It feels intuitively weird to you because you are a monolingual Indo-European.
In fact, these two circumstances are extremely different!
In one, there's a definite event, in which an agent performs an action that moves an object.
In the other, there is no event - there's a process. And it's not a definite process, but a generalisation. There is no agent. There is no action, properly speaking. And the object is not moved. On the other hand, the object IS altered, which it isn't in the first case. Semantically, these two propositions are utterly unalike.
And your second case is still at the more 'transitive' end of what inanimates can do - it's part of why many languages have animate words for 'water' (or two different words, one animate and one inanimate). It gets worse!
Consider instead: "water dissolves limestone". Here we have an additional point of overwhelming different: both objects are fundamentally altered by the process. This is very different from a prototypical intransitive (neither object is altered) and from a prototypical transitive (one object is altered while the other is unaltered). So really there's three kinds of relation: pure transitives like "the man smashed the skull", pure intransitives like "the man observes the rabbit", and what we might call 'symmetricals' like "the ink dissolves into the water". Indo-European languages tend to want to compress all three semantic forms into a single syntactic pattern - so, for instance, "the man observes the rabbit" is more grammatically similar to "the man eats the rabbit" than it is to "the man observes". But many languages remain truer to the semantics, and don't conflate all dyadic relations into a single prototypically transitive construction! Particularly languages with a concept of animacy...
And in the particular case of these 'symmetricals', it's worth noting a real peculiarity: subject and object are arbitrary. If the semantics are purely symmetrical, they're completely arbitrary: "the handle is glued to the case" and "the case is glued to the handle" are semantically equivalent. But even in asymmetrical reations, the subject and object can just be flipped by using the opposite verb: "the wall supports the roof" and "the roof burdens the wall" are semantically equivalent, they just have verb suppletion to invert subject and object.
English happens to often prioritise one verb in each pair. So in English, the moat encircles the castle, but the ocean subsumes the raindrop. But that's arbitrary: neither argument is semantically more the subject or object than the other, and other languages can use the opposite verb.
Similarly, in "the electron absorbs the photon", it's misleading to think of the electron as agent and photon as patient, because both are equally affected by the action, and neither is more responsible for it. It's a semantically balanced process (maybe 'balanced' is a better word than 'symmetrical'). A different language could just as easily have "the photon enters-and-utterly-pervades the electron". Or could use an entirely different sort of construction.
This is unintuitive for you because you are a monolingual Indo-European
The "ease" that arises from English using the construction that is natural to use in English arises entirely from that construction being natural to use in English. Speakers of languages where you just say "photon electron qua", where "qua" is a posposition meaning "two uniting as one that acts as the second while the former appears to disappear", will find that the weird rigmarole with which English first pretends there's an action taking place, then requires a verb for it, then requires one item to be prioritised over the other, then requires the 'superior' item to be treated as though it were a person and the 'inferior' item as though it were an object that was altered but not removed from existence by the 'action', to be a ridiculously un-easy and misleading metaphorical construction for an ultimately simple event!The ease with which English expresses notions like "the electron absorbed the photon" makes it well-suited to scientific writing. Without that ease of use, I feel like writing about physics or any field where inanimate objects regularly interact would become rather onerous.
