Typically in comparing Latin with modern Romance, the merger of neuter into masculine is one of the loss of distinguishing suffices. You can't distinguish between bonus and bonum if the final constant is lost and they are both bonu (or bueno in Spanish).
But some words didn't distinguish masculine from neuter by just a consonant. Take iste (m) and istud (n). Remove final consonants and you have iste and istu. So in Spanish there's este and esto. The different vowel gets preserved, so you can distinguish between masculine and neuter still, where with most words you cannot.
Early Proto-Indoeuropean is theorized to have 2 genders - animate/inanimate, but by the late PIE period it evolved to the classic three.
Of course, this is largely theoretical like you said, and I'm sure you could dig up some dissenting papers if you wanted to.
For example, there are some adjectives in Latin and Ancient Greek, which only have two declensions instead of three: one common declension for masculine/feminine, and one for neuter.
For example: naturalis (masc), naturalis (fem), naturale (neut), as opposed to novus (masc), nova (fem), novum (neut).
Edit: Oh, I translated this as alive/not alive. Guess, this was originally down to soul. Not advocating to go down that path again…
"Champagne", "Vitrine" as exemple.
The ""new"" gender neutral a woke fashion that has disregard for people with dyslexia or people who needs software to read (blind).
It's also worth noting that grammatical gender evolves over time. Originally, most Indo-European languages had the three classes of feminine, masculine, and neuter. But, for example, French's loss of the neuter class is pretty recent. We're also seeing modern efforts to drive these processes with, for example, the usage of the gender-neutral term "Latine" instead of "Latino" or "Latina".