Podcast Simon Hodgkins article

Alex: 00:00:00

Hey everyone, before we dive in, I want to mention that our discussion

2

: 00:00:03

today is inspired by an excellent article

written by Simon Hodgkins, Chief Marketing

3

: 00:00:08

Officer at Vistatec, Founder of the

Think Global Forum, based in Ireland.

4

: 00:00:14

Elizabeth: Absolutely, Simon's

insights really set the stage

5

: 00:00:17

for a great conversation.

6

: 00:00:18

Let's get into it!

7

: 00:00:19

Alex: Hey, have you noticed how

there's been an AI boom recently?

8

: 00:00:23

Companies like OpenAI are

splurging billions, but they're

9

: 00:00:26

still not making a profit.

10

: 00:00:28

Kind of surprising, right?

11

: 00:00:30

Elizabeth: Absolutely, it's like

even the big players like Anthropic

12

: 00:00:34

are grappling with the same issue.

13

: 00:00:36

Companies downstream, like Anysphere

with Cursor, face challenges too.

14

: 00:00:41

Pricing struggles in a high-demand

industry seem a bit unusual.

15

: 00:00:45

Alex: Exactly, and Ed Zitron even

calls it the "subprime AI crisis."

16

: 00:00:50

What do you think about that metaphor

comparing it to the: 2008

17

: 00:00:55

Elizabeth: It's a thought-provoking

analogy, isn't it?

18

: 00:00:57

It suggests we've invested too

heavily in an idea we don't fully

19

: 00:01:00

grasp, assuming it can only succeed.

20

: 00:01:03

Alex: Yeah, it's a bit concerning.

21

: 00:01:05

But there's more to it.

22

: 00:01:07

We're not really facing a

collapse but rather dealing with

23

: 00:01:09

misalignment, overhype, and the

tech's awkward adolescence phase.

24

: 00:01:15

Elizabeth: Speaking of numbers, OpenAI is

reportedly facing a $5 billion loss this

25

: 00:01:19

year despite raising over $60 billion.

26

: 00:01:22

Seems like a lot, right?

27

: 00:01:24

Alex: For sure!

28

: 00:01:25

People don't realize each

ChatGPT query costs money.

29

: 00:01:28

Even with GPT-4o optimizing

things, the cost of training

30

: 00:01:31

and maintenance is sky-high.

31

: 00:01:34

Elizabeth: That's true, and

the cost cascades to other

32

: 00:01:36

companies using OpenAI's API.

33

: 00:01:38

It's like a food chain, and when

expenses go up, everyone feels the pinch.

34

: 00:01:42

Remember what happened with Cursor?

35

: 00:01:45

Alex: Oh, definitely.

36

: 00:01:46

They had to hike their prices drastically

after their significant funding round, and

37

: 00:01:51

users were definitely not happy about it.

38

: 00:01:54

Elizabeth: Before we jump to conclusions

about an impending AI crash, we need

39

: 00:01:58

to ask what exactly we're measuring.

40

: 00:02:00

High valuations don't

guarantee immediate value.

41

: 00:02:03

Alex: AI is still finding its way,

much like YouTube did back in: 2006

42

: 00:02:08

The real gains right now seem

to go to infrastructure giants

43

: 00:02:11

like Microsoft and Nvidia.

44

: 00:02:14

Elizabeth: There's merit in viewing

this as a business model misfire

45

: 00:02:17

rather than a classic bubble.

46

: 00:02:19

Many assumptions about value

capture are proving premature.

47

: 00:02:23

Alex: I get that.

48

: 00:02:24

And here's the uncomfortable truth:

Is AI genuinely useful at scale, or

49

: 00:02:28

are we getting ahead of ourselves?

50

: 00:02:30

What's your take?

51

: 00:02:31

Elizabeth: Well, I think AI's potential is

huge — people need to learn to 'drive' it.

52

: 00:02:36

There will be experts and

consumers, and where expertise

53

: 00:02:39

meets AI, that's where value lies.

54

: 00:02:42

Alex: Agreed.

55

: 00:02:43

It's still serving curiosity more

than necessity right now, mostly

56

: 00:02:47

aiding and creatively assisting

rather than being mission-critical.

57

: 00:02:51

Elizabeth: That's valid.

58

: 00:02:52

But if AI doesn't prove commercially

viable quickly, the financial strain

59

: 00:02:56

could become a significant issue.

60

: 00:02:58

Alex: And about alignment — there's

a mismatch between AI companies'

61

: 00:03:01

goals, like OpenAI's focus on

AGI, and their customers' needs.

62

: 00:03:06

Seems like an odd place to be

if you're eyeing consistency.

63

: 00:03:09

Elizabeth: So, what are

the potential scenarios?

64

: 00:03:12

We're already seeing signs

of a mini "AI winter."

65

: 00:03:15

What's next, Jamie?

66

: 00:03:16

Alex: We might see a shake-up among

startups dependent on LLM APIs.

67

: 00:03:20

However, a killer AI-native application

could emerge, transforming the landscape.

68

: 00:03:26

Elizabeth: Or open-source tech might

eventually outperform the proprietary

69

: 00:03:30

models, cutting dependency and costs.

70

: 00:03:33

That could shift the

balance significantly.

71

: 00:03:36

Alex: Agreed.

72

: 00:03:36

Alternatively, AI could seamlessly

integrate into existing services, making

73

: 00:03:41

it indispensable but behind the scenes.

74

: 00:03:44

Elizabeth: So, is AI really facing

a "subprime moment," or is this

75

: 00:03:48

just another step in its evolution?

76

: 00:03:50

Alex: Zitron may be onto something.

77

: 00:03:52

We've likely overestimated

short-term returns, but this

78

: 00:03:55

risky investment fuels innovation.

79

: 00:03:58

Elizabeth: Exactly.

80

: 00:03:59

Like the early internet days,

it may take time to monetize.

81

: 00:04:02

But AI isn't going away — it's

merely refining its utility.

82

: 00:04:08

Alex: Indeed, Alex.

83

: 00:04:09

It's a matter of figuring out

practical applications rather

84

: 00:04:12

than chasing lofty dreams.

85

: 00:04:14

Only time will tell where AI's heading.

86

: 00:04:17

Simon's insights were spot-on

and really got us thinking about

87

: 00:04:21

these challenges in a new way.

88

: 00:04:23

Thanks everyone for joining us, and

special thanks to Simon Hodgkins,

89

: 00:04:27

Chief Marketing Officer at Vistatec,

Founder of the Think Global Forum,

90

: 00:04:32

based in Ireland, for the thoughtful

piece that inspired today's discussion.