HomeVideos

David Sacks Predicts the Regulatory Capture Playbook to Ban Open Source AI, Step by Step

Now Playing

David Sacks Predicts the Regulatory Capture Playbook to Ban Open Source AI, Step by Step

Transcript

55 segments

0:00

I got bad news for you, Chamath. An open

0:02

source ban is coming. They're not going

0:03

to call it that. They're going to say

0:04

that we simply have to apply the same

0:06

standards to open models that we apply

0:08

to closed. Here's how they do it

0:09

step-by-step. Let me explain how

0:11

regulatory capture actually works. So,

0:13

first of all, you got to get this

0:14

regulatory apparatus. Dario wants an FDA

0:16

for AI, but he doesn't have enough

0:19

political support for that. So, instead,

0:21

they do this Trojan horse of a FINRA for

0:23

AI. They call it self-regulating. It's

0:26

not really. But anyway, that gets them

0:28

off the ground. Now, they've created the

0:29

standard-setting organization. Now,

0:31

they've got pre-release model testing.

0:33

Then, the pressure grows to codify that

0:34

in law. So, that happens next. And then,

0:37

what they do is they say, "Look, all

0:38

these standards need to apply equally to

0:40

all models." But here's the problem with

0:42

that. Open models and closed models are

0:44

technologically different. Once you

0:46

release an open model into the world,

0:47

you can't roll it back, and you can't

0:49

monitor exactly how people are using it

0:51

because they run it on their own

0:52

hardware. Dario says this is what makes

0:54

open models dangerous. So, what they're

0:55

going to do is they're going to have the

0:57

standard-setting body say, "Well, we

0:59

have to set the standards for AI

1:00

safety." By the way, Dario and OpenAI,

1:03

they're going to fund the whole thing.

1:05

They're going to contribute all the

1:05

compute. They're going to be behind it.

1:07

They're going to be the ones

1:08

coordinating with the government

1:10

officials because frankly, people in

1:11

government have no idea how to monitor

1:14

and control and set standards for AI

1:16

safety. Technologically, this is way

1:18

beyond them. So, they're going to go to

1:19

these companies and say, "Tell us how to

1:21

do it."

1:22

And so, what will happen is the

1:24

standards will get set, and then it'll

1:25

be a very simple matter of fairness to

1:28

say that the standards need to apply to

1:30

open as well as closed models. The open

1:33

models cannot comply in the same way,

1:35

and gradually they will be shut out of

1:37

the market.

Interactive Summary

The video outlines a strategic process of regulatory capture in the AI industry, where major players like OpenAI and Anthropic are allegedly pushing for standard-setting organizations that would eventually force open-source AI models to adhere to the same compliance rules as closed models. The speaker argues this will effectively kill the open-source market because open models cannot comply with the same monitoring and control standards as proprietary, closed systems.

Suggested questions

2 ready-made prompts