scripod.com

Ep 91: Top AI Analyst Unpacks Today's AI Hype Cycle

Shownote

Benedict Evans, one of tech's most widely-read analysts, joins Jacob Effron. The conversation centers on Benedict's core thesis that comparing AI's scale to past platform shifts (the internet, mobile, PCs) is analytically useless, and that the more product...

Highlights

In this episode of Unsupervised Learning, Benedict Evans discusses the economic and structural implications of AI, arguing that comparing it to past platform shifts like the internet or mobile is analytically unhelpful. He focuses on how value will accrue, the jagged nature of AI capabilities, and the uncertainty surrounding its physical limits.
00:00
Where will value accrue in AI?
04:03
Historical analogies offer insights but not predictive power.
10:17
It's amazing but doesn't fully work yet.
25:05
Defining AGI as catastrophic doesn't prove it will happen.
30:24
Foundation model labs are like TSMC, not Windows.
33:56
AI will increase software volume by making development cheaper
42:18
Complex tasks require significant effort and are not commodity
53:06
People imitate forms without understanding the underlying cause.
57:30
Novel behaviors require entrepreneurial invention, not just better technology.
1:01:28
Consumer use cases must be invented, not spontaneous.
1:03:48
Sora lacks a sustained use case
1:06:27
AI is an enabling technology, not just text and pictures

Chapters

Intro
00:00
Is AI Bigger Than the Internet?
01:31
Barriers of Getting From Demos to Daily Use
10:10
Why Job Predictions Fail
20:15
Where's the Moat?
25:52
Will Models Eat the App Layer?
33:55
When Average Isn't Enough and Models Don't Work
39:25
Reflections on OpenAI
45:58
Consumer Usage Is Still Shallow
55:04
What's Required for More Enterprise Adoption
58:51
Opinion on Sora
1:03:47
Quickfire
1:06:27

Transcript

Benedict Evans: You can wave your hands and say, no, this is like electricity, Okay, fine, it's like electricity. Well, what happened with electricity? And there was a time before electricity, and people before electricity weren't dumb either. Jacob Effro...