[webmasterraj]

The major difference between what's going on now and what happened in 2000 is transparency. Not with what's happening at companies, but what investors are thinking. That's important, because investor sentiment is the proximate cause of any bubble bursting.

In 2000, it was easy to gauge investor sentiment – you just had to look at stock prices, which are real-time, precise indications of how investors as a whole feel. That meant it was easy to tell when the bubble started to burst (arguably on March 10, 2000, when the Nasdaq peaked).

What's hard today, and maybe why there's a lot of these "is it a bubble or not?" stories, is that investors don't know what other investors are thinking. Sam and YC have consistently been open, but they're more the exception that proves the rule. The vast mass of investors are holding their cards close to their chest.

The net effect is to prolong an increasingly unsustainable situation. I think it's highly likely that most investors already secretly think we're in a bubble. And that they also believe that other investors haven't realized it (everyone assumes they're the smart one). So we're stuck in this phase of "make my last buck before all the other idiots realize what's going on."

[nostrademons]

Secrecy actually makes it more likely that terms reflect correct valuations, as long as each of the firms have independent decision making power and see a wide variety of deals. See eg. The Wisdom of Crowds, which found that a group makes better decisions than an individual iff each individual relies on their own data and comes to their own decision. If one individual unduly influences others' decision-making (eg. the "YC stamp of approval" which inflates valuations at demo day), then you get information cascades that can lead to exaggerated boom/bust cycles. If the parties left in the dark lack decision-making ability and independent information channels (eg. Theranos), you get information asymmetry and a market for lemons. But a market where each firm sees a continuous stream of deals and independently decides on each one without any regard for its competitors is pretty much the definition of a competitive market, which implies prices should converge toward accuracy very quickly. I'd argue that that's much closer to the situation now than it was in 2000.

[webmasterraj]

This is a really good point, and explains why later-stage investing seems so wacky vs early and mid stage. The decisions to invest in later-stage startups, eg Uber, seem more influenced by other investors and the fear of missing out, while early to mid stage might steer closer to the independent decision making model.

[InclinedPlane]

That's one major difference, the other is the maturity of the industry.

In 2000 the online tech industry was fledgling, and most of it was vapor. The bubble was bigger than the entire industry. Today tech has grown across the board and the web has become a key pillar of that growth. Trillions of dollars in real value has been built and added to the economy of the world. As big as any bubble might be, the industry is so much larger and on such firmer footing today that a collapse wouldn't have nearly the same impact.

[ethbro]

I hadn't ever really thought about it this way, but it's a good point.

Fundamentally, one would expect some relationship between total "digitized" (?) economy value (where digitized means able to be controlled, measured, and/or analyzed by software) and the value of startups.

And I don't think many people would argue computer-industry merging hasn't been happening across different industries fairly quickly.

[InclinedPlane]

The "web economy" isn't solely digitized though. Sure you have things like, say, youtube, but you've also got Uber, Amazon, AirBnB, etc, companies that fundamentally rely on the internet and software but which have a big dependency on physical things.

[ethbro]

That's my point though. The degree to which "average physical thing that is depended on" is integratable with software at any point in time.

What is Uber worth with self-driving cars? Or with more granular locking standards (this car will authenticate and unlock for me because its driver added me)?

What is the traditional side Amazon worth without the ability to automatically route packages using laser readers and codes? Or with the ability of packages to self-navigate to their destination?

What is an AirBnB worth with an adjustable room? I walk in, all my preferences are imported into the room: illumination amount, light temperature, room temperature, music? Or, again, if the locks know me because of my digital keys rather than swapping physical ones?

If these businesses have a heavy dependency on physical things (as a lot of tech businesses do), then the degree of integration between that thing and software defines how much value can be delivered. Obviously you can increase this yourself with devices, but that's much more expensive than relying on an increasing average.