What if everyone who's calling this a bubble has forgotten what a real bubble actually looks like?

The first live video stream on the internet was a coffee pot. In 1991, a Cambridge researcher

pointed a camera at the breakroom pot so he'd know whether there was coffee before walking downstairs.

People called it a toy, a gimmick with no serious application. The coffee pot webcam, in no small way, became Netflix.

The pattern repeats.

Every major technology wave starts with use cases that look trivial,

and every time skeptics confuse silliness with insignificance.

30 years later, hundreds of billions of dollars are pouring into AI infrastructure.

Consultants estimate the current spending would require AI revenue to grow 40x by 2030 to justify it.

The dot-com comparisons write themselves.

But the dot-com crash wasn't just overvalued stocks.

It was a fiberglut financed by WorldCom, a company with $40 billion in debt that was cooking its books,

compounded by 9-11.

The company's funding today's AI buildout have hundreds of billions of cash on the balance sheet.

Comparing valuations isn't the same as predicting systemic collapse.

That distinction matters.

This conversation examines what separates a speculative correction from an economic crisis

and why this moment may look more like the mobile or cloud booms than dot com.

Martine Casato is a general partner at A16Z and a few months ago, he joined the Wall Street Journal's bold names podcast.

We're sharing that discussion here.

You know, it's interesting you were in San Francisco and Silicon Valley during the last

bubble bursting.

...