Novacene

The Coming Age of Hyperintelligence

James Lovelock 2019 Tigard Lib 570.1 Lov

This is probably 99 year old James Lovelock's last book. I like it a lot, although it is not as bold as it could be.

• Novacene is Lovelock's neologism for the geological era after the Anthropocene, which he dates from the invention of coal-powered machinery.
• Lovelock contends that humans are absurdly unlikely, and that we are the only intelligence currently existing in the entire universe. Extinction is also likely (war, asteroid impact, climate destruction, or eventual solar overheating).
• Lovelock considers himself an intuitive engineer, an agnostic with Quaker roots. The purpose of the universe ("tantamount to a religious statement") is to produce and sustain intelligent life. The cosmic anthropic principle and Gaia were two hypotheses that pointed a seemly way to the future.
• "Cyborg" coined by Manfred Clynes and Nathan Kline in 1960, a self-sufficient cybernetic organism made of engineered materials.
• No war between humans and machines because we need each other.
• The Newcomen steam engine was developed because it worked cheaper than human or horse power.
• Moore's law, doubling every two years, has continued for at least 40 years. He mentions a "possible" silicon stall, followed by a huge diamond chip advantage.

my note:

I would date the Anthropocene from the use of metal agricultural tools and fire to change land use and release CO₂. That's a small quibble. Compared to the multibillion-year tenure of Earth life, or even compared to a fraction of a Milanković cycle, these are practically simultaneous events.

In Lovelock's Novacene, machine intelligence displaces human intelligence as Earth's master, reversing climate change for its own survival. The unmentioned assumption is that machine and human will remain discrete unitary entities. As individuals devolve essential mammalian traits (like navigation and trust) to machines, it seems that a merger has already begun. The few humans who guide this merger will chart the path that other people and their machines will follow.

Lovelock writes about sunscreens to cool the Earth, but most of the sunlight reaching the Earth's mostly-ocean surface does not drive photosynthesis and carbon capture; artificial lifeforms (designed, modelled, and tested by imaginative bioengineers with the help of computers) could reduce atmospheric CO₂ by 30 ppm per year. That could feed other artificial lifeforms that lift nutrients from the ocean's floor, as well as spin gigatonnes per year of superstrength carbon fiber for mega-structures.

While 174e15 watts of sunlight reaches the Earth's surface ( and electric launchers could use 5% of that for 100 megatonne per second space launch ), two billion times more power is emitted by the Sun, mostly launched into the empty void. Feeding that power into machine computation, and emitting the waste heat into 2.7 Kelvin deep space, would power vastly more Novacene machine intelligence than Earth's puny share. If that capture and emission occured at 50 AU distance, where the black body temperature of the capture system might be 60K rather than 300K, and the energy per computation reduced by 5x, the Arrhenius chemical decay rates reduced by factors of billions, then "intelligence per watt" would be perhaps 1e20 higher than the mere 100 billion watts of human brain on Earth today.

Kuiper belt AI will be amused by many things, but the antics of biolife and humans down here on our damp corrosive mudball will certainly be worth the tiny amount of sunlight we consume and the waste heat we generate. If we focus on expansion into space (shielded wheel habitats in zero gee), and build the necessary AI support systems, then we've made the first steps towards the massive expansion of cyberlife Out There rather than competing with us Down Here.

MoreLater