There are several lessons here. The most politically salient is that in manufacturing, as in cooking, it is possible to “lose the recipe.” And with an accelerating pace of technological progress, it is possible to lose it in an alarmingly short span of time. This is perhaps the strongest argument for some form of industrial policy or trade protection: the recognition that the national value of manufacturing often lies not so much in the end product itself, but in the accumulated knowledge that goes into it, and the possibility of old processes and knowledge sparking new innovation. Of course, innovation is itself what killed the high-end cassette player. But many otherwise viable industries have struggled under the free-trade regime.
The fact is that technology is not embodied in a drawing or set of drawings or any set of instructions. It is embodied in human knowledge. One of the key problems in the industry is the loss of control a customer or prime has when they let a contractor develop the ‘data package’ and ‘product’ with no significant oversight. While the customer or prime may ‘own’ the IP because they paid for it, the fact is that the majority of the capability is embodied in the people and culture of the contractor not in any set of information.
The Hellenic world had machines as complex as early clocks and steam engines of a sort but lost the recipe in a few generations or less. Various complex building skills and wooden machines, metalworking and early chemistry were discovered then lost again and again because the data package was in human brains and examples. This is why the printing press and its ilk were so incredibly important to technological lift off. Along with a culture of progress and invention.
We are far ahead of that world but as above, not above losing the recipe of a complex technology. This is one of the drivers behind Computer Aided Design, Analysis, Documentation, Fabrication. Our cybernetic tools have the ability to record the data package in detail at least for certain classes of things so that we should be able to maintain the ability to replicate things. Making special, small run, even one off technological objects rational rather than nutty.
But at the same time I think that it is likely that the artisanal ethos and products will remain relevant and even increase in value as people shift away from a mind/economy/culture of scarcity to at least sufficiency and if we survive and expand into the universe eventually richness. These transitions will be extremely difficult because they are at odds with many tens of thousands of years of genetic/mimetic coding of our behaviors based on small group hunter gatherers and kin group bonding. Those transition will be enabled by machines that fabricate, even machines that invent. What will happen when humans loose the recipe for technological advancement, because too few engage in the complex enterprise of development??? Is that the point of the Rise of the Machine???
These craft and others such as craft like the voyagers continue to return immensely valuable data long after their primary mission is complete. One of the things NASA and other space science organizations struggle with is supporting these ships long after the original funding timeline is past. This is a great problem to have and by and large the money is found since these are very cheap deep space projects in the big picture.
So my title, the economy of ‘outer space’ is all about data, science, prospecting right now. These are valuable assets that we need to support to provide returns orders of magnitude greater than the cost in the sense of other ways of getting that data, data that is both live affirming in its fascination and valuable as part of the bedrock of our understanding of the universe.
A really cool concept. Miniature walking rovers that can explore tiny spaces, a single test to the moon this year with plans for swarms (small ones) in the not too distant future. Tbe video animation from Spacebit is worth a couple of minutes.
The Boston Dynamics robots are at the point that they can do most things a human can in regards to locomotion. It is unclear how much beyond balance and moving is local to the robots as the thoroughly bounded arena makes clear but the basics of the body frame is there. Ability to manipulate the environment other than in the most basic way has not been demonstrated by Boston Dynamics but other companies are making huge strides in manipulators. Ability to sense and understand the environment is another huge step. Except that the sensors exist (autonomous cars etc). Leaving understanding the environment beyond a very limited ‘world.’ And that takes a brain, and that seemed a long way off….except is it?
GPT3 would appear to be on the threshold of general purpose artificial intelligence. In the article it is noted that GPT3 is a brain in a box with no ability to sense or manipulate the environment without human intervention. But ‘wrapping’ those abilities ‘around’ GPT3 appears all but trivial. Given its ability to learn on its own would a Boston Dynamic’s wrapped GPT3 become something close to the robot of our dreams and nightmares. It certainly appears so.
Atlas’s is battery powered, I think, to the tune of an hour or so. GPT3 is instantiated on a huge computer network but both of those limitations are receding every day as computing power and battery storage continue to improve driven by their broad application across the tech scene.
Five years from now it would seem likely that the general purpose android robot will be a real thing. If built in quantity like say a Tesla 3 are you looking at $30K a pop? What does that lead to?
I want to make sure they understand that I for one welcome our dancing robot overlords.