Satoshi added many mempool filters himself, also mentioning spam explicitly (famous Lady Gaga discussion with Andresen). Virtually all Core developers for years agreed with and built on this approach, including current champions of the removal path (you can find mentions in this sense of Sipa, Gmax, Achow), often mentioning spam explicitly, with the main exception of Todd, who for many years argued for a mempool=blocks fee-rate-based approach, now culminating in the LibreRelay fork. To some degree, full-RBF was an instance of this debate (while not spam-related) where most devs (including Luke-jr actually agreed with Todd. Slowly, more people moved on Todd side (the only explicit reference I've seen so far about the change of heart is the tweet from Achow about "growing up").

In 2021, an unintentional bug shipped with Taproot (but not connected with Taproot itself) defused one such filter (datacarriersize) for some txs. In 2023, shitcoin scammers abused the witness discount to create a 4Mb block (impossible to create with non-spam txs), using txs that would have bypassed the filters thanks to the bug if broadcast (but it's noteworthy that the spam was not broadcast, it was sent directly to a colluding mining pool, paid offband, so even non-broken filters wouldn't have stopped it). After that, it seems like an important priority for the current Core mantainers is to move Core towards a more LibreBitcoin-style. Personally, as an illiterate pleb who listens to podcasts, I consider the technical arguments in this sense legit and overall directionally sound, even if they don't justify the prioritization/rush, especially considered the widespread opposition by many power-users, builders, entrepreneurs, advocates/activists.

Some people, including myself, are under the impression the the rush/priority is more driven by ego struggles, social dynamics and power games which have little to do with technical mempool policy debates: avoiding (inevitably-painful) self-reflection on the 2021 bug, justifying social proximity by many current contributors to the shitcoin scammers involved in this kind of spam (eg: Rijndael), not hurting feelings of colluding mining pools currently funding Core development but also monetizing OoB payments for spam (eg: Mara), regulating old grudges with maverick/outcast/contrarian devs (eg: Luke-jr, especially after the conflict with him escalated since the FBI hinted at a Core meetup as the origin of a nasty cybercrime at Luke's expenses), venting envy by many contributors that said maverick/outcast/contrarian devs received investment money (eg: many of them wrote to Dorsey publicly to try to convince him to remove his support to Luke-jr efforts towards mining decentralization), overall political devides (including a Bitcoin version of the overall "DEI" and "woke" related cultural wars).

Regardless of my (and other people's) interpretation (cf last paragraph) of the reasons mempool policy tweaking is considered a priority (I have to add there are two weak but not-irrelevant arguments for that: Todd made a case a mempool=blocks approach is safer for L2 unilateral exits, some devs made the case that because of the current spam attack we should actively engage with harm-reduction by opening up the less damaging attack surfaces like op_return to hijack some pressure from the more damaging ones like UTxOset), the effect of the conflict was also a very bad management of the GitHub repo, with obviously good-faith, balanced, nuanced, polite and insightful comments hidden as "spam" (ironic), disagreeing parties with lots of skin in the game banned from the repo, contentious PRs closed only to be insta-reopened to let some ACK from friends sneak in and close them again, etc. That also escalated further the conflict.

There are some things happening off-line as well which escalated things: for example some fairly neutral but not "aligned" developers have been excluded from in-person Core meetups, after having been criticized for not taking a strong enough position in this debate. Some of us get to know about these things from personal conversations, but the involved parties often ask not to publish details, in order to avoid further retribution. So we end up with some public outspoken figures (including for example Mechanics) raising red flags about an overall Core problem, but most viewers not being able to verify any details. Then it becomes a religious/tribal choice of "camps": every argument on one side is a "lie". On my personal account, I can say that some philosophical issue (people claiming absurd things about the nature of spam or censorship) triggered me more than the policy debate itself. I could have kept more cool if less people called filtering "censorship" (utter self-contradictory nonsense) or spam "impossible when valid/expensive (utter self-contradictory nonsense).