Will any private message on Manifold be shared with the public without the consent of any of its participants by 2025?
16
178
290
2025
20%
chance

I'm uncertain that DMs will stay private forever

- @Austin

Resolves NO if this hasn't happened by the end of 2024.

This counts Manifold intentionally choosing to release the data, and also count them doing so unintentionally due to someone exploiting a security vulnerability. (If someone reports such a vulnerability, this resolves YES, even if they don't actually publish the data.) It counts even if it was only a single user or a single message affected.

It does not count something that's the fault of an individual user, such as them leaving themselves logged in to Manifold on a shared computer and someone else taking a screenshot.

Get Ṁ200 play money
Sort by:
predicts YES

See Austin's comment explaining what he means in more detail

I think it's quite unlikely (<2%) that we would deliberately and retroactively publish everyone's DMs. I'm primarily musing about us changing this mechanism sometime in the future (with ample warning). Again, probably the best reference class is what happened with bets.

More caveats:

  • I could imagine deciding to publish a DM if eg I noticed someone making concrete plans to bomb Manifest 2024

  • Historically, our record at keeping things private hasn't been great (see private groups), so caveat emptor

  • If you care about private messaging, I'd suggest eg exchanging contacts on Signal or sth instead of relying on Manifold DMs. DMs are currently a pretty minor feature, so you may be better served by a product whose differentiating feature is its privacy

https://manifold.markets/SemioticRivalry/is-it-bad-that-manifold-publicizes#t6z1WYCknGRRU6LvAbCJ

predicts YES

I'm assuming that if such a change is implemented, it would not count as a positive resolution because the participants would have consented?

bought Ṁ70 of NO

@jack Yeah, this market is only about the current private DMs.

bought Ṁ20 of YES

Created a related market that captures a similar but slightly different question:

@jack This is exactly the same as mine but about only a single message rather than any message, right?

bought Ṁ50 of YES

@IsaacKing That's the main difference, there are some other probably minor differences in specification.

In particular, here are some example events in which your market resolves YES while mine doesn't:

  • Evan's scenario "A selection of relevant DMs were released by the admins as part of the conclusion of a fraud investigation"

  • Bug reveals some DMs but not all

  • Admin accidentally shares DM between two other people

Screenshot of the original discord update channel announcement:

@EvanDaniel Which I suppose poses an obvious question: does this market include "A selection of relevant DMs were released by the admins as part of the conclusion of a fraud investigation"?

@EvanDaniel That'll count, yep.

bought Ṁ10 of YES

edit: answered below

Should say shared by Manifold (not just someone screenshotting)

Probably also clarify whether unintentional vs intentional sharing counts

Oh, I guess "without the consent of any of its participants" does indicate that screenshotting wouldn't count.

bought Ṁ10 of YES

Here's a loophole: suppose Alice and Bob have a Manifold DM. Charlie sneaks onto Alice's computer and screenshots the DM and posts it publicly. That would resolve the question yes as worded now, which is not what you want.

Therefore I still suggest the criteria should be the DM being revealed by Manifold.

Also still should clarify whether unintentional vs intentional sharing counts