Like many changes, stuff doesn't happen until there's somebody that has the time, the means, and the intent to push for it.
The familiars were left as they were for a long time because other devs had their own projects to tackle, and familiars were a low priority - not to mention, and as evidenced, nerfing the pixie would be met with some community backlash.
Nerfs, however necessary, are largely unpopular, so it's not particularly alluring to update a system that is going to be received negatively.
I can tell you my rationale for pushing for it -
And that is simply that the pixie was too good.
You'd have to shoot yourself in the foot if you wanted to use a pseudodragon, because the option between a funny talking lizard that is only a cosmetic accessory, or a literal pocket rogue? You'd pick the pixie every time unless you wanted to make a statement.
In a RP server, having one mechanical choice that is objectively superior than the others is a design failure, and something that needed to be corrected as a matter of principle.
The consequences that it brings - having to gear up/invest ranks/rely on a friendly rogue - are all acceptable to me.
As an old time mage player, and one that is very fond of familiars, I wanted a system that gave more space for variety and flavor, for thematic choices, rather than those purely mechanical. So this is an agreeable result.
You could also observe "but why is the system not doing X, or also doing Y", and that is a matter of scope - the longer, bigger, and more nuanced a system is to design, it also becomes harder to write, to review, to approve, and to bugtest. I didn't want to code a mechanical monstrosity of a system that would take a small manual to understand, get stuck in review hell for 3 years, or spend the next 5 months bugfixing.
Even with this rather simple system we are seeing a lot of bugs related to language (that I am tackling as I figure why they are happening).