Startups, Bubbles, and Unemployment

[Tim May]

On Tuesday, August 27, 2002, at 06:36  AM, Matt Curtin wrote:

> Tim May <tcmay at got.net> presumably wrote:
>
> [I'm not sure if Tim actually wrote the following, as a splat
> proceeded the "Back" but nothing else in the paragraph.  Apologies in
> advance if I'm misattributing.]

This is a problem as articles get bounced around, sometimes to some of 
the many similar lists run by one person.
>
>> Back in the 1970s and early 80s, high tech start-ups were difficult
>> to do. It took some really good ideas, or at least the departure of
>> a talented group of people who had already developed
>> something. Examples like Sun and Cisco are examples of where the
>> core technology had already been developed (Stanford, in both cases)
>> and where the companies could begin to SELL PRODUCTS almost
>> immediately.
>
> I cannot help but wonder how the cycle of "start company, get
> financing, hire friends, build stuff, retire rich" was influenced by
> increasingly Draconian intellectual property agreements and whatnot.
> I was too young for gainful employment in the 1970s, but maybe some
> folks who worked in high technology at that time could shed some
> light.  I'm particularly interested to know whether this "we all of
> your thoughts" business came about as a result of people going out on
> their own to sell products they developed on their own time while in
> another technology company's employ.

I was there, at Intel, and will give you my thoughts on this.

* When there were about 10 chip companies, extensive patent 
cross-licensing deals had been struck (mostly in the 1960s). There 
_were_ a few lawsuits (notably Fairchild suing others for violating the 
"oxide isolation" patents), but not many.

* The focus on patents was minimal. Trade secrets were much more 
important. I can tell you that Intel had several tricks ("Getter III 
V-over-I," "Pyroglass," "Dip-back," special equipment custom-built, 
proprietary CAD systems long before CAD companies existed, etc.) that 
were not patented, just kept secret. Something crypto folks ought to 
appreciate.

* The focus on patents was so low that our patent guy, Ed Taylor, kept 
his office in LA and only came up to Santa Clara a few times a year. 
Mostly he discouraged the filing of patents, and certainly middle 
managers did not encourage engineers to waste time trying to get 
patents. I got one during this time. A friend of mine was told not to 
bother to file for a patent on something that other chip companies later 
(two years later) were trumpetting as their patented invention.

* Things began to change for many reasons.

*First, the entry of the Japanese, Taiwanese, and Koreans into the 
market. They had not be party to the cross-licensing deals.

* Second, some American companies only peripherally involved in chips 
attempted to sue chip companies for violating their patents (a notable 
example being the Hughes Corporations suits against chip companies doing 
ion implantation...Hughes claimed that a patent they had gotten on ion 
implantation meant huge royalties were due them).

* Third, patent portfolios came to be seen as valuable in takeovers, as 
part of the "good will" that could be charged for. Some large companies 
bought small companies just to acquire their patents.

* It is not likely that patents and IP have had much effect on movement 
of people between companies. Just my opinion....I see as much moving as 
ever.


> (I know not what Stanford's intellectual property rules are, but at
> Ohio State, if you develop something significant while working there
> and go off to take it to market, they're almost certain to want a
> piece of that action.

Stanford has gotten pieces of the action from many companies. This shows 
up in their holdings of stocks in those companies, reported on the Form 
144 Insider pages on Yahoo and other places. In the case of CISCO and 
Sun, I don't know all of the details, but there were no lawsuits against 
either. As I said, my understanding is that Stanford said "Go ahead" to 
those seeking to commercialize their projects.

> Somewhere during all of this time, the whole "push technology"
> stupidity came about, and people wanted to know how our product
> stacked up against "other push products", when it wasn't really
> comparable to what most of the other guys were doing.

Ah, "push technology," the buzzword of around 1997, just after set-top 
boxes and just before whatever.

> Bootstrapping a company is hard if you don't have a bunch of money to
> throw at it.

And, I would argue, even harder if you DO have a bunch of money to throw 
at it!

(I watched friends burn through $8 million as they sat around thinking 
cool thoughts. More germanely to this list, look at several crypto 
start-ups who threw tens of millions at problems.)

> It is possible, but requires serious financial
> discipline, throughout the company.  People who expect to walk into a
> job with super deluxe office space, a big support staff, and a huge
> salary just aren't going to find a place in a company that's getting
> itself off the ground.  Scott McNealy makes a lot of money now, but if
> that's what he demanded from Day One, he wouldn't be at Sun.

And Sun, like many other successes, had very spartan office arrangements 
in the beginning. Then they got big. And bloated. Now they have an 
entire "campus," a much more campus-like campus than most companies 
have. But their stock is at $4.25 and the end is coming (IMO).

>
>> To close on a less bleak note, this is a fine time for people to
>> "get back to basics."
>
> Crypto still has serious problems to solve.  Key management is still a
> big problem.  User interfaces are a big problem.  Integration with
> non-hacker tools is a big problem.

I have differing views on these points. But that's for another time.
--Tim May
"Aren't cats Libertarian? They just want to be left alone.
I think our dog is a Democrat, as he is always looking for a handout"  
--Unknown Usenet Poster