Simply put, UX copy should *never* be tied to technical implementation, reflect technical debt, or simply be left "as-is" to atrophy in-product as feature areas evolve for the sake of documentation pruning efforts.
If the feature area is "properly lit" with supportive UX copy that informs users what a feature set does and how to engage with it, old support documentation becomes irrelevant and newer support documentation can be updated and simplified for the new approach ... eyes v10 release 😉
Since you're looking for thoughts on this topic, my approach would be to explore the user actions, the whys, that lead to the desire to "filter" thoughts in the first place. Search is one tool, it does what it does, "add search results to selection" is another tool (driven from search box context) and it does what it does. Both of these function quite well for situational ephemeral tasks.
This "reports" feature set in my mind could begin to push towards a far richer "dynamic filtered brain view" that I'd love to see further explored. I've done a ton of user experience envisioning in this area and find it a quite unexplored space as presently implemented.
The image below is a quickly doctored and focused glimpse into one portion of a product UX exploration that my partner and I were previously working through in which we landed upon a "curation" conceptual data structure which would be quite powerful if implemented well:
The idea herein would be that a collection of "filtered thoughts" could be gathered together (based upon their underpinning relationships and/or based upon specific secondary meta-search criteria similar to what is currently expressed in the reports section) and then this "curation" construct could then be pinned and/or saved as its own top-level "thought" node to recall and/or re-invoke said bespoke filtered thought collection at a later time.
The potential power of TheBrain's "reports", historically speaking, has largely been lost on me personally as a TheBrain user as they feel quite ephemeral as implemented which the previously noted search and add to selection features already handle quite efficiently. The idea of saving and/or sharing a "filtered view" of my brain with my future self or another user via an online brain by way of a single "curation thought node" makes a good deal of sense to me when driven by a specific need to share a subset of my brain.
To personalize a hypothetical user experience journey of this curation construct a bit further ...
... let's say, for example, as an author, I've generated a collection of top level "thoughts" within my brain to act as a compendium for my latest publication. The publication itself traverses (in a traditional linear manner) a great deal of content within my brain. In order to give a new reader a more dynamic and multi-dimensional glimpse into the highlights and/or a chapter by chapter compendium dataset to consider as they read, I'd like to assemble a curation from my master brain that serves to anchor a reader's journey into my broader brain by way of a tailored subset that I deem relevant. Knowing that my curation has some some dynamic filtration criteria, I can publish my curation once and trust that as I continue to build out my brain over time, new and relevant thoughts will continue to surface for future readers as they visit my compendium brain via this curation node.
Anyhow, a few rough bits to consider ... anyone inclined to dig in a bit deeper on this with me, please don't hesitate to reach out directly 👍