In order to understand and plan for the future, we need to know how much change will occur and when things will happen. Understanding the present and near future is informed by identifying and understanding what is already undergoing sustainable and rapid change. There also needs to be quantification wherever possible.
Nextbigfuture is working on a systemic framework for analyzing the future from 2021-2100. What are other major parameters?
Classifying the rate of change:
Explosive change is 41% or more change every year, doubling in less than 2 years
fast change is over 20-40% change every year, doubling every 2-3 years
moderate change is 8%-19% change every year, doubling in 4-9 years
slow change is 2-7% change every year, doubling in 10-35 years
virtually no change is -1 to 1% change, almost no change for 40-100 years
decline is -2% or less, halving in 35 years or less.
China had moderate change GDP growth in 39 years out of the last 60 years (1961-2020). There were only four years from 1978 to 2013 that did not have more than 7.5% GDP growth. China was doubling its economy every 7-9 years. China has slowed to 5-6% GDP growth which is doubling every 12-15 years. It is only exceeding US GDP growth by about 2%. It would take about 35 years for China’s economy to double relative to the US economy.
India has had moderate change GDP growth in 10 years out of the last 30 years. India had 3 years of greater than 7.5% GDP growth rate from 2010-2020. India is doubling its economy in 12-20 years. India is also exceeding the GDP growth of the USA by 2-3% each year. It will take 25-35 years for India to double relative to the USA. China is still maintaining a 5-6% GDP growth. India is not closing the gap with China and India is not consistently outperforming ASEAN countries like Vietnam and the Philippines.
It does not appear that there will be any country that will have consistent moderate change GDP growth (7.5% each and greater) from 2020-2030.
Tesla has explosive change with 50% year over year growth.
Amazon has fast change with 20-40% year over year growth.
Broad Classifications of the Timing
near future is likely by 2030
mid future is likely in 2031-2049
far future is 2050-2100
very far future is 2101 and beyond
* Civilization is majorly altered
This would be like the invention of farming. Changing people from hunter-gatherers to builders of cities.
The potential civilization altering possibilities for the 21st century are:
Superintelligence – broad AGI, very useful multiple narrow super-AI, strong molecular nanotechnology, space travel beyond the scale of commercial aviation, regenerative extreme life extension
The major technologies would need to have some level of quantification. There would need to be some log scale or magnitude classification of different super-technologies.
Beyond human-level AGI is too vague. Just being a lot faster is not as profound as better at true innovation. The speed and productivity would need to categorized and defined.
We have some quantification of narrow superintelligence AI. There are some way to quantify the AI that play chess and go.
The best chess AI is currently MuZero. In 2019 DeepMind published MuZero, a unified system that played excellent chess, shogi, and go, as well as games in the Atari Learning Environment, without being pre-programmed with their rules. It surpassed AlphaZero after 1 million training steps. MuZero used 16 third-generation tensor processing units [TPUs] for training, and on 1000 TPUs for self play (for board games, with 800 simulations per step) and 8 TPUs for training and 32 TPUs for selfplay (for Atari games, with 50 simulations per step).
AlphaZero used 64 first-generation TPUs for training, and 5000 second-generation TPUs for selfplay. As TPU design has improved (third-generation chips are 2x as powerful individually as second-generation chips, with further advances in bandwidth and networking across chips in a pod), these are fairly comparable training setups.
Stockfish is a computer program that beats the best human players. Alphazero beats Stockfish. Alphazero won 28 games, drew 72 and lost none. MuZero beats Alphazero.
AI has completely solved the game of checkers (draughts).
There is also quantification around the AI that help people drives cars and for self-driving cars. The metrics are miles driven per human intervention, miles driven per accident and miles driven per fatal accident. Tesla Autopilot currently helps people to be ten times safer on miles per accident metrics.
The altering of civilization can be quantified. Is per capita income being increased ten times? Does humanity expand to colonize earth orbit? the moon? Mars? Asteroids? Beyond?
* Century defining – This would be like the industrial revolution level (steam engine)
* How we live every day has major change.
* Trillion-dollar or more impact.
Written By Brian Wang, Nextbigfuture.com