Last night I attended the Computer Hjstory Museum event featuring the founding team of Sun Microsystems – Andy Bechtolsheim, Bill Joy, Scott Mc Nealy and Vinod Khosla – with Sun’s current Chief Scientist, John Gage moderating. Scott Mc Nealy actually referred to John as the “Fifth Beattle”. If you have never visited the Museum, I *really* recommend coming down over the week-end to take a tour of the magnificent exhibits that contains a wealth of computer antiques. And if you like it, you can join as member (as I did) to support their effort.
I have tried to capture the gist of the panel’s conversation. This was pretty challenging since they changed subjects pretty often, and kept on making jokes about each others. It created a great ambiance – sort of an old school pal reunion, but a difficult context to liveblog. For a long time I have been an admirer of the team that made Sun a driving force in computing and networking – until the company lost its mojo in the late 90’s. I started developing on Sun workstations back in 1989 (a sun 3 110 and Sun 386i “Roadrunner” with 4MB of RAM and no disk) when I joined Effix Systemes to create a Unix trading room desktop application – that was before Sun adopted X11 as its GUI environment, and before shared libraries existed in SunOS forcing you to link with the kernel – those were good times.
“Would you invest in Sun ?” was the first question – since 2 of 4 founders are VCs at Kleiner Perkins Caufield and Byers. Vinod recalled that Sun’s original business plan was 3 or 4 pages long. Bill mentioned that the initial round was quickly wrapped up – in a parking lot. The debate was then whether Scott would quit his job (he was then making $40K of annual salary) to join Sun full-time – which he originally refused, and had to be talked into it by Vinod who explained that as a founder, he could not remain employed somewhere else.
Sun (Stanford University Network) came up as a joint idea of Andy, Scott and Vinod who then approached Bill to join them as the Unix guru and write the operating system code, based on Unix BSD (the famed SunOS). One of the key choices made by Sun early on – unlike other competitors – was to adopt open systems and standards that would make development and interoperability easier. At that time, companies did NOT want components and peripherals to interoperate as proprietary elements drove prices, and margins, up.
Contemporaneous (circa 1982) computers were Vax mini-computers, using nascent windowing systems at PARC, and having limited resources (80MB disk and 2MB memory for a Vax 750/780). The initial Sun1 had the same resources, for a 10th of the price. Scott was originally responsible for building boxes according to orders coming through a pool of (6) phone lines. He had developed a simple ERP application (in vi) to enter order and track deliveries.
Sun revenues went through an impressive revenue ramp in its first 6 years: $8.5M, $39M, $110M, $210M, $450M, $1B. Vinod attributed partially that success to the fact that they focused on building a $0B company – meaning that the infrastructure, the processes and the ambition were aimed at getting to $B of revenues.
Moving to the issue of immigration, and the quota forced by the administration on H1Bs, Scott pointed out that US companies had strongly benefited from the work of US immigrants. Andy, or James Gosling, were not born in the US and contributed hugely to the innovation of the global IT industry. He then joked by asking Andy how many billion dollars he paid in taxes, pointing out that he was hardly a burden to the US society.
Discussing long term corporate longevity, Vinod stated that Dec only featured 3 different innovations throughout its history, and this was enough to sustain it for 30 years – until its dies. He likened Dell to the new Dec, not innovating a whole lot on the technology front.
As a technology optimisit, Vinod believes that there is no limit to technology innovation per se, just limitations that human impose to themselves because they do not think “out of the box” enough. This remark was made in the context of a discussion about Moore’s law, and the fact that so many years – it was still governing components innovation . Scott continued by claiming that IBM and Sun are the only two cmpanies investing heavily in computer architecture.
Talking about the Microsoft relationship, Scott explains that whilst he enjoyed his run-ins with Gates/Ballmer, it was time for the two companies to start working together for the benefit of the industry, and to comply with regulatory pressures. Andy made a related comment on how difficult it was to give away stuff for free – referring to the process Sun had to go through to open source Solaris. Now Sun is gunning for Red Hat, as the provider of a great, scalable, high performance operating system.
One of the new innovation challenges is power consumption and its optimization – a strength of the new line of Sun servers, as opposed to sheer performance. Limiting energy comsumption, renewable and alternate energies is actually a key concern of Vinod’s personal investments.
There was then a long Q&A session that I did not capture, but for one question: which role models did Sun founders look up to. Vinod mentioned Intel founders, Andy referred to Albert Einstein’s genius, Bill elected the two Johns: John Gage and John Doerr (now his partner at Kleiner Perkins Caufield and Byers), and Scott… noone.