# Making \*TOPIC\* atoms a special case (trust circle + curator paradigm)

**URL:** <https://atlas.discourse.group/t/making-topic-atoms-a-special-case-trust-circle-curator-paradigm/1258>\
**Category:** 2. Semantic Engineering\
**Created:** [April 16, 2026, 2:38am UTC](https://atlas.discourse.group/t/making-topic-atoms-a-special-case-trust-circle-curator-paradigm/1258 "2026-04-16T02:38:10Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![smilingkylan](https://sea1.discourse-cdn.com/flex019/user_avatar/atlas.discourse.group/smilingkylan/32/1135_2.png) [@smilingkylan](https://atlas.discourse.group/u/smilingkylan)\
**Post date:** [April 16, 2026, 2:38am UTC](https://atlas.discourse.group/t/making-topic-atoms-a-special-case-trust-circle-curator-paradigm/1258/1 "2026-04-16T02:38:10Z")

</div>

With Hive Mind’s decision to pursue trust circles as the preferred user experience it begs the question of how best to structures the related atoms. The following is our idea:

`[address] - trusted for - [topic]`

1. `address` - self-explanatory
2. `trusted for` - I’m proposing to use the TextObject version of “trusted for” because it most succinctly conveys the relationship while being a unique enough phrase that is unlikely clash with other meanings  
Term ID: **0xe55923557a4bc0ce15f91945eb2d4fca92e1a6fa8074378bcd8400eeb8b79176**
3. `topic` - again I propose using the TextObject version of the topic name “Sports” or “Crypto”, specifically capitalized to distinguish it from any other atom that happens to use that same word.

### Topic Capitalization

I intend to give topic atoms distinction from other types of atoms because they will play such a pivotal role in the trust circle ecosystem (along with curators). For that reason I am proposing to use first-letter capitalization as the convention for a topic atom. So instead of “tech” it would be “Tech” or “sports” would become “Sports”… this matches convention of this atom representing the NAME of a topic.

### Atom Type TextObject vs IPFS vs stringified JSON

I heavily favor TextObject type atoms because they force users to be much more concise, they often save on visual space on the front-end. Additionally, they encourage atom creators to choose unique strings as their identifiers (nested data structures breaks this expectation) which makes them easier to search for.

### Topic Curator Self-Nomination

I have mentioned this elsewhere but I am also proposing to use `[address] - trusted for - [topic]` for individuals to declare themselves as a curator for that topic by staking on this triple for their own address. So if @smilingkylan wants to become a curator for “Sports” then he will stake on the following triple:  
`smilingkylan.eth - trusted for - Sports`

Let me know your guys thoughts on this semantic structure for trust circles + curators

---

<div class="post-metadata">

**Author:** ![billy](https://sea1.discourse-cdn.com/flex019/user_avatar/atlas.discourse.group/billy/32/3_2.png) [@billy](https://atlas.discourse.group/u/billy)\
**Post date:** [April 21, 2026, 12:54am UTC](https://atlas.discourse.group/t/making-topic-atoms-a-special-case-trust-circle-curator-paradigm/1258/2 "2026-04-21T00:54:30Z")

</div>

Hey @smilingkylan - we’re also workshopping this right now.

We should have some updated thoughts on this from our end by EOW.

In the meantime, if you haven’t checked it out, here is a repo with some of our most recent data structure thinking:

> **[GitHub - 0xIntuition/intuition-data-structures](https://github.com/0xIntuition/intuition-data-structures)**
>
> Contribute to 0xIntuition/intuition-data-structures development by creating an account on GitHub.

Also, some questions for you in the meantime:

1. What are some other Predicates besides [trusted for] that may be appropriate here? Would love to work with an inventory of creative thoughts for what we’ve got - [trusted for] is nice and concise and relatively unambiguous, but not perfectly unambiguous… Spitballing some other variations: [trusted in], [trustworthy in] [trustworthy for], [trusted in the context of], [trustworthy in the context of], [trusted in context], [trustworthy in context], etc.
2. The Topic Atoms are actually extremely important, in my mind. These can be generally applied across every application in a variety of places - i.e. tagging things to categorize them, etc. - so, of course, anyone can create a new Topic and it would be great to have some sort of standardized Topic format that we all conform to, but it would also be great to have a standard Topic ontology. Do you have any thoughts on the initial Topics you might want to start with?