I actually don't buy this.<p>Lurking in the background is analogy to e.g. car manufacture. There was an era of fomentation, followed by an era of consolidation. Eventually, we wound up with a handful of automakers.<p>But the market for software is fundamentally different. For one thing, there is not the same reliance on mass labour, nor is there significant material cost; nor is there the need to maintain a sales network.<p>Sure, there are always network effects which replicate some of the momentum/inertia of a physical-goods market, and yes, ZIRP was one heck of an accelerant, and it's not coming back soon.<p><i>But</i> software has not finished eating the world. It might not even be halfway through, as AI has opened up new regions of the (if you will) 'carcass' for software to ingest (much as fire and the cooking of food opened up new nutritional pathways for humans.) And even when software is <i>finished</i> eating the world, opportunities for disruption will remain plentiful, as incumbents will have a hard time maintaining moats.<p>Now, that said, I expect three things to happen in tandem: (1) the average software engineer salary will stagnate or decline; (2) credentialism, long foreign to the industry, will begin to creep back in; you'll see CS and Eng degrees rise in importance, along with more situation-specific certifications; (3) the average market cap of startups (specifically: SaaS startups) will decline, along with growth cadence.<p>This isn't death. It's just middle age. A touch of grey, as Jerry Garcia put it.