Vignette K: The Spring of 1992, IP Routing, and the Scott McNealy Lesson

[RAW USER SOURCE: Direct Narrative]

So when I met with Dave Marquardt, it was in the spring of 1992, whenever we were graduating. And I had a job offer to continue at Microsoft. In fact, I actually worked part-time my second year. I got one of the first copies of Remote Access Server, which allowed you to dial up remotely into the Microsoft network and filing systems. And I had some sort of job and some sort of deliverables. Basically, what I was doing is using it to learn. And that's why I didn't end up as a Henry Ford II Scholar, which would have been the top graduate. I was in the top five going into the final semester, but I blew off most of my classes and got a few P+'s instead of A's. So I would have been the top graduate at Stanford Business School, or so I was told by Basil Karampelas, my classmate and buddy, who got word from the registrar's office, or whatever they called it, like I shouldn't leave early to go to Europe, which was my plan. 'You need to come to graduation. You might be the number one graduate.' So it was like, 'You have to come. We don't know yet because they haven't finished the grades for the last semester, but if you, you're in, you're in the lead.' So come to graduation so we can give you your freaking prize. So I rescheduled my flight two weeks so I could actually go to graduation because of that. Anyways, I was talking to Dave Marquardt in the spring, and I had a job offer to work for Darby in doing the what became MSN. Darby's one of my best friends. I haven't seen him for a while, but amazing, amazing friend and huge stories with him. Wonderful guy. He worked for this asshole, well, I'm sure he's brilliant, but Rob Glaser, who ran the whole division, doing the Internet Superhighway stuff, which is just what they called in the popular press networking. They were just using the wrong protocols. ATM, there's like 50 different networking protocols. SONET, like it was crazy. It was like an alphabet soup of protocols, of networking protocols. And that's when I did my project, it's like, 'You should just route IP.' Internet Protocol, it works. In fact, I heard that from a professor at Stanford, what was his name? I forget his name. He was the guy that did hybrid networks, and he's actually famous because he did some huge, huge wins. I was one of the investors in freaking Google. One of the first two investors in Google. What was his name? David something, Cheritonov? I forget. He was in the engineering school because I got tired of talking to business school guys. I already knew everything, so I started hanging out in the engineering school, which is right across the quad. But anyways, I digress as I often do. So Dave Marquardt told me his story over lunch. I can picture the place. Beautiful restaurant. He has very He's a very elegant gentleman. And he taught me a lot over the years. I've spent tons and tons of times with him at TBI, and he was just a great mentor. Never invested, but he came close, and he gave me encouragement all along. Anyway, he said, 'You know, in my class,' he was class of I forget what his class was, but 'In my class,' you know, he asked me, 'Who's going to be the best, like the top graduate in your class?' In my class, it was Scott McNealy. And he said, 'Nobody would have thunk it.' He said, 'They, you know, he was bottom 10% in the class academically going in, playing golf, chasing girls, drinking beers. He just wasn't a serious guy, but, you know, he ended up as the CEO of Sun Microsystems, so he was the most successful of Dave Marquardt's classmates at the GSB, according to Dave.' And he asked me who did I think would be the top person in my class. And I thought about it for a while. I actually thought set for like 30 seconds. I said, "You know, there's probably 20 people that have the right innate capabilities to be the top person, but 15 of them will end up wasting their lives working for either something like Goldman Sachs or McKinsey because those are high prestige, relatively high earning occupations that are safe. But five people would really go for it. And of those, four would end up in a yuppie homeless shelter because they'd strike out and fail, and then one would make it. One of those five that really took chances would be the top person." Now, at that time, I already had a job offer from Microsoft to continue or to work for Darbee Williams in the group that was going to do what became known as MSN. At the time, it was a different architecture. It hadn't really been defined yet, but it was the new internet architecture, or the new information superhighway, as they called it. And but I also had an opportunity to work for Blyth Software, spelled B-L-Y-T-H Software. It's a product had a product called Omnis, which is the only cross-platform, basically Mac, Windows, cross-database. It worked with Oracle, Sybase, Microsoft SQL Server, RDB from DEC, and and also So it was basically the only cross-platform across everything development tool in client-server with the client-server architecture. And I had met them because they're one of our partners, and they asked me to go squeeze them because they'd signed up for some minimum some sort of distribution deal reselling SQL Server, and they hadn't hit their target, so they owed us money. So I went and met with them. I was like, "Why would you?" And they told me basically to threaten to sue them to squeeze some money. I was like, "Hey, they're barely alive, you know." So I got to know the team there and the product and say, "Hey, this this could be really interesting because I could tell that with the progress that PowerSoft was making and PeopleSoft, I could just tell that the client-server market was about to take off, and Blyth Software, even though it was a relatively small company, it was already public. And, you know, from my experience in the financial in the financial world and in Wall Street, Cornerstone Research, said once a market goes hot, these idiots on Wall Street just they just do a sector play. So they if it looks like a comp, it'll go up with it.

[AI DISTILLATION: Contextual Synthesis]

Graph Nodes & Edges: [People: Dave Marquardt, Basil Karampelas, Darby Williams, Rob Glaser, David Cheriton, Scott McNealy] | [Institutions/Companies: Stanford GSB, Microsoft (MSN), Sun Microsystems, Google, Goldman Sachs, McKinsey, Blyth Software (Omnis), Oracle, Sybase, DEC, PowerSoft, PeopleSoft, Cornerstone Research] | [Concepts: Remote Access Server (RAS), Henry Ford II Scholar, IP Routing vs. ATM/SONET, Academic Merit vs. Execution, Sector Play, Cross-Platform Client-Server Architecture]. Synthesis: In the spring of 1992, Jonathan’s trajectory at Stanford GSB intersected with the bleeding edge of early internet architecture. Working part-time for Microsoft via an early Remote Access Server (RAS), he sacrificed the chance to graduate as the Henry Ford II Scholar (the top academic spot, which he led going into his final semester) to focus on raw, applied learning. While evaluating Microsoft's 'Internet Superhighway' initiative under Rob Glaser and Darby Williams (which eventually became MSN), Jonathan identified a critical architectural flaw: the telecom industry’s reliance on an 'alphabet soup' of complex protocols like ATM and SONET. Influenced by cross-quad conversations with Stanford engineering legends like David Cheriton (a foundational investor in Google), Jonathan recognized that simple IP (Internet Protocol) routing was the superior path forward. Over a mentorship lunch, Dave Marquardt reinforced the distinction between academic perfection and real-world impact by sharing the story of his own GSB classmate, Scott McNealy—a bottom-10% student who went on to become the wildly successful CEO of Sun Microsystems. Challenged by Marquardt to predict the most successful graduate in his own class, Jonathan articulated a stark assessment of institutional capture: of the twenty most capable students, fifteen would sacrifice their potential for the "safe" prestige of firms like Goldman Sachs or McKinsey. The ultimate outlier would emerge from the five who embraced extreme risk. Applying this logic to his own career, Jonathan weighed his offer to remain at Microsoft against a volatile opportunity at Blyth Software. He had originally been sent by Microsoft to threaten Blyth with a lawsuit over missed SQL Server reseller targets. Instead, he recognized the disruptive potential of their Omnis product—a uniquely cross-platform, cross-database client-server tool. Drawing on his Wall Street modeling experience from Cornerstone Research, Jonathan correctly anticipated that as the client-server market ignited behind companies like PowerSoft and PeopleSoft, public markets would execute a 'sector play,' indiscriminately lifting comps like Blyth alongside the vanguard.

Vignette L: David Cheriton, Andy Bechtolsheim, and the ARPANET Roots of Bill Joy