> Purposes are deduced from behavior, not from rhetoric or stated goals.
A light rewrite of the slogan "the purpose of a system is what it does".
I hate that slogan. It's supposed to be profound, but it's actually really, really dumb.
The whole point of the word purpose - its purpose if you will - is to be able to talk about the difference between what a system does and what it ought to do. To say the purpose of a thing is what it does, is to render the word completely redundant.
If the stated goals of a system don't match what it does, then either you have a dysfunctional system that needs to be fixed, or those stated goals are lies.
In these days of AI, it's all the more important to think clearly about purpose. Just because a system does a thing, there's not necessarily anyone who wants that thing. And there's definitively not any "spirit of the corporation" or "spirit of the nation" or "spirit of the language model" which wants it. Only persons want things.
If it bothers you, you could work around the semantics and talk about its empirical purpose or observed purpose, vs intended purpose, rather than using purpose and expecting it to be interpreted in a particular way.
I think the jarring alternative definition of "purpose" is an intentional mental hack: to break people out of orienting themselves around design intent, when processes often go out of alignment with that. It's easy for humans to fool themselves. Taking a dispassionate, empirical view when considering what a process is doing, or analyzing what its range of behaviors is, can illuminate many problems.
They do not need to be lies, but they may be. They may also just be a mismatch between the level a statement is generated at and the level it is consumed at, the “level” here being something we could call a system’s functional context.
Additionally, a statement is generated by a system and reflects itself a separate purpose in its generation. A system can exist without statement of purpose, so why was it made? The statement has an intent and an audience. It is therefore necessarily more constrained than the system itself, and for the systems we are referring to, it is markedly less complex than the system itself. It cannot hope to capture the full function, which is what can serve (or not) a purpose.
I’ve always interpreted this saying to point in the direction you’re actually going. That a machine does not have a “purpose”. Its creator might have had a purpose in mind, but when talking about the machine one should focus on what it actually does. This avoids needing to fend off attacks about how noble the creator’s intention for their machine may or may not have been, which could otherwise consume the rhetorical oxygen better spent.
Yes, that's part of the point. If the stated purpose of a system is to "improve vehicle fuel efficiency" but the observed behaviour is to give car companies a way to avoid doing much of anything about fuel efficiency, or if the stated purpose is to "encourage EV adoption" but the observed behaviour is "funnel money to certain rich people" then you are better off assuming that the behaviour you observe is the intent, and the publicly stated purpose is marketing spin.
On the other hand, things change over time. Google started out organizing the world's information, and gradually turned into surveillance advertising. They got rid of "don't be evil" so there are no explicit lies. They aren't a dysfunctional search company, they are a functional surveillance-advertising company, that thing they do - update your expectations so you believe that's their purpose and don't believe that they are "trying to be good but misguided".
The quote you replied to is used to educate a reader how to understand an existing system. For example, it helps one understand why a system is in a state when as you say 'there's not necessarily anyone who wants' it that way. Rivers are polluted because there are some factors in the system which enable it.
Also, as the video I linked in the previous comment says:
'our first instinct is to blame people in the system and that is a low leverage response...actions people take are possible, feasible, and rational because of the universe in which they exist.'
I am not sure if you've read the book. But even though you've not used terms like feedback loops, etc. I think what you are saying is not much different from what the book says.
> The whole point of the word purpose - its purpose if you will - is to be able to talk about the difference between what a system does and what it ought to do. To say the purpose of a thing is what it does, is to render the word completely redundant.
as it takes normal every day systems that we think we can depend on (banks, the army, phones) and sarcastically highlights some of the worst things those systems do vs what we're told they do. The chorus is that slogan.
A light rewrite of the slogan "the purpose of a system is what it does".
I hate that slogan. It's supposed to be profound, but it's actually really, really dumb.
The whole point of the word purpose - its purpose if you will - is to be able to talk about the difference between what a system does and what it ought to do. To say the purpose of a thing is what it does, is to render the word completely redundant.
If the stated goals of a system don't match what it does, then either you have a dysfunctional system that needs to be fixed, or those stated goals are lies.
In these days of AI, it's all the more important to think clearly about purpose. Just because a system does a thing, there's not necessarily anyone who wants that thing. And there's definitively not any "spirit of the corporation" or "spirit of the nation" or "spirit of the language model" which wants it. Only persons want things.