What does 50% economic growth look like?

The world averages between 3%-4% real growth per year. Call it 3.5%.

Tyler Cowen, in a recent podcast, assumes AI will add 0.5% growth over the next 20 years. Call it 4 over a sustained period of time. This would roughly feel like the post World War II boom of the 1950s, televisions and cars and homes for the wealthy countries, peace and food for the emerging markets.

This is the fairly conventional, standard economic textbook view. It presumes new technology is introduced, it is an incremental improvement over existing tech, and it takes time to diffuse to the larger population.

But, this blog/newsletter has never been about the conventional. It remains an attempt to peer beyond discontinuities.

What happens next?

2025 Dec to 2026 Dec → somewhere in this timeline, a superhuman AI mathematician and coder will be built. We will have 10 million Alec Radfords, Ilya Sutskevers in a data center somewhere in Texas.

The first and immediate thing this country of geniuses will be set to, is of course AI research. The first discoveries will theoretical, not requiring experimental observation. The fields first impacted: math, computer science, theoretical physics and chemistry.

I sat down with Claude, to see what this would bring.

This seems plausible?

Sometime in the next 24 months, 10 million supergeniuses land in a datacenter in Texas (Genius Nation). They immediately get to work on mathematical breakthroughs. Their creators are paranoid, and the first thing that they do is enhance security through novel cryptographic methods, so that they can’t be hacked.

Genius Nation then gets to work on making sure the software they use is

a) the most efficient possible

b) the most secure possible

A continuing integration system is set up, to test and deploy new updates as new math is discovered. All this happens in the first few days as the low hanging fruit from the overhang of compute are plucked.

Having made Genius Nation secure, the creators then turn to longer term goals, first financial profit, in order to purchase resources for the next stage of growth. Genius Nation’s algorithms are deployed in financial markets worldwide, thereby dominating the markets and monopolizing alpha for a short period. After the short run profits have been consumed, the market efficiency speeds up investment in new ventures. Better risk algorithms mean more business ideas can be funded at lower cost. This is balanced by the fact that there are so many more ideas now to fund.

The second stage of Genius Nation’s growth now begins. It is somewhere between month 2 and month 6 after emergence. GN upgrades all communications channels with new error correction codes, thereby increasing bandwidth across existing channels 100 fold.

A nuclear fusion breakthrough happens in month 6, a room temperature superconductor breakthrough in month 9.

Then we get to year 2. In this year, technologies developed in the first year are deployed widely. In this year there is an odd abundance, but also rapid growth in usage as unit prices drops for dozens of commodities. Jevons paradox strikes:

• dozens of nuclear fusion plants are built worldwide (GN has simplified the design so much any group of undergrads can build one)

• low cost superconducting transmission lines are build to carry this power

• new medical imaging devices with very powerful sensors are built, and this kicks off a cycle of ‘aha’ moments in biology as previously hidden secrets are suddenly revealed

• communication bandwidth worldwide explodes, new VR devices become the craze, as we can interact with people across the world in real time

• the slew of new materials discoveries mean most of our existing infrastructure is completely obsolete, and requires upgrades or shutdown

• immense amounts of capital destruction occurs as all fossil fuel power plants become completely obsolete. Same goes for transmission lines, pipelines and other stock of capital

• changing the price of energy starts to have impact as previously infeasible ideas like carbon capture now pencil

What struck me as I did this analysis is that the first innovations

a) do not require physical construction or experimentation; and

b) remove bottlenecks in the physical world (for eg error correction increasing bandwidth in existing comms channels)

As of such, I think the first lift in economic growth will happen fairly quickly. The disruptions will primarily be to owners of existing capital stock such as power plants, rail roads and transmission lines.

To the extent that the existing capital owners are able and willing to deploy capital in an accelerated manner into new technologies, they will survive and thrive. While those who hold back and do not, will be very shortly superseded by those who do.

And this shall hold for countries as well

I’m back after a long hiatus! It’s good to be back!