The Productivity J-Curve [pdf] (2018)

I did a write-up on the history of the J-curve and some 2026 macro data that supports it for generative AI. The short version is that in 2026 the U.S. productivity is climbing again after a decade of stagnation.

https://lightsight.ai/blog/j-curve

(disclosure: my company’s blog)

This paper clearly shows why you shouldn't use MS Word to typeset research.

This completely breaks down under the current reality of AI investment, as players large and small are no longer price-takers. The marginal costs of investment are not constant because we have finite supplies of GPUs, TPUs, memory, hard drives, and power. The Hamiltonian in equations 5 and 6 needs to account for this.

If you are one of those that are amused by attempts to synthesize paradigms, here's one that superposes J-curve on the hype curve

https://www.financialprofessionals.org/training-resources/re...

Q: The J-dip is where capital stock is just about to overtake investment growth, why should it lag the hype trough where presumably value overtakes interest ?

FYI about terminology before people who don't read the paper comment

1. GPT means general purpose technology or any sort of new technology that has a compounding effect on productivity, not the OpenAI model.

2. Productivity in this case means economic output, not the colloquial definition that means "hard work". If it takes 5 automotive factory workers to assemble a car manually but 2 with industrial automation, then the latter are more productive than the former despite expending equal amounts of effort.

3. The crux of this paper is that existing economic metrics are not able to adequately measure the impact of IP and R&D driven innovations in the larger economy. For example, think about how it took 20-30 years for traditional econometrics to fully gauge the impact of digitization and industrial automation that began in earnest in the 1990s and early 2000s.

[flagged]

[dead]