Re: Suppressed inventions

Carl Feynman (
Wed, 30 Apr 1997 11:43:59 -0400

At 01:20 PM 4/29/97 -0700, Robert Schrader wrote:
>On Tue, 29 Apr 1997, Carl Feynman wrote:
>> I estimate the device could be sold for $80 with a manufacturer's profit of
>> $10. The average customer that buys this gizmo will save $300 per year on
>> phone bills, which will cost AT&T about $70 per year. It is worth it for
>> AT&T to pay the inventor any amount between $10 and $70 per unit to make
>> this gizmo not be sold
>> --CarlF
>Were I in such a fortunate position, I would not accept the first offer.
>Even presuming that ATT makes their best offer first ( which from a game
>theory perspective is probably unlikely ), some other big company might
>beat it. So it would be to my advantage to go public and solicit offers.

Oh, I'd contact all three big long-distance companies. But AT&T has by far
the most to lose, because they have the largest market share, so they'd
probably pay the most to suppress it.

>Also, why are we being so generous selling a device for $80 that saves
>the buyer $300 per year? Does it only last for three months? It seems
>if it has a mean lifespan of several years minimum like most appliances,
>the minimum selling price should be upwards of $500.

Well, I'm simplifying. Actually the right business model is to sell it for
$80 and then charge the customers half the money they save every month-- $6
or so. But even if we couldn't charge them by the month, it would still be
the right thing to charge $80 for them. Consumers don't buy money-saving
devices unless they pay for themselves in six months or less. It's crazy,
but my business-savvy wife says this is a well-known fact of retailing. I
knocked another factor of two off that because at least at first, they will
be unproven devices, and consumers will be wary of dropping $150 on them.

The reason it makes sense for AT&T to buy this invention is the irrational
short-termism of consumers, and the high information cost of convincing
consumers that it really worth, as you say, upwards of $500. Unlike the
consumers, AT&T is long-term, and willing to be convinced by e.g. a
prototype deployment with 100 customers, which would only cost $150,000 by
our estimates.

In a market meeting classical criteria of efficiency, this kind of
suppression would not make sense: the customers would pay what the device
was worth, and the former supplier (AT&T) would be screwed, because they'd
lose just as much money by buying the patent as by losing customers to the