GO Corporation

~ the last words ~

GO closed in July 1994. I was ten years old and four years away from my first modem, so everything below is reconstruction from the outside. I include it because the shape of this failure keeps turning up later, in products i did live through.

Jerry Kaplan founded the company in 1987 alongside Robert Carr and Kevin Doren, and what they were building was a computer you wrote on. Not a computer with handwriting bolted on top of a mouse driven interface, which is what everybody else eventually shipped. The whole operating system, called PenPoint, assumed the pen. Documents behaved like paper in a notebook. You crossed a word out to delete it. There was no desktop metaphor because there was no desk.

For a while it looked like it was going to work. GO was one of the most heavily funded startups of its era and the press liked it enormously. PenPoint ran on x86 tablets from IBM, where it carried the first use of the ThinkPad name, and from NCR, and on AT&T's EO Personal Communicator. The staff list reads like a directory of people who went on to run things elsewhere, including Bill Campbell, Omid Kordestani, and Mike Homer.

None of it turned into revenue. Kaplan's own account puts it flatly: the company generated no meaningful sales in its lifetime. Consumers in 1993 did not want a tablet, and the vertical customers who did want one wanted it cheaper and sooner than GO could deliver.

~ the record ~

~ the funny part ~

GO split itself in half in 1991, spinning the hardware out as EO Inc. AT&T bought EO in 1993, mostly because it wanted a shop window for its own Hobbit microprocessor. That sale raised cash GO badly needed and immediately created a new problem, because EO now answered to AT&T and started looking at rival operating systems while GO sat on the software side watching.

So in January 1994, out of cash, GO agreed to sell itself to AT&T as well, which put both halves back under one roof. Two weeks after the deal closed, AT&T cancelled the Hobbit line. The company had spent years narrowing its bets onto that chip and had just handed itself to the only party that could switch it off. There was no reason left for AT&T to fund either half, and by July it did not.

~ the next chapter ~

Kaplan wrote the whole thing down in Startup: A Silicon Valley Adventure, which is still the book people hand you when they want to explain what venture capital actually feels like from inside a company that is losing. Then on 29 June 2005, eleven years after the doors shut, he filed an antitrust suit against Microsoft claiming its engineers had taken GO's work under an NDA and used it in Pen Services for Windows. Judge J. Frederick Motz threw it out on the four year statute of limitations, and the Fourth Circuit agreed in 2007. The merits were never argued. The court's position was that GO had been on notice since 1992 and should have sued while it still existed.

What GO got right was almost all of it except the year. Handwriting as the primary input, gestures for editing, a document model rather than a file system, all of it turns up again in the Newton, then in Palm, then eventually in a slab of glass that sold hundreds of millions. Being early is not a lesser version of being right. It is its own way of being wrong, and $75 million is roughly what it cost to find that out in 1994.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »