HomeVideos

Why Jane Street Makes More From Claude Than Anthropic Does - Dylan Patel

Now Playing

Why Jane Street Makes More From Claude Than Anthropic Does - Dylan Patel

Transcript

44 segments

0:00

What if the AI models were literally as

0:01

good as a fully automated software

0:04

engineer? They're not currently there

0:05

yet, right? Like I think they're far

0:06

from just being able to fully automate

0:07

the job of like a full white-collar

0:09

worker. But white-collar workers earn,

0:11

you know, six figures or north of that a

0:12

year. And if you have a gigawatt that

0:15

can sustain a population of like, say, a

0:17

million white-collar workers, then off

0:19

the back of that you'd be many hundreds

0:20

of billions of dollars if you get like

0:21

full AGI per gigawatt.

0:22

>> I think the other aspect of this is most

0:24

of the value that these models generate

0:25

does not get given to Open AI and

0:27

Anthropic. Thankfully so far, it is

0:28

mostly just being given to the users.

0:30

>> Yeah, yeah.

0:30

>> Right? Jane Street, where they're like

0:31

one of Anthropic's biggest customers, is

0:33

generating way, way, way, way more value

0:36

out of the tokens they're paying for

0:38

than Anthropic is

0:40

uh generating in terms of profit. Or

0:41

Meta, who at one point was rumored to

0:44

be, you [music] know, as much as 10% of

0:45

Anthropic's business, they're generating

0:47

way more efficiencies by optimizing

0:49

their ad algorithms or what have you and

0:52

and getting engagement time 5% longer

0:54

and, you know, all these things. They're

0:56

making way more money off of using these

0:58

models than Anthropic. And so the

0:59

ultimately Sure, if you had a million

1:01

new software engineers, the cost per

1:03

software engineer would also fall. It

1:04

wouldn't be a pure

1:05

>> If you only had a software engineers and

1:07

not other type types of labor

1:08

>> but that's every kind of laborer, right?

1:10

That's the that's the whole point of AGI

1:12

is it collapses the value of labor.

Interactive Summary

The video discusses the potential economic impact of AGI, suggesting that while AI is not yet a full replacement for human workers, its ability to automate tasks could lead to massive value creation. The participants emphasize that the primary beneficiaries of AI models are currently the users—such as large companies—who generate significantly more value through efficiency gains than the AI providers make in profit, ultimately leading to a potential collapse in the value of human labor.

Suggested questions

3 ready-made prompts