBT it's chatty PT, talking to Chatsworth.

51

: 00:02:31

GPT, and it becomes a

little bit more autonomous.

52

: 00:02:33

I think there's another one

called auto GPT, um, as well.

53

: 00:02:37

And these are basically talking

back and forth to one another,

54

: 00:02:40

a little bit more autonomously.

55

: 00:02:41

So it is like maybe becoming more of.

56

: 00:02:44

A self propelling tool.

57

: 00:02:47

But for the most part, AI is just a tool.

58

: 00:02:49

It's just like a hammer.

59

: 00:02:51

It's like a saw you still

need a human to operate it.

60

: 00:02:54

So will our tasks change as data

analysts and data practitioners?

61

: 00:02:59

Yes, probably as they always have.

62

: 00:03:02

I mean, imagine the data analysts before

computers were invented now, I don't know

63

: 00:03:07

if that was, if they were really a thing

back then, probably not, but imagine

64

: 00:03:11

what they would be doing would be a lot

different than what we'd be doing today.

65

: 00:03:14

You know, for example, they'd be

doing a lot of hand calculations.

66

: 00:03:17

Maybe they'd be using that.

67

: 00:03:18

Like, what's that weird

invention, like the Abacus, right.

68

: 00:03:21

To keep track of all

these different things.

69

: 00:03:23

And now we have Excel.

70

: 00:03:24

Now we have SQL Those are

tools that did not exist.

71

: 00:03:28

Did you know, these people who

these mathematicians did, they

72

: 00:03:32

lose their job when Excel came out?

73

: 00:03:34

No.

74

: 00:03:35

Did their job nature change?

75

: 00:03:37

Yes, probably.

76

: 00:03:38

They probably started to

having to use the tool.

77

: 00:03:41

So in my opinion, this is

just a hammer and everyone

78

: 00:03:43

should start using the hammer.

79

: 00:03:45

There's no reason whether you're trying to

land your first data job or whether you're

80

: 00:03:49

an experienced data professional that

you should not be using some form of AI.

81

: 00:03:53

You need to be otherwise, they're

going to get left with the times.

82

: 00:03:57

This is a new tool.

83

: 00:03:57

It's going to be revolutionary

and you should feel enabled.

84

: 00:04:01

In fact, you should feel emboldened

by this because all of a sudden, the

85

: 00:04:04

barrier to break into data analytics

has even dropped a little bit further.

86

: 00:04:09

Because a lot of the

technical requirements can be

87

: 00:04:13

fulfilled with CQL or sorry.

88

: 00:04:16

Can be fulfilled with chat, GPT and AI.

89

: 00:04:20

And what I mean by that is.

90

: 00:04:22

It's basically all, all chatty

bitchy and AI really is, is a

91

: 00:04:26

more effective Google, right?

92

: 00:04:28

Like in the past, if you didn't

know how to do something in

93

: 00:04:30

CQL, you could Google it.

94

: 00:04:31

And within a few clicks, you

could probably find your answer.

95

: 00:04:34

Really I'll chat.

96

: 00:04:35

GPT is, is a more effective Google.

97

: 00:04:39

And I'm not even going

to say more effective.

98

: 00:04:40

I'm going to say faster, but it's not

necessarily guaranteed to be right.

99

: 00:04:44

A hundred percent of the time.

100

: 00:04:45

So you do have to be a little bit

concerned and worried about that.

101

: 00:04:48

But really we've been

doing this for years.

102

: 00:04:51

We've been posting stuff and questions

on stack overflow and getting help.

103

: 00:04:55

We've been Googling stuff,

finding different EHRs and finding

104

: 00:04:58

the answer inside of Google.

105

: 00:05:00

So really what I see chat GBT

is coming into play is it's just

106

: 00:05:04

going to make us more effective.

107

: 00:05:05

It's a tool that's going

to make us more effective.

108

: 00:05:08

And for you, as someone who wants

to break into the data world, you

109

: 00:05:10

should feel really excited because

now it's going to be a little bit

110

: 00:05:13

easier to break into the field.

111

: 00:05:15

You don't have to be memorizing all SQL.

112

: 00:05:17

You don't have to, you know, memorize

all this different statistics

113

: 00:05:20

stuff you could ask Chachi.

114

: 00:05:21

GPT and pretty reliably get an answer.

115

: 00:05:24

Right, but that is a little bit scary.

116

: 00:05:25

Cause every thinking, well,

that makes me obsolete.

117

: 00:05:27

But once again, a tool has

to be wielded by a human.

118

: 00:05:30

And humans are really good at

knowing when we should do tasks.

119

: 00:05:36

W how we should do this task

and what those tasks should be.

120

: 00:05:38

Now we can use AI.

121

: 00:05:40

To kind of expedite that whole

journey, the whole process.

122

: 00:05:44

But it still requires a good human

brain to link the business to the AI.

123

: 00:05:49

It will not be the case in the future.

124

: 00:05:50

I think so.

125

: 00:05:51

We'll, we'll chat GBT and AI get

better and need humans less probably.

126

: 00:05:57

But until that happens, let's not fret.

127

: 00:05:59

What's going to happen at

midnight when it's only 9:00 AM.

128

: 00:06:02

Right.

129

: 00:06:02

And we can get good at using AI.

130

: 00:06:05

We can get good at using chatty PT today.

131

: 00:06:07

And then we're going to have job security.

132

: 00:06:10

All right, we're going

to have job security.

133

: 00:06:12

So AI will not take your job,

but a human that uses AI might.

134

: 00:06:17

And so to be future-proof all

you need to be is a human that

135

: 00:06:20

knows how to use AI really well.

136

: 00:06:23

And what does that mean for you?

137

: 00:06:24

Like what can you do

today to get good at that?

138

: 00:06:26

I think you just go onto chat

GBT and start playing around.

139

: 00:06:30

Start trying to ask different questions,

you know, try to ask different SQL

140

: 00:06:34

questions or different Python questions,

or maybe even some Excel questions.

141

: 00:06:37

And just start messing around.

142

: 00:06:39

All right.

143

: 00:06:39

You can kind of get a feel for

what it can do and what it can't

144

: 00:06:42

do, what some limitations are,

but don't be worried about it.

145

: 00:06:45

If you have never opened it up

before, if you've never tried

146

: 00:06:47

doing data analytics on it.

147

: 00:06:48

I don't don't fret.

148

: 00:06:50

There's some tasks that Chatsy

beauty is fantastic for.

149

: 00:06:54

And there's other tasks that Chatswood

beauty is actually not very good at, for

150

: 00:06:57

example, if you S chat GPT to add, like,

let's just say 3 million, 246,392 plus 7

151

: 00:07:08

million, 564,123 or something like that.

152

: 00:07:15

Jeff.

153

: 00:07:15

It's a simple math problem, right?

154

: 00:07:16

Computers are really good at that.

155

: 00:07:17

Judgey beauty will most

likely get that problem wrong.

156

: 00:07:20

And the reason is.

157

: 00:07:22

It doesn't know how to do math.

158

: 00:07:24

It just knows how to read

what math has been done.

159

: 00:07:28

And so it's usually when, if you, if you

were to put that in chat to PT right now,

160

: 00:07:32

they'll probably get the first numbers.

161

: 00:07:33

Right.

162

: 00:07:33

And the last numbers.

163

: 00:07:34

Right.

164

: 00:07:35

But like the hundred thousands

will probably be all jumbled up

165

: 00:07:37

because chatty BT has used written

math to understand when these

166

: 00:07:42

numbers are added to these numbers.

167

: 00:07:45

You know, then we should probably start

with these numbers and we should probably

168

: 00:07:48

end with these numbers, but it's actually

not that good at doing the middle section.

169

: 00:07:52

So for instance, it's obviously

not going to take the job in any

170

: 00:07:56

sort of large number arithmetic.

171

: 00:07:58

Now, computers are better

at humans than adding.

172

: 00:08:01

If we can tell the computer what we need,

but that brings me to my next point.

173

: 00:08:04

Is that the bridge between the tech and

the domain or the tech in the business?

174

: 00:08:10

Is still going to be

largely covered by humans.

175

: 00:08:13

Humans are what.

176

: 00:08:15

Humans are more capable to see what's

going on in the real world and relate

177

: 00:08:20

it to math related to statistics

related to data and make those

178

: 00:08:24

connections and also make those choices.

179

: 00:08:26

Right, because at the end of the

day, humans, most of society is

180

: 00:08:30

still a human decision made society.

181

: 00:08:32

We're making business decisions.

182

: 00:08:35

You know, even when I was at

Exxon mobile, when we had AI

183

: 00:08:38

tools, they weren't really AI.

184

: 00:08:40

They were very data, sciency

machine learning tools.

185

: 00:08:42

It would suggest stuff, but

it's still down to the human to

186

: 00:08:45

make those decisions overall.

187

: 00:08:47

Because once again, it's a tool it's

not going to replace us and less.

188

: 00:08:51

We don't use it at all.

189

: 00:08:52

So I encourage all of you guys to go

play with GPT, go test for yourself.

190

: 00:08:57

Go try it out.

191

: 00:08:58

Don't be afraid of it.

192

: 00:09:00

It's going to be okay.

193

: 00:09:01

Your job's not going anywhere

and life's going to continue on.

194

: 00:09:05

That being said, if you want help

in your data journey landing,

195

: 00:09:08

that first data job hit me up.

196

: 00:09:10

I run a program called

data analytics accelerator.

197

: 00:09:13

It's a 10 week bootcamp that will

help you land your first data job.

198

: 00:09:16

Uh, by helping you build nine

portfolio projects, learning the

199

: 00:09:19

most important skills and teaching

you how to network effectively with

200

: 00:09:23

recruiters, hiring managers and peers.

201

: 00:09:25

So that you can land that data job fast.

202

: 00:09:27

If you're interested in that you can

click on the link down below once again.

203

: 00:09:30

Thank you guys for listening to the show.

204

: 00:09:31

See you soon.

Share Episode

Shownotes

Transcripts

Follow

Links

Chapters

Video

More from YouTube