Ashish — the cross-country divergence is an area I'm keeping an eye on. Within-US convergence, but top 20 countries pulling away from the rest: you frame it as demand-side inequality — "institutional culture, work patterns, organisational willingness to experiment." That's correct, and it's stickier than access.
I also see a topology-dependent disruption — AI bites where economies are already legible, digitized, SaaS-eaten. Your interface tax point reinforces it: software devs have near-zero friction, dentists face enormous cognitive load. The geography of adoption may follow the geography of prior formalization.
I wrote something adjacent — where new work appears in the cracks as AI stalls at the edges. Different angle, same underlying question: where does the friction live, and who ends up working it? https://rajeshachanta.substack.com/p/the-last-meter-economy
The Econ Index offers the first real data set for testing the topology thesis at scale. Would love to see someone map adoption rates against proxies for economic legibility — share of workforce in SaaS-adjacent roles, digital payment penetration, etc.
Lovely post, thank you for sharing! @Hollis Robbins made a related point about education in an excellent essay she wrote back in November (https://hollisrobbinsanecdotal.substack.com/p/last-mile-education). The choice of ‘meter’ and ‘mile’ is something I find fascinating (the length itself and Imperial/SI units, both!)
Thanks. Hollis's piece is a perfect complement to mine. Her "last mile" is my "last meter" applied to education.
And Hirschman is someone I haven't explored — unbalanced growth inducing investment at the bottlenecks. That may be what's happening: AI creates imbalances, and new work appears where the system strains.
Wow. Thanks, for this Ashish.
After reading this, I went and built a self-reflection tool with the help of Claude. The tool runs locally, never leaves the machine.
Premise: Self reflection based on the conversation with your AI assistant.
https://github.com/criatvt/claude-mirror
Ashish — the cross-country divergence is an area I'm keeping an eye on. Within-US convergence, but top 20 countries pulling away from the rest: you frame it as demand-side inequality — "institutional culture, work patterns, organisational willingness to experiment." That's correct, and it's stickier than access.
I also see a topology-dependent disruption — AI bites where economies are already legible, digitized, SaaS-eaten. Your interface tax point reinforces it: software devs have near-zero friction, dentists face enormous cognitive load. The geography of adoption may follow the geography of prior formalization.
I wrote something adjacent — where new work appears in the cracks as AI stalls at the edges. Different angle, same underlying question: where does the friction live, and who ends up working it? https://rajeshachanta.substack.com/p/the-last-meter-economy
The Econ Index offers the first real data set for testing the topology thesis at scale. Would love to see someone map adoption rates against proxies for economic legibility — share of workforce in SaaS-adjacent roles, digital payment penetration, etc.
Lovely post, thank you for sharing! @Hollis Robbins made a related point about education in an excellent essay she wrote back in November (https://hollisrobbinsanecdotal.substack.com/p/last-mile-education). The choice of ‘meter’ and ‘mile’ is something I find fascinating (the length itself and Imperial/SI units, both!)
And yup, what you suggest is certainly worth exploring. This strikes me as a related idea, and I’ll spend some time on chewing on it (https://en.wikipedia.org/wiki/The_Strategy_of_Economic_Development)
Thanks. Hollis's piece is a perfect complement to mine. Her "last mile" is my "last meter" applied to education.
And Hirschman is someone I haven't explored — unbalanced growth inducing investment at the bottlenecks. That may be what's happening: AI creates imbalances, and new work appears where the system strains.
Thank you for both the links. More to chew on.
Thank you both! My first, groundbreaking “AI and the Last Mile” piece was from 2024, and got this wonderful nod: https://www.forbes.com/sites/peterhigh/2024/12/31/top-10-tech-articles-of-2024/