Neurofarms
From Biohack
According to the two themes (see getting up to speed), self-modification is a form of self-creation, especially in the case of brain implants. Now, with a brain implant, you don't just randomly modify your genome or decide to delete a neurotransmitter receptor from the mix, or decide to add in a random new polypeptide or protein since this will alter your mind to the point of no return. So, you need to do sandboxing of new functionality, and taking a hint from software development practices, never integrate a test feature into a product release (the current running brain) without extensive testing.
The brain has many billions of neurons, many trillions of connections, possibly thousands of genes uniquely responsible for the brain (how many? see the Allen Institute and brain atlas projects for mapping mRNA activity throughout different slices of the brain). Testing these on the scale of experiments that we do today is basically impossible since we can't physically provide enough manual labor with the number of people willing to contribute, and even with machines carrying out some of the tasks, there's only a limited amount of possibility space that can be explored.
So, the idea of a neurofarm is to provide a scalable architecture for experimenting on neurotissue. There fundamental unit will be the neuropod: a slice of neural tissue in a dish with a microelectrode array, a nutrient bath (with nutrient circulation and so on), etc. The neuropod will have a basic definition, but there will be types of neuropods, such as some to experiment with different neurochemicals, some to experiment with different genomes and viral modifications, others to experiment with the longevity of neurons, others to test the biocompatability of different electrodes, etc.
How do we scale up a neurofarm? It is definitely possible to have a neurofarm that can be managed by a few humans. It is definitely possible to have a small farm of rats with brain implants. After all, look at zoos, or look at Google, a giant computer supercluster managing hundreds of thousands of machines (some estimate a few million by now), and Google is growing with some dx of 20k petabytes of data per day, we don't know the computational growh though. But at this scale, only a few different combinations of changes can be tested at the same time. Mathematical/computational neurobiology (see computational biology for software) hasn't reached the point where we can do analytical creation of new neurobiological circuits in the mind, and it may never (but let's hope, and we can contribute to those projects in the mean time). But otherwise, the scene looks set for a giant neurofarm to be constructed for sandboxing of tests to brain tissue.
The basis of the neurofarm idea is to scale up the number of neuropods, to make the manufacturable, and to scale up the number of nodes manufacturing neuropods via KSRM self-replication. Where would we get all of these resources and an environment suitable to self-replication? Interestingly enough, asteroid mining might provide some interesting opportunities for getting enough minerals for this sort of project. There's no specific need for humans to be in direct physical contact with the experiment -- it can be automated, as long as the data is fed back down to the surface of the planet for further analysis and review, making sure that there is automated life support systems still functioning, making sure that the asteroid mining bots are still getting enough food to the neurons (perhaps through a synthetic ecology, see asteroid synbiosphere project stuff).
Contents |
Neurofarm logistics
- Rate of growth equations
- Physical manufacturing of the neuropods
- Getting neurons up into orbit, as well as DNA synthesizers and DNA sequencing.
- Data downlink
- Need to deploy something to get into LEO and then orbit
- SugarShot2Space
- LOX (DIY fabrication -- otherwise this costs too much)
- Orbital neurofarm would require LOX-in-orbit manufacturing capabilities (luckily, you can access oxygen from that side of the earth atmosphere). This would be used to stabilize its position and so on.
- Might need to make only one rocket, but not even the experts a la John Carmack and Henry Spencer have been able to do this yet, really. Although NASA has done asteroid landing before with a satellite, so that's the start of asteroid mining, especially with cyanobacteria and the moontank project to make bacteria that are ready to eat through asteroids.
- What would be the optimal size of a group of neuropods?
You can probably play around a bit with the neurochem kit.
Neuropod design
- Neuropod standards/specifications.
- Experiments?
Is the neurofarm idea a blackswan?
If self-replication is a blackswan, then so is the neurofarm idea, to the extent that it requires self-replication, although the 'sandboxing by hand' methods for testing of potential self-modifications is not a blackswan, since this has been shown for decades by neuroscientists to be legit.
Note that if we can use cells to do self-replication and the equipment necessary to run a neurofarm experiment, then all is well, but the problem with this is that we generally cannot use cells to make machinery, send radio signals from orbit, or be used for remote cytodiagnostics. So this particular implementation idea has to go to the backburner for a while (perhaps indefinitely).
Do not mess with the peripheral nervous system
The peripheral nervous system (somatic nervous system, autonomic nervous system and the enteric nervous system), particularly the ANS with the sympathetic division and parasympathetic division, meaning cardiovascular blood vessel dilation, heart, breathing, other important functions for keeping the body going. Consider this a read-only zone. Modifying anything else might, at worst, produce a seizure or something, not drop you dead on the spot.
