Bryan Cantrill reduces Sun Microsystems’ failure to one sentence: the company had become bored with the mechanics of running a business.

His evidence is a 2005 episode that was, on paper, a vindication of Sun open-sourcing Solaris. A startup growing like a weed — one of the companies pioneering what we would later call cloud computing — was running its infrastructure on OpenSolaris and wanted to buy a large amount of Sun hardware. The strategy had worked. The customer was standing there with money.

  • Sun could not get the customer on the phone. When Sun did pick up the phone, it tried to sell them the wrong product.
  • Dell’s contrast is the whole point. The startup filled in a web form in the middle of the night and the next morning got a call from Steve, the local account executive. Under two weeks later the servers were in the datacenter, leased on the company’s financials with no personal guarantees, and 95% of the work had been done for them: “I felt like a Big Company. I felt like Steve worked for me.”
  • The startup wrote the whole experience up publicly in 2006 so that Sun could see it. Cantrill read it while squatting with the Fishworks team in a corner of abandoned Sun office space, and describes his heart sinking: strategic success, operational failure.
  • Sun kept going a few more years and the people inside it tried to right the ship. It wasn’t enough.

The personal ending is the argument’s coda. Cantrill left Sun in 2010 and joined that startup; the startup had the good sense to hire Steve from Dell as well; years later the two of them started Oxide together. The team’s nostalgia t-shirts for defunct computer companies, he says, are not really nostalgia — they study these companies to learn what they got wrong. “We honor them best by learning from both — to be at once inspired and warned!”

Most of the 244-comment thread on Hacker News is first-hand memory of buying from Sun — but it also carries an inside account of Sun’s internal structure, a vendor-side counterpoint, and one argument with the essay’s thesis.

What the thread adds

  • bcantrill — the author appears in his own thread, mostly to narrow the blame. He objects to reading the piece as engineering blaming sales (“there were some extraordinary sales folks at Sun”), points readers at the 2006 blog post for the structural analysis, and puts the failure in leadership: “Scott visited customers where they lived – and Jonathan more or less didn’t” (Sun’s two successive CEOs).
  • kens — nine years inside Sun, 1995–2004, with the internal version: the real competitor was other groups inside Sun, and the divisions wanted incompatible things. Solaris wanted to run on SPARC, x86 and PowerPC to reach the largest market; the SPARC division wanted none of it. Bubble-era revenue let all of it be ignored until it couldn’t be.
  • coreyh14444 — hands-on corroboration of the buying experience from the late 1990s: Sun and DEC forced you into live sales meetings and endless quote revisions, while the rails and power cords for a new Alpha server could cost more than a delivered Dell server. Reply sschueller adds the symmetry — Dell has since become the vendor that charges whatever it thinks it can get.
  • jonathaneunice — the counterpoint from the other side of the table. Buying from Sun was a mess, but selling to Sun was a joy: it paid like clockwork (net 35) while preaching hyper-efficiency. The failure was in what Sun did with its own money, not in how it dealt with its suppliers.

Where the thread pushes back

  • schmichael — reads the piece as engineering blaming sales and asks where the voice of sales is, noting hardware margins were under attack from Dell and from the cloud that Sun’s own customers were building. As a sales rep handed a startup’s order, they write, “I probably don’t have much incentive to pick up.” Cantrill replies in-thread that the complaint isn’t about the people: sales were as frustrated as everyone else, and leadership was “not interested in running a business.”
  • jedberg — disputes the premise by degrees: Sun was “never interested in running a business” rather than bored with one, and after pushback edits to add that the company was already being crushed by the early 2000s. Replies supply the hardware counterpoint: otterley reports frequent repair calls on Sun gear versus “solid-as-a-rock” Dell and Compaq, while mattbillenstein says commodity x86 single-thread performance was “amazing compared to the Sparc boxes” and that Sun “were eaten from the bottom by faster cheaper hardware.”
  • rbanffy — the missing on-ramp: why would anyone invest in SPARC and Solaris when a generic x86 box with Linux does 99% of it for far less, and why develop on one platform to deploy on another? No credible desktop offering, they argue, is a long-run wound — a critique reply nickpeterson turns back on Oxide itself, asking whether a six-figure hardware purchase competes with near-zero startup cost on AWS.

On reading comments as evidence: HN handles are pseudonymous and the site publishes no per-comment scores, so this is a slice of HN’s own ranking, not a consensus. The theories about what doomed Sun — structure, incentives, commodity hardware, leadership attention — are quoted as commenters’ arguments, not as findings.