> On the other hand, how long did it take Microsoft to figure that out - remember the Windows 95 TCP/IP stack?

Having a good TCP/IP stack core isn't about figuring out that plugging into a national computer network is one of the key values of a computer, its about knowing which national network will be the key one (and, more specifically, which technology stack that one will be built on.)

The web -- the bit that made the Internet win that battle as the consumer-facing network that would matter -- was fairly new when Windows 95 was released and hadn't even been invented when work on Windows 95 began. At that time, dialup, largely-text mode offerings that didn't rely on a TCP/IP stack not only dominated local BBS's (some of which were part of national interBBS networks), but also represented some of the biggest (inter)national consumer-facing services to which computer users were likely to connect (like CompuServe, which had already secured leadership in that space when Jobs gave his 1985 interview.)

> The point is, plenty of smart people missed this boat for years, and I don't think it was at all obvious in 1985 that getting on a network would be the driving use case for widespread computer adoption by "non-techie" people.

It was widely accepted. Its true that in 1985 it wasn't at all obvious that the Internet or any specifically TCP/IP-based network would be the specific network, the recognition that connecting to a large-scale network was a compelling feature and likely to become the most compelling feature was widespread in 1985.

It did have a tcp/ip stack but it wasn't installed by default, it was one of the optional extras.

What it didn't have was a web browser, you needed the plus pack for that.

trumpet winsock and hours of troubleshooting different configs :) irq settings etc