Would a mapping where sensors / motors / preferences (or maybe even beliefs, partially) are considered as not-part-of-the-agent and, instead, their agent-facing inputs/outputs are considered as actuators/perceptions, be more simple and, thus, more plausible?

I meant "interest" in the sense of "enjoy interaction with" or "enjoy a relationship with".

I think the problem here is one of motivation to self-modify. For example, it's one thing to want to self-modify to like spicy food (possible, but potentially unpleasant) or become bisexual (possible for some, probably, but not for others), but other self-modifications are less desirable - for example, I wouldn't want to be more interested in "normal people" even if it would increase the number of people with whom I'd have relationships.

Side safety note: always "set" a time-limit on your self-hacks (some pre-defined moment at which you force yourself to resurface from your compartmentalized-mode / hacked-end-goals / inconsistent-state and re-consider the usefulness of it before diving back in with yet another timelimit if deemed worth it).

Also, this use of the "terminal goals" is still making me uncomfortable (even if it's not "terminal values"), especially with a sort-of-disclaimer that is not on the top. Would be quite nicer to make up a different term for it (e.g. "end-goals").

And, in general, with any successful self-motivation (especially with these methods that can, indeed, be super-effective) you are running into increased danger of pursuing mostly-useless end-goals.

Bridge hypotheses don't seem to me to be the kind of thing one ought to even try to get rid of. The idea of having a physical layer is a powerful one, and you'll need to bridge from it to your perceptions.

On thought, my response is that no circular argument can possibly be rational so the question of if rationality is binary is irrelevant. You are mostly right, though for some purposes rational/irrational is better considered as a binary.

AIXI ignores models of the world and only knows about perceptions. There is an analogous agent who only "knows" about the world. For this agent, every perception is represented as a change in the model of the world.

"Better" isn't a function of the real world anyway- I'm appealing to it because most people here want to be rational, not because it is objectively better.

What do you mean by "rational" is not a binary?

If you have no non-circular basis for believing in induction, surely it is irrational?

Could the “correct” bridge hypothesis change if part of the agent is destroyed, or, if not, would it require a more complex bridge hypothesis (that is never verified in practice)?

For an agent that can die or become fully unconscious, a complete and accurate bridge hypothesis should include conditions under which a physical state of the world corresponds to the absence of any introspection or data. I'll talk about a problem along these lines for AIXI in my next post.

It's similar to a physical hypothesis. You might update the hypothesis when you learn something new about death, but you of course can't update after dying, so any correct physical or mental or bridging belief about death will have to be prospective.

Is it supposed to be possible to define a single “correct” bridge mapping for some other agent than self?

I'm not sure about the 'single correct' part, but yes, you can have hypotheses about the link between an experience in another agent and the physical world. In some cases it may be hard to decide whether you're hypothesizing about a different agent's phenomenology, or about the phenomenology of a future self.

You can also hypothesize about the link between unconscious computational states and physical states, in yourself or others. For instance, in humans we seem to be able to have beliefs even when we aren't experiencing having them. So a fully general hypothesis linking human belief to physics wouldn't be a 'phenomenological bridge hypothesis'. But it might still be a 'computational bridge hypothesis' or a 'functional bridge hypothesis'.

Is the location of the agent in a world a part of the bridge hypothesis or a given?

I'll talk about this a few posts down the line. Indexical knowledge (including anthropics) doesn't seem to be a solved problem yet.