The conventional story about AI: we're at the beginning of a fundamentally new era. A counterargument worth examining says the opposite - AI is the end of the current era, not the start of the next one.
An analysis drawing on economist Carlota Perez's framework for technology revolutions argues that AI is the mature deployment phase of the digital wave, not a new wave in its own right.
How Technology Waves Actually Work
Perez, a Venezuelan-British economist, identified a consistent pattern across five major technology revolutions: early textile mills, railways and steel, electrification, automobiles and mass production, and computing. Each follows the same arc.
First, an installation phase: a new technology emerges, capital rushes in speculatively, infrastructure gets built fast, and a financial bubble forms. Then a crash. After the crash comes the deployment phase - the technology is now cheap, mature, and widely distributed, and genuine productivity gains finally arrive. Eventually the wave ends and a new one begins, usually built on a different physical or scientific foundation.
By this reading, the digital wave started in the 1970s. The dot-com crash of 2000-2001 was its inflection point. Cloud computing, mobile internet, and e-commerce were the deployment phase - finally delivering the productivity that digital technology promised but couldn't yet achieve during the bubble years.
Where AI Fits
Under Perez's framework, AI isn't a sixth wave. It's the final chapter of the fifth.
The argument: AI makes existing digital infrastructure more productive. It automates tasks that computers could theoretically handle but couldn't do economically. It reduces the cost of generating text, code, images, and decisions by orders of magnitude. These are real gains - but they're gains layered on top of chips, internet infrastructure, cloud platforms, and software that already existed. AI didn't introduce a new physical technology; it made the existing one dramatically more capable.
Compare this to what genuine new waves looked like historically. Railways didn't incrementally improve horse travel. Electricity didn't marginally upgrade gas lamps. New waves introduce fundamentally different productive capabilities. AI, by this analysis, is doing something more like what mainframes becoming personal computers did - making an existing technology far more accessible while remaining within the same wave.
The genuinely new wave - most likely in biology, energy, or materials science - hasn't arrived yet.
What This Means for Practitioners
If this analysis holds, a few things follow.
Tools like Claude and ChatGPT make knowledge work faster and cheaper in the same category as spreadsheets made financial modeling faster and cheaper. Valuable, without being civilization-altering. The productivity gains from using AI effectively are real - they're just deployment-phase gains, not paradigm-shift gains.
The risk is over-indexing on AI as a destination rather than a tool. Companies figuring out where AI genuinely reduces cost or improves output will benefit. Companies building strategies around AI changing everything are making a different and riskier bet.
The weak point in this argument is that Perez's framework was built to describe completed historical waves. Identifying which phase you're in from the inside is genuinely difficult. It's possible AI is actually the installation phase of a new wave, with a correction and deployment phase still ahead. But the practical test is simple: are you earning a return on AI adoption, or are you speculating on AI transformation? Those are different bets with different risk profiles - and right now, most organizations haven't clearly chosen which one they're making.