This word does not exist

Do you do inference on every visit?

What is the cost of being on the front page?

I'm procrastinating on other stuff, so I'll elaborate, but not because I have any responsibility to.

Both you and the author are guilty of using imprecise language as if it were precise language.

>imposing activity, as opposed to forbearance from activity.

When you use imprecise (ie natural) language, what you are really doing is using generic language. The natural language is a template for deciding whether a given formal language construction can replace it. I claim that I can choose a definition for activity, such that these two concepts two are logically equivalent, that is heuristically compatible with the natural language. For example:

Time moves forward, and whatever you do while that happens is your activity. the ways to require me to perform some activity in a set of activities are equal to the ways to require me to perform some activity in the complement of some set of activities.

In fact, I think that you might have some trouble defining activity in such a way that those two things are different, but the point is that there is ambiguity.

This isn't a minor point. The author uses natural language constructions as if they were logical constructions all over the place, and then further builds propositions out of them as if he were using formal constructions.

>a positive-sum maximization of individual agency and control over their lives to the greatest extent possible short of anyone’s agency infringing on anyone else’s, is universal consent of the governed....

A sensible logical interpretation here is that all of the agents must be completely isolated from each other. Each one is maximizing their own utility in their own black box. If they weren't, then they would each be naturally infringing upon each other's agency to not be observed by each other by observing each other. We have explicitly excluded that case, so we must be talking about isolated agents. In this case I agree with the premise, even. The ultimate in democracy is indeed the (pointwise) maximization of each agent's utility. As a logical conclusion, this is fine. It more or less follows from the assumptions and I have no immediate problem with it.

However, if we are asked to additionally assume that these people live on earth and are subject to earthlike constraints. This clearly violates the premise, as you even point out yourself.

"Oh no", a strawman might say, "it is bad faith to instantiate these natural language definitions with example formal definitions that yield contradictions!"

What else am I supposed to do with informal reasoning? But fine, supposing that we really do have a good way to define "infringing on each others agency" so that people can live on earth together without doing it. What even is the sum of their agencies or whatever measure is being used? In my earlier example, I just needed a partial order, and the agent's isolation from each other was enough to guarantee me a maximal element, all of a sudden I have to both combine different people's utilities into a single number and then order the result of that combination? And this supposed ordering happens to have a generic maximal element whenever no one is infringing upon anyone elses agency? I would describe this as unintuitive.

If the author has an argument for or proof of this fact, they better make it more explicit than this, rather than just talking loosely about how ant colonies solve pathfinding problems (stigmergy).

On a more constructive note it is true that I disagree with the author's premise. It seems that the anarchists solution to the prisoner's dilemma is either to simply pretend that its an edge case and doesn't happen often enough to worry about, or to redefine social utility to prefer the situation where both playes defect because it is a Nash equilibrium, and therefore must be ideal.

Going even further, I can get universal consent to impose a policy of punishing defectors. This is a constraint on free agency, however, so universal consent and "free agency maximization" have nothing to do with each other.

I'm not sure what you're referring to by "valent phonemes." The standard treatment of a syllable, not only in English but cross-linguistically, is that a syllable is composed of an cluster of consonants at the beginning (the onset), a single "nucleus" phoneme that is most often a vowel or, somewhat less often, a sonorant consonant like 'r' or 'l', and another cluster of consonants at the end (the coda). The specific ways that syllables can be formed in a particular language are governed by the "phonotactical rules" of that language. These are rules like the English rule that an "ng" sound may only occur immediately after the nucleus, or like the Japanese rule that a coda may only be "n" or "" (the null coda).

a simple blocklist might be just the cheaper and easier solution. After all, blocklists were, and still are being used to filter human output as well. And fail-safe. A person with a hobby project does not have a legal/pr department to deal with the consequences of AI having a bad day

I was kind of referring to valence electrons - so think “edge phonemes” - but I don’t think that came across.

This is a really interesting explanation of something that I kind of understand, but want to have a deeper understanding of. Any good references you can think of offhand?