The annual ESA meeting functions in a lot of different ways. There are the obvious: the sharing of ideas and work, the discovery of new ideas, methods or sources of inspiration, networking and job finding, social reunions. But it also functions as a kind of report on the state of the field (and that's not even considering sessions meant to explicitly do this, like the panel “Conversations on the Future of Ecology”). The topics and methods presented say a lot about what ideas and methods are timeless, what is trendy, and over many meetings, where ecology appears to be going. If you go to enough ESAs, you are participating in a longitudinal study of ecology (or at least your subfield).
I went to my first ESA five years ago in Albuquerque, NM. One of the things that struck me was that there were two Community Assembly and Neutral Theory sessions and many talks in those focused on tests of neutral theory, particularly looking at species abundance distributions (SADs) and various iterations of neutral models. There are usually still one to two sessions called Community Assembly and Neutral Theory, but five years later, I don't think I saw a single talk that looked at SADs for evidence of neutral theory (and only one or two talks that were named to explicitly include neutral theory). Instead, the concept first introduced by Hubbell has morphed from "neutral theory" in to something slightly more general, designated "neutral dynamics". This gets used in a lot of ways – most precisely, neutral dynamics are in the spirit of neutral theory, suggesting that population demographic rates are similar, allowing long-term co-occurrence. Sometimes this is cited with reference to equalizing fitness effects in a Chessonian framework, where similarity in fitnesses prevents exclusion despite overlap in species niches. But it also seemed to get used in a default sort of way, as the explanation for why niche differences between species weren't discovered by a study, or else "neutral" was used interchangeably with "stochastic". In any case, the pattern appeared to be a move from highly specialized and precisely defined usage of the term, to broader incorporation of the concept that had suddenly acquired several, often less precisely defined meanings. Instead of being the central focus of a few specialized talks, neutrality was commonly invoked as a minor theme or explanation in many more talks. It is not what I expected, but its continuing usage suggests that neutrality has developed a life of its own.
Other topics similarly seem to have taken on separate lives from their initial application; even over the short time I've been attending ESA. For example, sessions focused on simple applications of ecophylogenetics methods (overdispersion, clustering, using different systems) were relatively common 3-4 years ago, while there wasn't a single contributed session specifically named for phylogenetics this year. There was however many sessions in which phylogenetic work formed the backbone of talks that were about broader questions, including in the "Evolution, Biodiversity, and Ecosystem Function" session and the “Coexistence of Closest Relatives: Synthesis of Ecological and Evolutionary Perspectives”. In the best case scenarios, it seems like even over-hyped approaches may be used with more nuance in time, as people recognize what information these methods can and cannot provide.
Sometimes it did seem that there is a lag between when critiques of certain methods or ideas are expressed and when they actually get incorporated into research. I could be wrong, but it seems this is most common where the research is focused on particular study systems or species, and methodology may be driven more by precedent in the literature and criticisms may take longer to infiltrate (since they aren’t the main focus of the work anyways). And unfortunately, the topics and sessions which appear to be timeless are those on human-related applications (restoration, climate change, invasion). Those pressures are sadly unchanging.
*The great thing to do would be map out changes in keyword frequency over the ESAs that have archived programs. Unfortunately, I don’t have the time/motivation.