In the post he mentions Sun Microsystems' problems, and worries aloud about other patent attorneys using the Internet to undercut his application costs.
This leads me, once again, to offer my explanation of what is happening in terms of Moore's Law.
The original Moore's Law held that chip density could double every 18 months into the foreseeable future. These exponential improvements have been adopted throughout the industry, and chip-makers like Intel have conquered some Moore's Law limitations, so those improvements continue.
But Moore's Law has a corollary, Moore's Second Law. That is, as chip densities increase, manufacturing becomes exponentially more expensive to set-up. Plants that cost millions cost billions, and there seems no end to it.
That's what has bitten Sun. It's the same force that is biting AMD. The danger in this case doesn't come from free, it comes from monopoly. If only one company, or one group, can afford to make chips they gain a stranglehold on the market, and monopoly profits.
So far, thanks to chip foundries, which allow companies to share these costs, that has yet to happen, but Moore's Second Law results in creative destruction nonetheless. Silicon Graphics is dead. Sun needs a buyer. All hail Moore and Schumpeter.
Of course we make up for it through volume. Once the plant is established, no matter its cost, yields rise, production takes off, and the price of each part continues going down, even as its capabilities go up.
But software is mainly subject to Moore's Second Law, not its first. As it grows more complex development costs rise. All we have done with tools, from Assembler to Cobol to Java, has only mitigated this a little. Programmers still code by hand. The complexity of projects keeps rising.
Open source is the best tool yet developed for fighting this Moore's Second Law impact on software. By working together, by looking at one another's code, by transferring it instantly online, we are creating shared productivity that did not exist before. It's a shared monopoly, a software chip foundry.
In order to gain these benefits, companies forego monopoly rents. If you want everyone to help you make your tool better, you can no longer claim that it is only your tool. The more of these "rights" you give up, it turns out, the faster your tool gets better.
Costs don't go down to zero. They are shared. Those with the greatest needs for better tools put in the most, and seek to monetize this investment in various ways. The rest of us reap the benefits.
But that's not a "race to zero." It's Moore's Law in action, working against the pernicious effects of Moore's Second Law, applied to software.