Talk:Craver CTF Essay
From Wise Nano
Contents |
a suggested modification for safely allowing recursive assembly
Nanoblocks are pre-fabricted, sophisticated blocks of atoms (molecules, I guess) that are larger than 100 nm in scale, at least. Some are computers, some motors, some structural, etc. Each can be tagged by type, and even by (semi)uniqueID, for forensic purposes. these are sold in various packs through existing physical delivery infrastructure - in markets, corner stores, shipped cross country, etc. Think razor blades, where the fabs themselves are razors.
nanoblocks are manipulable only by special tooltips. One kind of tip is required to assemble blocks, while another is required to disassemble them. For a lot of applications, "nanoscale", or even 100 nm scale, isn't necessary - even sub-millimeter scale would be sufficient. with such blocks, most products would be more like microtech - even the smallest devices would be visible (albeit barely) to the naked eye. It would still be enormously useful.
L1: assemblers and recyclers begin as L2 products for general public use, assembling an array of L1 blocks that do not include assembly or disassembly tips, prohibiting the manufacture of L1 products that can recursively assemble or disassemble any other product. Recyclers of L1 products remain L2 products, but are publically available in controlled forms to prevent fabrication of malicious free-range disassembly devices (see below).
After an initial trial period of limited L1, assembly tip nanoblocks can be released, allowing for recursive fab, but not for fab of free-range disassemblers.
Note that the initial assemblers (which are themselves L2 products which assemble L1 products) must be fabbed or recycled by L2 facilities. after the assembler tip L1 blocks are released, the second-generation assemblers will be L1 products that assemble L1 products (including recursive assembly of other L1/L1 fabs)
L2: assemblers and disassemblers are only usable by civil authorities in restricted high-security facilities. they work at the atomically precise level, able to assemble and disassemble any product. products include: recycle bins that can disassemble L1 products (only), for distribution to the public; free-ranging nanobots that can disassemble L1 products (only), for use only in warranted circumstances, or when carting away loose, lost L1 blocks (cleanup);
The best part: no damn annoying DRM, and no accompanying dangerously false sense of security.
Does this stop foreign crash programs creating their own atomically-precise fabs or defabs? no. What does? What it does do is keep the really dangerous (ie, really small) applications out of the hands of civilians, including, most importantly, small terrorist groups and well-meaning klutzes. the vast majority of beneficial apps are covered at the macroscale, while withholding most of the dangerous applications to civil authorities. Because civilians will be appeased by the available products, the motivation for civilians to crack the system will be reduced.
Note: having separate assembly and disassembly keys takes care of grey goo on two fronts: first, nanoblocks do not occur in nature. goobots might make a mess out of your can of raw blocks, then go hunting for your neighbors, etc, but reasonable precautions on storage of raw blocks can keep pesky goobots out of it.
second, disassembly tips are only available on L2 products, which can't be disassembled by L1 products, so no L1 products can disassemble existing products, as they would require if they were to attempt to recycle existing L1 nanoblock products in the wild to use to reproduce themselves.
So yeah, there are a number of capabilities you have to keep locked up with civic authorities; but there is a lot you can release.
--Nato Welch http://n8o.r30.net/
A few responses Chris Phoenix, CRN 19:08, 10 May 2006 (CDT)
The ID could be totally unique for each block. Only takes a few atoms to represent 10^100 code values.
You mention assembly and disassembly as though they're symmetrical. They're not symmetrical; assembly uses well-ordered feedstock, while disassembly has to deal with unknown atom placement. You come close to suggesting a universal assembler and disassembler, and there ain't no such thing.
Large blocks are one of several ways that consumer products could be made relatively safe. Other ways include low power density, medium-low strength, and no sharp edges. You want to be careful not to permit a flywheel bomb, a lethal hummingbird UAV, etc--or at the very least, make them hard to build accidentally!
Two keys may be a good idea...
Using two "key" manipulators is worth considering. I didn't include it or think deeply about the possibility, because it adds some complications:
- if there's only one key, it can be very simple, with correspondingly simple lock mechanism - a flat bar that gets twisted to activate block to block locking might suffice. If there are two, they must be quite different - adding complexity to the "locks".
- instead of a 'near impossibility' (can't build anything as small as the key out of nanoblocks), security now rests on making the "recycle lock" unable to be "picked" using the "assemble" key or several keys cleverly manipulated. Assuming nanoblocks need to be key-able from any face, this could mean a significant increase in complexity.
- Nanoblocks might be deliberately designed to make it impractical to build a reliable fabber from them - protecting the commercial market for the fabbers. In that case, the "key" concept might just be an elaboration on a general scheme to prevent the main benefit (fabber duplication) made possible by having separate keys. It's even possible that nanoblocks could be designed to be one-use by consumers, but easily recyclable by the manufacturer - they'd still be cheaper than non-reusable nanoblocks, but the manufacturer gets a continuous long term cash flow.
I'm not saying it isn't worth doing - just that there were deeper issues than seemed worth going into in the essay, especially given the fairly high probability that any commercial maker of nanoblock fabbers will design to prevent them from self-duplicating. In that context, I felt that maintaining the ability to recycle the nanoblocks was a more important point to make.
grey gobblers and utility string
I don't like the name "lumpy goo" for Malicious nano block fed self replicators because I don't think that they would be goo - like.
I prefer the name Grey Gobbler. It sound ominous, the grey is for the machine aspect, and gobbler focuses on what it is doing --- eating ---- potentially everything you have ever fabbed. So now we can talk about two types of nano-tech monsters, grey goo - which eats the biosphere, and grey gobblers - which eat your nano-tech.
Utility string is an attempt to cross nano factories with utility fog.
In Chris Phenix Primitive Nano Factory paper, he mentioned the possibility designers not dealing with 100 nm nano blocks but with larger virtual materials made up of patterns of nano-blocks. This got me thinking about what is more complex than a nano-block but still simpler than just about anything else? If a nano block is a point then the next step up is a line. The more I think about it the more i realize what a difference an extra dimension can have! Not blocks or fog lets but strings or fibers.
Lets rethink our basic design unit. We want it to be easily reused by us but not by some robot looking to grow / reproduce. If we make each unit much bigger than the 100 nm block and imbed some intelligence into it we can solve both problems. I think that the basic unit should be a hexagonal fiber 1-10 microns in diameter and ~1,000 microns in length. At each end of the fiber there is a standardized connector that transmits information and energy from one fiber segment to the next. Each fiber segment should have the capability to store and process information and energy. You have thousands of cubic microns to play with, you can have a set of common capabilities to all fiber types and still have plenty of room for fiber segments to be specialized for particular tasks.
By making the making the basic unit linear, much larger, and intelligent the task of disassembly is greatly simplified. The de-fabber contacts the object to be disassembled and supplies it with the disassembly code. The object sends a message along all the fibers that make it up telling the fibers to disengage from all neighboring fibers, thus causing the object to fray into a pile of string that is cleaned, tested, sorted and stored for later reuse.
Utility string: good idea
You could certainly fit a lot of smarts into a 1x1000 micron fiber. The overall volume would be the same as a 10-micron cube--i.e. a smallish human cell. This seems a reasonable size for most human-scale products.
I'm not sure what advantage a fiber has over a microblock. You don't necessarily get higher tensile strength. Fibers sliding past each other may be a problem due to environmental dirt; for shape-changing, a 10-micron block could certainly be made to deform its shape by a factor of 2 or more.
I don't know whether it would be easy or difficult to clean such a fiber for re-use. A 10-micron block would have slightly less surface area (600 vs. 4000 square microns).
Chris Phoenix, CRN 09:03, 12 June 2006 (CDT)
Why micro fibers not microblocks
For the sake of this discussion we want to focus on the situation in which the manufacture of the nano components are separated from the manufacture of human scale products.
Ok lets say that you are going to use Micro-blocks, you manufacture them at a secure site and you are going to ship them to homes all around the world. How do you package them? you want hem to be easy to use and undamaged in transit. The simplest solution I can come up with is to put them all in a row. In other words make them into a fiber or a string.
Once you accept that the blocks are going to be in a row, you take advantage of the linearity to simplify the design of both unique human scale products and reusable micro-scale components.
Anti lumpy-goo product coating?
To prevent malicious fabbers from eating all your nanoblock products, could a coating or spray be affixed to the surface area of nanobloack products? It would function like anti-graffiti sprays presently work. I know a malicious diamond nanoblock could cut a nanoblock product's diamond coating. But all that would be necessary would be for the electrical properties of the coating to be altered by the malicious nanoblock virus. The coating would be doped or consist of conducting CNTs and connected to an indicator of sorts (and a battery). A nanoblock virus would alter the voltage across the coating and trigger the inticator. Presumably a person could then do something about the lumpy-goo once made aware of it.
(added Oct/21/06) It looks like there is already something like what I have suggested above, being researched for nanocomposite applications: http://nanotechweb.org/articles/news/5/10/9?alert=1
Phillip Huggan
