Technology Installation Cycle

Technology installation cycle is the source’s Carlota Perez-style frame for how major technologies move from early introduction through financial excitement, bubble risk, broader coordination, and maturity. In E162.康波周期中的AI：新技术总在萧条期爆发，bad times make good people, the guest contrasts this innovation-cycle lens with Kondratiev Cycle analysis: one tracks how a technology diffuses, while the other asks how economic waves turn.

Applied to AI, the episode places the field closer to an introductory or installation stage than to a mature deployment stage. That explains why the source can be structurally optimistic about AI while still warning that an early financial bubble may break before the technology becomes widely embedded.

141. Freda的投资札记第2集：Tokenmaxxing、把电机塞进蒸汽机、接力赛变篮球赛、孤独、人的连接 adds the AI Economic Diffusion version of the same pattern. Freda / Friday uses the electric-motor and steam-engine analogy to argue that productivity does not arrive when a new technology is merely inserted into old workflows; it arrives when factories, firms, software, and teams are redesigned around the new capability.

So are we in an AI bubble? Here are clues to look for. adds the post-bust infrastructure question through Productive Bubble Spillovers. The episode compares dot-com fiber to possible AI data centers and R&D: an installation-stage bubble can destroy investor wealth while still leaving assets that later deployment may use.

Infrastructure lessons from the dot-com bubble turns that comparison into a reported infrastructure walk. Paul Vixie’s account of dot-com fiber construction and Dark Fiber shows the installation stage in physical form: cable was laid before demand fully arrived, companies failed, and later applications used the capacity.

Bytes: Week in Review - Are we in an AI bubble? adds David Kirsch’s “infrastructure as timekeeper” version. The source compares AI with railways, electrical distribution, and aviation: infrastructure can be installed, but broad value depends on airports, norms, communications, business models, market design, and organizational routines that may develop more slowly than capital spending.

真正改变世界的技术，为什么一开始都不被看好？| S10E16 adds the emotional-history layer. 汪波 compares present-day surprise, optimism, and fear around ChatGPT with earlier reactions to telegraph and telephone, while the MOSFET / MOS Transistor and Moore’s Law stories show that the installation cycle can begin with both technical skepticism and overlarge social hopes.

AI 发展了 4 年，把应用发展没了？｜AI 年中复盘 adds a stage-question version through 曲凯 / Qu Kai’s use of Carlota Perez. The episode asks whether the AI wave, nearly four years after ChatGPT/GPT-3.5, is still in an explosive installation phase, has entered a frenzy phase, or is approaching a turning point where token prices fall, intelligence becomes cheaper, and application value returns to the center.

Key Claims

• Installation-stage technologies can be technically important before their social, organizational, and investment returns are settled.
• Early finance and narrative can run ahead of deployment, making AI Equity Valuation Risk compatible with genuine AI progress.
• The useful question is not only “is AI real?” but whether infrastructure, workflows, business models, and institutions have absorbed it.
• The installation/deployment frame complements AI Investment Metrics by adding a longer maturity question to tokens, ARR, CAPEX, and deferred revenue.
• Technology diffusion and economic diffusion can diverge: model capability can spread faster than organizational redesign.
• Reusable infrastructure can make an installation-stage bubble socially different from a purely speculative collectible boom, even when public-market investors overpay.
• The installed base matters: Dark Fiber shows how deployment can lag installation by years, turning unused capacity into later economic infrastructure.
• Kirsch’s aviation analogy adds that an early visible use case may be closer to demonstration than mature deployment.
• Early reactions to a technology can swing between dismissal and utopian expectation before real deployment boundaries become visible.
• The model-versus-application rotation can be read as an installation-cycle symptom: capital first prices infrastructure and model control, then later asks where broad application deployment and user value appear.

Connections

• Carlota Perez — theorist invoked by the episode for technology-cycle staging.
• Kondratiev Cycle — long-wave economic-cycle frame paired with this innovation-cycle frame.
• AI Investment Metrics — observable near-term business signals.
• AI Equity Valuation Risk — market risk when prices assume deployment before it is proven.
• Depression Driven Innovation — macro condition in which new technologies often begin installation.
• AI Economic Diffusion, AI Organization Design, and Agent Native Software — episode 141’s productivity-absorption extension.
• Productive Bubble Spillovers, Dark Fiber, AI Compute Continuity, and Data Center Debt Risk — post-bust infrastructure extensions.
• David Kirsch, Tech Bubble Conditions, and AGI Narrative — Marketplace Tech’s historical bubble-score and narrative-risk extension.
• Disruptive Innovation, MOSFET / MOS Transistor, Moore’s Law, and ChatGPT — S10E16’s early skepticism and emotional repetition extension.
• AI Application Market Trough, 曲凯 / Qu Kai, Carlota Perez, and AI Commercialization Pressure — 42章经’s question about whether the AI wave is approaching a Perez-style turning point.