The Invisible Trillion · The Other Ledger
What This Chapter Doesn’t Say
I want to close by being clear about what this chapter is not arguing.
It is not arguing that technology is net-good and therefore we should stop worrying about the costs. You have just spent three chapters on the costs. They are real. The energy, the water, the mining, the e-waste, the attention economy, the concentrations of power that come with concentrations of infrastructure — all of it is real, and none of it is magically cancelled out by the benefits on this ledger.
It is not arguing that every new technology is good. Some technologies are bad and should be banned or regulated. Some technologies are mixed and need careful stewardship. Some technologies are so transformative that we can’t even yet tell what they’ll turn out to be.
It is not arguing that you should be grateful. Gratitude is a strange emotion to direct at infrastructure. The people who built the infrastructure are not asking for your gratitude; they are asking to be paid, which is different. Your relationship with the infrastructure is simpler than gratitude: you live inside it, it shapes your life, and you have some say (small but non-zero) in how it evolves.
What the chapter is arguing is this. The infrastructure you have been worried about — the chips, the fabs, the data centers, the AI models, the whole stack — is, among other things, the infrastructure that feeds you, cures you, warns you, educates you, and tells you the planet is in trouble. The cost ledger is real. The benefits ledger is also real. Neither one erases the other.
The interesting question — the grown-up question, which is the question this book has been building toward — is not whether to have the infrastructure. It’s how to shape it. Who decides what it does, who pays for it, who benefits, who bears the costs, what it’s allowed to be used for, what it’s not. Those decisions are being made now, by specific people, in specific institutions, through specific processes.
The rest of this book is about that question, applied to the buildout happening right now: who lives next to it, how much of it we actually need, what it would look like if it were built well, who pays the bill, and who decides.
A note on sources: the medical statistics are from the CDC, NIH, and the American Cancer Society. The AlphaFold figures are from DeepMind’s published papers and the EBI AlphaFold Protein Structure Database. The weather forecasting accuracy figures are from the European Centre for Medium-Range Weather Forecasts (ECMWF). The M-Pesa and Aadhaar impact studies are from academic work; William Jack and Tavneet Suri’s work on M-Pesa poverty reduction is a good starting point if you want to follow up. The claims about AI-accelerated software development and the self-driving safety record are developed, with their caveats, in the buildout chapters that follow. Where specific percentages appear, I’ve used the middle of credible estimates; individual studies vary.
Questions to talk about
No answer key. These are for arguing over, with someone who read the same chapter.
- The author names the tonal trap in advance — that listing good things turns a ledger into a brochure. Did he avoid it? Point to a sentence either way.
- Costs and benefits are both large and both real. Why is holding both harder than picking one, and why do most books pick?
- 'The interesting questions are how to get more of the second while paying less of the first.' Pick one benefit from this chapter and one cost from an earlier one, and propose a trade.
All 19 chapters of The Invisible Trillion · Library · All courses