I’ve observed that business processes tend to follow three phases.

• innovation: an opportunity is found or a problem is solved.
• refinement: improvements are made. The process becomes cheaper, more reliable, and generally better.
• externalization: there are few identifiable improvements to be made. Value capture can possibly be increased, and opportunities to externalize costs are exploited.

In the refinement stage, there are still plenty of opportunities to reduce costs and improve value capture (or yield) without anyone getting hurt. A process that took six months, when performed the first time, might be reduced to two. Genuine waste is getting cut. It’s not zero-sum. However, returns diminish in the refinement stage as the process improves, and the available gains are made.

At this point, talented and creative people look for new opportunities to innovate. Average or risk-averse people look for other processes to refine, and that’s fine: we need both kinds. Vicious and malignant people, however, start making false refinements that externalize costs and risks. Pollution increases, companies become more brittle, and ethics decline. The world becomes sicker, because in the cops-and-robbers game that exists between cost externalizers and regulators, there are just far more of the former. That’s how one gets corporate processes like employee stack ranking and, in software, the bizarre cult that has grown up around two-week “sprints” (as an excuse that allows management to demand rapid but low-quality work). These appear to yield short-term gains, but externalize costs within the company, diminishing the quality of the work.

Most of the sleazy activities for which business corporations are (justifiably) despised are performed in the externalization phase, when the company starts running out of viable refinements and declines to innovate. It may not be (and probably is not) the case that true improvements cease to be possible, at this point. What seems to happen is that there is a shift in political power between those who seek genuine improvements and those who mercilessly externalize costs. Once this happens, the latter quickly and often intentionally drive out the former. They are often able to do so because genuine improvements to processes usually take time to prove themselves, while cost externalizations can be devised that show profits immediately.

It is, at this point, no longer novel to point out that venture-funded startups have ceased to back genuine technical innovation and have, instead, become a continuation of the process (starting in the 1980s) by which the staid corporations and bilateral loyalty of yesteryear have been replaced by quick gambits, degraded working conditions, and a lack of concern by these new companies for their effects on society. This is what “disruption” often looks like.

The realization that I’ve come to is that Silicon Valley, by which I mean the aggressive and immediate financialization of what purports to be technological innovation, is now deep into the third phase. It still “changes the world”, but rarely in a desirable way, and most often by externalizing costs into society in way that regulators haven’t yet figured out how to handle.

While Silicon Valley is marketed, especially to “the talent”, as an opportunity for people to break free of old rules and take on established interests, it’s actually better thought-of as a massive and opaque, but precisely tuned, genetic algorithm. The population is the space of business processes, up to the scale of whole companies. The mutation element occurs organically, due to inexperience and deflected responsibility– when rules are broken or scandals occur, a 23-year-old “founder” has plausible deniability through ignorance that a 45-year-old financier or executive wouldn’t have. The crossover component is far more destructive: corporate mergers and acquisitions. In this way, new business processes are generated in a way that is deliberately stochastic and, when seemingly cheaper processes are discovered, they’re typically snapped into existing businesses, with losses of jobs and position to occur later.

In the three-phase analysis above, Silicon Valley had its innovation phase in the 1970s and ’80s, when new ways of funding businesses, and looser attitudes toward risk, began to take hold. Business failure in good faith was destigmatized. No doubt, that was a good thing. The refinement phase began in the late 1980s, brought with it the golden age of the technology industry in the 1990s, and ended around 2005. Silicon Valley is now in the externalization phase, exemplified most strongly by Y Combinator, an incubator that openly champions the monetization of reputation, self-indulgent navel-gazing, disregard for experience (a polite way of saying “ageism”), and a loose attitude toward executive ethics. The genetic algorithm of Silicon Valley still operates but, in the Y Combinator era, it no longer produces more efficient business processes but, rather, those that are cheaper and sloppier.

The innovation and refinement phases of Silicon Valley are long gone. Cost externalization– the replacement of trusted corporations with good benefits by a “gig economy” culture of itinerancy, the pushy testing of regulations by founding billion-dollar companies that flout them, and the regression into a 21st-century Gilded Age– has replaced the old Silicon Valley and driven it out. This makes it the responsibility of the next generation’s innovators to come up with something entirely new. It is clear that Paul Graham and Y Combinator, as well as those who’ve subscribed to their culture of flash over substance, will have no place in it.