Miscellaneous notes on the relationship between the legibility of Great Society and the automation thereof by computer.
George argues AI and automation are at odds:
… the vast majority of use-cases for AI, especially the flashy kind that behaves in a "human-like" way, might be just fixing coordination problems around automation.
AI, from this perspective, is something like “the computational overhead of metis”.
Thus we end up with rather complex jobs; Where something like AGI could be necessary to fully replace the person. But at the same time, these jobs can be trivially automated if we redefine the role and take some of the fuzziness out.
A bartender robot is beyond the dreams of contemporary engineering. A cocktail making machine, conveyer belt (or drone) that delivers drinks, ordering and paying through a tablet on your table... beyond trivial.
I would like to return to this point. Is legibility just the simplest thing?
Policy and Statistical learning
TODO. Brief digression on how legibility and management looks as a statistical learning problem. We know that constructing policies is costly in data, and we know that administrative procedures frequently do not have much data from repeated trials of what works. We also know that coming up with policies (in a machine learning or in a political definition) is computationally challenging and data hungry. How does the need to bow to the ill-fitting bureaucracy of the Great Society resemble having to work with an underfit estimator of the optimal policy? What does that tell us about, e.g. optimal jurispudence? Possibly something. Or possibly the metaphor doesn’t work; after all, what is the optimisation problem one solves?
Lanier (2010) has a notion about “post-symbolic communication” as something that exists beyond the symbolic communication that modernity’s legibility favours, and I suppose the “pre-symbolic communication” possibly in the metis regime.
Suppose we had the ability to morph at will, as fast as we can think. What sort of language might that make possible? Would it be the same old conversation, or would we be able to “say” new things to one another?
For instance, instead of saying, “I’m hungry; let’s go crab hunting,” you might simulate your own transparency so your friends could see your empty stomach, or you might turn into a video game about crab hunting so you and your compatriots could get in a little practice before the actual hunt.
I call this possibility “post symbolic communication.” It can be a hard idea to think about, but I find it enormously exciting. It would not suggest an annihilation of language as we know it—symbolic communication would continue to exist—but it would give rise to a vivid expansion of meaning.
This is an extraordinary transformation that people might someday experience. We’d then have the option of cutting out the “middleman” of symbols and directly creating shared experience. A fluid kind of concreteness might turn out to be more expressive than abstraction.
In the domain of symbols, you might be able to express a quality like “redness.” In postsymbolic communication, you might come across a red bucket. Pull it over your head, and you discover that it is cavernous on the inside. Floating in there is every red thing: there are umbrellas, apples, rubies, and droplets of blood. The red within the bucket is not Plato’s eternal red. It is concrete. You can see for yourself what the objects have in common. It’s a new kind of concreteness that is as expressive as an abstract category.
This is perhaps a dry and academic-sounding example. I also don’t want to pretend I understand it completely. Fluid concreteness would be an entirely new expressive domain. It would require new tools, or instruments, so that people could achieve it.
I imagine a virtual saxophone-like instrument in virtual reality with which I can improvise both golden tarantulas and a bucket with all the red things. If I knew how to build it now, I would, but I don’t.
I consider it a fundamental unknown whether it is even possible to build such a tool in a way that would actually lift the improviser out of the world of symbols. Even if you used the concept of red in the course of creating the bucket of all red things, you wouldn’t have accomplished this goal.
I spend a lot of time on this problem. I am trying to create a new way to make software that escapes the boundaries of preexisting symbol systems. This is my phenotropic project.
The point of the project is to find a way of making software that rejects the idea of the protocol. Instead, each software module must use emergent generic pattern-recognition techniques—similar to the ones I described earlier, which can recognize faces—to connect with other modules. Phenotropic computing could potentially result in a kind of software that is less tangled and unpredictable, since there wouldn’t be protocol errors if there weren’t any protocols. It would also suggest a path to escaping the prison of predefined, locked-in ontologies like MIDI in human affairs.
I am not convinced, for reasons I might go into at some point.
Self fulfilling prophecies in
I have a sickening feeling someone will say something about nonstandard decision theories, but let us mention this anyway.