A weird, perhaps impossible plan to deal with all the extra C02 in the athmosphere

Currently, the CO2 levels are 400 ppm (part per million), but it used to be 270 or something.

This means there is a lot of molecules that could be put to better use.
I think about this sometimes, and yesterday I had an idea:

  1. Build a safe artificial super intelligence
  2. Instruct the super intelligence to design a system of billions self replicating drones that can scale up to the entire earth
  3. Capture CO2 in the rotating blades of the drones, release it at docking stations using moisture
  4. Use the material in automated factors for clothing, building bricks, solar panels, cars etc.

Technically, I imagine something like this being possible some point in the future, perhaps a few hundred years without super intelligence, or a few decades with a safe version. At the rate we are doing, CO2 levels will still be a problem in hundreds of years, because it requires enormous amount of energy to extract it and store it from the atmosphere. The sooner we can do get the levels down, the better.

Some scientists and engineers think of “trees”, but there is a fundamental problem of how large systems you need to capture CO2 fast enough. I imagine that swarms of drones could make it easier to deploy the system everywhere, and one can use simple radio signals to communicate with it, such as “keep away” to clear out drones from places where people live, or where the noise disturbs wildlife. We probably need billions of such drones to make it efficient.

Why you need a safe super intelligence

One reason you need a super intelligence for this, is because the system to produce the drones, docking stations and automated factories must be one integrated system. The drones themselves might require some advanced scientific research that normally takes decades. A super intelligence that can learn science using path semantics could derive the theories and conclusions automatically. The most complex connection is the utility of the captured CO2, which requires technology far beyond current level to be fully automated.

No industrial scale today can match up with the production you need to make as little long term damage to the environment as possible. If it can synthesize threads, we can make clothes of it. Instead of just storing the CO2, we could supply all humans with free choice of clothes and disposal of old ones. If it can make pellets or fibers, we could produce modular building bricks that can be used in housing construction. With large enough supply of materials, people could live anywhere on earth and choose how to live in a palace if they wanted to. You just have to be smart about the design and architecture, and make sure biological diversity is not harmed. The task of the super intelligence is to make the whole system functional and non-harming to the environment and to people.

To make super intelligence safe, one could make it non-human but based on an advanced form of path semantics. It would be able to reflect over its actions and sub goals. Since it has a characteristic non-human intelligence, it needs specific instructions, because path semantics is more like a knowledge base than an active agent. When it is not running, it will just turn itself off and not interfering with human activities. When it is running, it can communicate with humans and show the proofs, designed in a such way that it can not hide anything. The source code can be open and free for everyone. It does not need to be self modifying or emulating human behavior, which prevents large classes of risks.

Drone design

You need air flow to capture CO2, and drones get this simply by keeping themselves up in the air. The rotating blades could suck in some of the air and output lower concentrations. There will be billions of such drones all around the earth, but the atmosphere is big enough to not make it feel intrusive. Drones would have no camera, so there will be no surveillance. Instead, they can use simple radio signals to navigate and communicate. No internet connection, just signals for controlling the drones. The behavior of drones could be simple herd navigation logic. Design hardware for supporting the behavior and nothing else. No security holes, no Turing complete computer you can hack for other purposes. Color could be white to increase albido and reflect more rays back to space.

Advancing civilization through the automated factories

Imagine some person walking out in the desert and putting up a radio beam, emitting a signal to the drones to start assembling a factory. Drones are not very strong, but if they are many they can carry a little weight. To transport something at long distances they can take turns, since there are many enough drones nearby to always prepare a local swarm for the transport. A few weeks later, the factory is complete.

Drones themselves do not do the thinking, but propagate the message to some center where a safe super intelligence make plans and issue orders.

The design of the factory could be made such that it can be removed later without making harm to the environment. It could be fitted to the climate and weather patterns on the area it is built. If no one uses the products, it will automatically be deconstructed by drones.

Think of this as an earth-around free supply of products by need. People can move around, settle somewhere else, and tell the drones what they want.

  • Free clothes
  • Free housing
  • Free food

All this is possible because there is an excess of materials that has no better purpose. It is time to think differently about CO2!

Idea for funding this project:

  1. Make a deal with the Norwegian government about the excess CO2 in their atmosphere, such that the property rights to the collected material is ensured
  2. Sell cheap shares for lifetime free clothes, housing and food in case the problem gets solved. The money is labeled, so it can only be used for research, development, testing, economic scaling up, deployment and maintenance of the system. One share covers owner plus 6 optional people (e.g. friends or family, not bound by contract). First to buy gets first served.
  3. Sell extra, more expensive shares for “experimental” benefits. This is for researching other usage of the material. This could be medical equipment and treatment, vehicles, infrastructure, space travel. Money are used for optimizing shared experience among those with these kind of shares.
  4. Donations to “common good for all of humanity”. This could cover buying properties around the world for research facilities, maintenance and factories not target at personal consumption. Money used as best fit, but justified by analysis, with room for feedback and participation from community.

I believe some people might feel strange buying an insurance for the future just for themselves, but if they also can help 6 other people, they might feel better about it. Imagine you buying a share and after many years there is a terrible economic situation on the earth, but then the problem get fixed. Think of how happy the 6 people you choose would be.

On the other hand, e.g. if the share costs only 1500 USD and you could afford but didn’t buy it. Imagine how bad you feel seeing your friends, which you had an opportunity to rescue. People would line up in the millions to get shares. Yet, even with the help from a safe super intelligence who knows how long it will take. Decades? Do not expect that the world will somehow magically fix everything for everybody, it is probably not what is going to happen. Just look at the world and see where it is heading.

Perhaps there is a production limit and we fail delivering on the promise for more than N people. This could happen if there is a minimum amount of money required per person, even with help from a safe super intelligence. Setting the right price could be of crucial importance. Not too low, so you run out of money, and not too high so people will buy. I think the right price should be “yep, I can afford it” but not as low that everybody buys before even the research get started. After all, you need to cover the needs for 7 people for each share before moving on to the next one. The scaling up part could yield better return on investment, but also expecting diminishing returns when getting close to optimal efficiency.

For people that can not afford the share, they could donate and hope that everything turns out OK. The money used on this is separated from the others, such that even you don’t get as much personal benefit, there could some something.