SenseMaker

Context that finds you while you browse — on any page, from the people who actually know something about it.

  1. 1It follows the claim, not the platform. One discussion — on the tweet, on the Reddit thread, and on the article itself.
  2. 2Nuance lands on the point, not the page. A note covers the link; people add what it gets wrong sentence by sentence, where it actually matters.
  3. 3See how different rooms read it. Not one verdict — the people you trust, and the readings you'd never otherwise meet, side by side.

Four steps, about two minutes. The bar at the top tells you what to click next.
Concept from October 2020 — three months before Twitter announced Birdwatch.

Home

SenseMaker Search any claim or link JH
· you opened this from a post on x.com
nytimes.com · Oct 19, 2020 · James Gorman

Can Humans Give Coronavirus to Bats, and Other Wildlife?

Researchers are asking whether the virus can move back from people into wild animal populations.

JH
Add your read on this article…
nytimes.com/2020/10/19/science/coronavirus-bats-wildlife.html
SEARCH The New York Times SUBSCRIBE FOR $1/WEEKLOG IN
ScienceHealthClimateSpace & AstronomyAnimals
Trilobites

Can Humans Give Coronavirus to Bats, and Other Wildlife?

Researchers are asking whether the virus can move back from people into wild animal populations — and what that would mean for the species involved.

Little brown bats in a hibernaculum. Field researchers have suspended some winter surveys as a precaution.Nicholas Pfosi for The New York Times
By James Gorman
Oct. 19, 2020·4 min read ↗□▭

As the pandemic entered its second year, a quieter question began circulating among wildlife biologists: not where the virus came from, but where it might go next. Several agencies have paused hands-on fieldwork with certain species while the risk is assessed.

The distinction matters, because a virus that establishes itself in a wild population becomes almost impossible to remove.

Risk assessments published this year put the probability of human-to-bat transmission during routine fieldwork as low, but not negligible — a judgment that rests on assumptions about contact rates that are themselves poorly measured.

For species already under pressure from white-nose syndrome, even a modest additional mortality could matter. That is the argument for caution, and it is doing more work in the current guidance than the transmission evidence itself.

Researchers who spoke for this article were careful to separate the two questions. One is whether infection is biologically possible; the other is whether it happens often enough, in the wild, to establish a reservoir. The first has partial answers. The second does not.

What you just used

Four mechanisms, and why they might matter

None of these are new on their own. The argument is that they only work together, and that no product currently runs all four.

The note is drafted by AI and published by people

A model reads the sources and writes the draft. It stays invisible until enough people rate it helpful — including people who usually rate things differently from each other.

Why it might matter: there are far more contested claims than volunteers. Drafting is the part that has to scale; judging is the part that cannot be automated without losing the thing that makes it trustworthy.

Nuance attaches to the sentence, not the page

A note covers the link. Underneath it, contributors anchor to the specific passage that overreaches — so the correction sits where the problem is, not in a general verdict about the whole article.

Why it might matter: most articles aren't true or false. They're mostly fine with one step that doesn't hold, and a page-level rating can't say that.

Identity is verified; nothing claims you were right

Backgrounds come from third parties — an ORCID record, a university directory, a veterinary board. Nobody types their own credentials in, and no score anywhere says a person was correct.

Why it might matter: self-declared expertise recruits the self-promotional and repels everyone else. And scoring people on being right needs someone to rule on what is true — which is the thing this whole problem is about. The only performance number here is whether people who usually disagree found you worth reading.

You see the spread, not just your own room

Weight the people you find useful, and that shapes one view. The default stays the reading that holds up across people who usually disagree — and you can switch to see how a different group read the same piece.

Why it might matter: personal trust weights on their own are a filter bubble with a credibility skin. Showing the other readings next to yours is what turns that from a flaw into the point.

The reason all of this sits outside any one platform: a claim doesn't stay in one place. It appears as a tweet, a Reddit thread and the article itself, and a per-platform product has to fight the same argument separately in each. Follow the claim and you only have to win it once.