Really Wild Stuff

Stratospheric Microballoon Artificial Lifeforms

Playing with artificial lifeforms is hazardous. Some worry (perhaps too much, perhaps not enough) that the artificial genes we create will outcompete and replace natural genes. Others look at the vast multiplicative effects of genes expressed as the proteins and compounds we can extract from artificial lifeforms and see vast productivity with minimum resources drawn from nature, compared to the old and inefficient techniques of bulk chemistry.

I look at artificial life forms as non-agile and early binding. Early binding describes ideas and decisions that must be made early in a design project - integrated circuit process design, for example. Late binding is like new revisions of patched software, produced immediately before use. Is there some way to produce lifelike structures that are late binding, agile constructs that can produce new materials almost as soon as you can think them up, and stop producing them equally quickly?


We assume that lifeforms are self-contained and genetically complete replicators, just like the natural ones produced by Darwinian selection. While most lifeforms depend on communities of differing lifeforms to thrive and reproduce, they do not rely on externally supplied genetic information for essential steps in their replication process. But that might be just the ticket for artificial lifeforms that must not reproduce in the wild without intentional human intervention.

Ribosomes copy DNA information into RNA. RNA information is used to construct proteins. What if we built a different kind of RNA, which does not transcribe DNA, but instead contained externally supplied information? What if artificial lifeforms were missing large chunks of genetic information, essential for metabolism and reproduction, but had some means of using externally supplied information to accomplish these tasks instead?

Imagine a different kind of ribosome, controlled by millimeter-wave radio signals rather than DNA, a "radiosome". We can broadcast public key radio signals into a cloud of artificial lifeforms containing these radiosomes, and send them instructions to make RNA and in turn make proteins, either as end products or as the tools for reproduction. Without the right signals and the right private key encoding them, reproduction ceases. With different signals, the artificial lifeform immediately starts producing different RNA and eventually end products.

Broadcast Exogenes

attenuation graph from Electronic Design magazine

A good operating band might be 200 GHz, which is attenuated 12dB per kilometer in the sea level atmosphere. At higher altitudes, the air is thinner and the attenuation drops proportionally. At 12 kilometers altitude, the air is 4 times thinner and the attenuation is only 3 dB/km. With a density lapse rate of 8 km, the total attenuation of a signal arriving from space will be 24dB, a factor of 250. On the ground, the attenuation will be 96dB, a factor of 4 billion. A signal strong enough to reach an artificial lifeform at the tropopause will not be strong enough to reach the ground.

200GHz is interesting because if some degradation mechanism drops the frequency, attenuation increases. If the frequency drops a lot (say to 80GHz), the attenuation vanishes, but signal bandwidth drops a lot, too. This minimizes the possibility that the system may evolve towards a ground survivable one capable of interfering with existing ecology. Indeed, we might move to 360GHz to put one more attenuation peak between our design and nature.

Habitats for Artificial Life

If we grow organisms on the ground, we displace the native organisms that were there beforehand. Most deserts have native organisms adapted to arid conditions. There are few places on land where we can grow something new without displacing nature, with consequences we are only beginning to understand.

Meanwhile, recent increases in CO_2_ trap more heat in the atmosphere, and without some magic compensatory mechanism, this can cause undesirable long term surface heating (averaged over time and space, this does not change the wide statistical variations that produce extremes). If we slightly attenuate the light reaching the surface, we can add some artificial compensation and select optimum temperatures. One predicted effect of warming is the thermal expansion of the oceans, raising sea levels. Most of the ocean is relatively lifeless, with carbon capture rates a tiny fraction of land and of the coastal margins where most ocean life can be found. Sunlight absorbed in the ocean deserts produces evaporation and cloud formation. Most of the resulting rain falls back on the ocean, with the evaporation and precipitation cycle mostly acting to transport heat energy back out to the tropopause to be radiated into space.

If we slightly reduce the sunlight reaching these ocean deserts, at times and in places that do not result in rain on land or otherwise affect the active biosphere, we can gather terawatts of light with small or nonexistent effect on living nature. The effects will not be zero, but they will be much smaller than crop agriculture for the amount of sunlight gathered. We can use that sunlight to produce energy, chemicals, and food.

Hydrogen Mining with Artificial Balloons

William Mook of Mökenergy proposes 20 meter balloons that capture and concentrate sunlight to produce hydrogen. The proposal is evolving, but one version involves floating balloons that roam the upper atmosphere collecting sunlight, then returning to the ground to offload hydrogen into pipeline networks, and pick up more water. While I don't really understand how altitude cycling will work, or the range of navigation possible by changing altitude into differently directed wind layers, the idea is interesting to think about. The big expense will be manufacturing and maintaining billions of large balloons - they won't grow themselves.

Artificial Microballoons

But what if the balloons could grow themselves, and create desirable products with directed molecular biochemistry rather than fixed function macroscopic apparatus? A millimeter-scale balloon would be big enough to receive information from W/m2 bursts of radio energy, and small enough to be a single artificial cell. It might crack water into hydrogen, or it might perform the energetically simpler task of making and/or collecting methane as lift gas. That behavior, and the materials of the envelope, and many other aspects of construction and behavior could be programmed on the fly, with only a hard-coded kernel of DNA containing public key and error correction behavior. Most of the RNA/genetic information stored in the cell would not be permanent, but downloaded for the needs of the moment and replaced when those needs changed. The artificial organism would not be viable in the wild; indeed it would soon die and disintegrate without frequent software updates.

The tropopause could have vast clouds of these artificial organisms, absorbing water, carbon dioxide, and whatever particulate pollution makes it up there. We might have to artificially seed these clouds with other trace elements. When they complete their synthesis tasks, some are programmed to descend, and pass through subsequent phases of their artificial life cycles.

Trees, Ants, and Spent Balloon Collection

With detailed modelling of the atmosphere, we can predict precisely where a small cloud of microballoons will descend and arrive at the ground. We can program the descent to occur only when we can be sure the wind will blow the small cloud through a particular small forest containing a particular well-characterized species of harvester ants. The balloons will get caught on and stick to the leaves of the trees, and emit chemicals designed to attract the ants so they collect the balloons and carry them back to their nest.

A signficant percentage of the balloons will contain "ant food", feeding the nest and rewarding the foragers. Other balloons will contain industrial product, and will stop signalling the ants when the balloon passes over a collection strip. The desired behavior (mad handwaving here) is to harvest all the microballoons in the forest, deposit a large percentage of them for collection, and use the rest to feed the anthills. After a few kilograms of material has been collected, it is harvested by small robots and gathered into a series of larger collection stations. Nature will behave differently than it would without the artificial microballoons, but far less differently than it would if we replaced the forest with a farm, or heaven forfend, solar cells.

The balloons could also deliver specific-target biocides or nutrients to the forest. While we should be very careful about interfering with natural cycles, it may take centuries of well-thought-out interventions to help forests recover from previous centuries of habitat destruction, invasive species, and other disruptions of the original forest community. This will be an ongoing learning process, supplemented by vast amounts of observation, computation, and experimentation.

Project Persephone

Most of that experimentation will occur in solar orbit, far from earth, so that the chances of unintentional contamination of the biosphere by those experiments is very small. Resources and balloon programming information will be standing by to contain and remediate whatever might possibly slip through the barriers.

Without complete genomes and life cycles that can evolve those genomes, and with dependence on artificial materials that nature does not normally provide, our artificial life will need to be protected from far more robust Darwinian life, not vice versa. A big advantage of widely separated, small artificial organisms in stratospheric and space environments is that they are protected by distance and physics from nature's predators, allowing them to be less robust and easier to design. We will not deploy test organisms in the space environment until after they have been thoroughly modelled with computation. We will not deploy artificial organisms to the stratosphere until they have been exhaustively tested, by the trillions, in orbit. We will have the wealth of space resources needed for lavishly excessive testing, not merely for safety but for optimization and improvement of future designs.

Just as Server Sky will be the first step towards filling the solar system with computation, the Japanese Project Persephone will be the first step towards biological experimentation in space. Experimental systems that weigh grams or even milligrams, working in arrays, managed by and communicating through server sky arrays, will provide venues for safe biological experimentation at a fraction of the cost of performing those same experiments safely on earth.

Even if access to space remains more expensive than terrestrial access, the advantages of vacuum and microgravity can more than make up for the costs, especially for microsystems vulnerable to terrestrial contamination. Computation and microbiology are two easily forseen near-term uses of space, but as we develop other new microsciences, they will also find their natural home in space.

RwsMicroballoon (last edited 2015-01-16 04:56:35 by KeithLofstrom)