On April 12, 2025, Ethos added live LLM-based analysis to the process of writing reviews and vouches. The feature was built to encourage people to add more detail to what they write about others while they are still writing it.
For anyone who reads reviews on Ethos to decide whether to trust an account, this matters. A review with specifics says more than a short one, and the change aimed to get more of those specifics into the network.
A late-night addition
Ethos announced the feature on X with a nod to the hour: “thought we were done for the night? the devs don’t sleep.” The team presented it as one more change shipped after it had seemed the night’s work was finished.
What changed
Ethos is a network where users write reviews and vouches. Reviews and vouches are the raw material of that system.
With this update, an LLM analyzed reviews and vouches live, during authoring. The purpose, as Ethos described it, was to help encourage users to add more detail before they submitted.
Ethos’s post did not go further into how the analysis worked or what it showed users, beyond that goal.
What Ethos planned next
At the time, Ethos said the analysis would be taken further. The plan was for it to be:
- Stored, rather than only shown while someone was writing
- Displayed alongside activity info, so others could see it next to a review or vouch
Ethos described this as coming soon. The post does not record when, or whether, that next step shipped.
Separating signal from noise
Ethos framed the goal as helping to separate signal from noise. On a network where credibility comes from what people say about each other, not every review carries the same weight. A detailed account of a real interaction is more useful to a reader than a vague one.
Nudging authors toward detail at the moment they write addresses that at the source. Storing and displaying the analysis, as planned, would extend it to readers, giving them more context for judging the reviews and vouches they see.
This post first appeared on X.
