so if X start calling a tweet something else (for example, an X-post), this market would resolve to No?

I still don't fully understand the resolution criteria. If it would resolve today, how would it resolve?

@ErwinRossen Hmm, no response after three months doesn't instill confidence in this markets proper resolution. :-(

I can still find mentions of Twitter, but mostly in archived help documents and not the application proper. The legal documents like Privacy Policy and Terms of Service appear to have been updated to say X.

The last clear reference to Twitter and Tweets is the ad analytics. The HTML <title> is "Twitter Analytics for Mira" and the page calls things tweets:

Though it seems to be broken. If this is the only page on the site that refers to things as tweets, it will have to be working and identify specific posts as tweets to count for "message referred to as a tweet".

The "new version of account analytics beta" just takes me to the main app.

They've got some work to do on logged out views.

@jskf Elon fired too many of his devs lol

Here's an interesting one:

Let's assume twitter.com changes its main domain to x.com, likely going to happen before or later. And let's now assume that they set up a redirect that redirects twitter.com visitors to x.com. Couldn't it be argued that that would qualify as "any website that refers to itself as Twitter"?

@JamesColiar This wouldn't count: The domain is used to locate websites but isn't part of the website. I'll add a definition for "website" that it's the visual rendering only, though.

"Tweets" are rebranded to "Posts":

Still plenty of references to specific messages as tweets, such as the "Tweet Analytics":

@Mira its just a matter of time till all the PRs get approved to change the strings hah

@Mira Are there still any instances of this: "Twitter is any website that refers to itself as Twitter", though?

And based on this: "A Tweet is either a message referred to as a Tweet or a publicly-visible reply to anything currently(as of 2023-04-04) named a tweet." Even if they remove all references to tweets, it wouldn't matter because 'posts' as replies to past tweets would still qualify as "anything currently (as of 2023-04-04) named a tweet." No?

@JamesColiar Yes, there are still references to "Twitter", and yes publicly-visible replies to tweets are still considered tweets. So if this market resolved today it would be YES.

For replies: If I use a private browser and it forces me to log in before I can see the reply-to-tweet, it wouldn't be publicly-visible.

oh? https://twitter.com/TurnerNovak/status/1683479012213944323

@Blomfilter I expect this will be received very poorly. Twitter Blue was, and this is an even worse way to monetize.

This market raises some technical questions.

Imagine a new internet protocol comes out which is great but which means domain names now have changed to a new format, and everybody quickly adopts it.

But the websites and content still exist, it's all fine. You can even still use today's hostname but you'll be redirected to something new like google.com.vr3

This twitter market is written in such a way that if the protocol change above happened, all current websites would be considered not to exist since it relies on the post-redirect URL rather than on the domains which it still maintains redirects from

I think that's a bit uncomfortable and not what the people will expect.

@StrayClimb

Right now, such a URL redirect would still resolve YES because my web browser accessing twitter.com.vr would redirect to twitter.com - a website that refers to itself as Twitter(the search bar). The locator is my web browser(which follows redirects), not nslookup.

Or here is the "Twitter Terms of Service": https://twitter.com/en/tos

They need to be more thorough in their rebranding for this to resolve NO.

@Mira that was just an example. It could be another change which technically modifies the URL internally, for example with punycode or for some new unicode-like system.

I mean more generally this situation: when there's an internal internet change which 99% of users might not know about or see, and all old links work --

In that situation I'd guess most users would feel the site still exists, but the way the definitions are written, this claim would consider that not existing.

@StrayClimb

99% of users might not know about or see

The website has to call itself Twitter. The branding is a clearly user-visible aspect of the site: That's why I showed screenshots that anybody could've taken to verify this.

The difference between the bird logo and X isn't hidden, it's something everybody sees because you click on it to get to your timeline view.

The important thing is that I can start by typing "twitter.com" into my web browser's address bar and land on a website that refers to itself as "Twitter".

@Mira Hmm, let's say most of the site uses the name "X" and there's a few leftover references to Twitter but it's clear that the principal name the site uses for itself is X. Do you intend to count that as a YES or NO? I had expected that to be a NO.

@jack For consistency with https://manifold.markets/Mira/will-twitter-still-exist-by-end-of#l2pFtIf9zkXuzbj61rRg , a single reference to the current site as Twitter as part of the main app would count:

If they change the name of "tweets", the name change must be thorough to resolve NO: Even a single reference on Twitter to a new message as a "tweet" in some obscure place like the ad analytics would qualify, as long as it's unambiguous that a specific new message is being referred to as a tweet.

The homepage currently has this casual reference:

I will be looking for any earnest reference to the site as "Twitter". Some example scenarios:

Things that would count(anywhere accessible from twitter.com by web browser):

• The main app referring to itself as Twitter

• The home/signup page

• "Twitter for Business", even though it is a separate product from "Twitter"

• Legal pages like their privacy policy, terms of service, etc. referring to Twitter, as long as those pages are "actively binding" and not just archived.

Things that wouldn't count include:

• An old blog post referring to the site as Twitter that was kept for archival purposes

• Users writing Tweets or other messages that refer to the site as Twitter, even if those messages are hosted on X

• The phrase "X(formerly Twitter)" (it used to be referred to as Twitter, but isn't anymore)

Things that would make me consider PROB, if they were the only such instance:

• A single phrase like "X - the new Twitter" (One interpretation: It's not the old Twitter. Another: It's a new version of the same Twitter)

If Twitter.com including links to old tweets redirect to the correct location on some other domain, will you consider Twitter not to exist?

This is already defined - the answer is NO if it's on a different domain (even with a redirect)

Twitter is any website that refers to itself as Twitter located at either the domain twitter.com or any subdomain of twitter.com redirected to past a login step.

A Tweet is either a message referred to as a Tweet or a publicly-visible reply to anything currently(as of 2023-04-04) named a tweet.

@jack Imagine if I tweet today. Then they rename "tweet" to post. But you can still visit what in my mind is a tweet, at the same URL (which redirects), and the content I wrote is still there. It's a bit absurd to say that Twitter doesn't exist at that point. It obviously exists and all my tweets are still there, although some people call them another name.

DNS isn't real anyway it's all IPs

I wonder how the market rating system will respond when everyone, still using their Twitter apps and seeing all their old tweets, is told that "Twitter doesn't exist" because of a redirect rule (although even then old tweets direct links will still work)

@StrayClimb Obviously I agree that the resolution criteria did not match the spirit of the question. But if you write resolution criteria, you have to follow them, or else what was the point of them? (Personally, I think it's reasonable to resolve N/A if the resolution criteria and the spirit of the question conflict, but most prediction markets just follow the resolution criteria.)

Yes, these resolution criteria contained an unexpected edge case that made it differ from the original intended question, and that is an issue that does occasionally happen. But it's far more common for underspecified questions to have issues and be ambiguous.

@jack I appreciate you're just following the text. I just wonder if we can include the title, too, in considerations? We don't have to follow American legal precedent here. Common sense says Twitter still exists. If I change username does my account still exist? To me duh yes 😆

@StrayClimb The general principle is that the title is a summary of the resolution criteria. It is short, so it cannot capture all the detail.

Also note that there is no contradiction here if you read the question title + text as a whole. The title uses the term Twitter, and the text defines what that term means. It is a reasonable definition, although not the ideal definition.

I 100% agree that a question "Will Twitter still exist" with no further clarifications should follow Twitter the entity regardless of renamings. But when there are further clarifications, you should abide by them.

@jack I read the description as defining the terms at time of market creation

IE if I sign a contract with someone who changes their name, the contract is still valid.

So by this reading, we defined a market on whether a "thing today known as twitter" still exists.

The thing changed name but still exists.

@StrayClimb I don't understand your argument. The question says

Will Twitter still exist

[...]

Twitter is any website that refers to itself as Twitter located at either the domain twitter.com or any subdomain of twitter.com redirected to past a login step.

The definitions is right there for everyone to read. A contract with this language would work the same way. Obviously this is different than a contract with the company Twitter - imagine something more like a contract for ownership of the domain twitter.com

@StrayClimb Somebody asked this exact question in a comment 3 months ago: https://manifold.markets/Mira/will-twitter-still-exist-by-end-of#rfXosVQy7Tdvn1FSB0Sq

The description was modified specifically to incorporate my answer on how I'd handle a rebranding.

@jack

Imagine this market:

---

Will the shIp Queen Mary still exist in 2030

The Queen Mary is a ship located in Long Beach CA

---

If the Queen Mary moves, is this false? To me, no. That was just defining the terms

On the Twitter claim the description described the entity we are discussing the existence of, at writing time.

We understand what that refers to and should evaluate whether the fears in the description have happened.

To me, no since all tweet content etc still exists

Same way the queen mary still exists even if it moves.

@StrayClimb I see what you're saying, and I agree that is a common way to write things. But a careful reading makes it clear that this definition is not about what Twitter is today, but about what counts as Twitter in the future:

• the phrase "any website" and the stuff "located at either the domain twitter.com or any subdomain of twitter.com redirected to past a login step." makes this quite obvious. It wouldn't make sense to write that if I were saying what Twitter is today - then I'd write "the social media website Twitter (located twitter.com)"

• "A Tweet is either a message referred to as a Tweet or a publicly-visible reply to anything currently(as of 2023-04-04) named a tweet." is similar - this phraseology makes zero sense if I were saying "this is what a tweet is today"

@Mira @jack okay, thanks

Your reading makes sense based on the definitions. But it makes the first paragraph of the description sound meaningless since the sense of that part of the description will be violated if we say "no Twitter doesn't exist" even though the fears expressed never happened

@StrayClimb I 100% agree there. As I said

Yes, these resolution criteria contained an unexpected edge case that made it differ from the original intended question, and that is an issue that does occasionally happen. But it's far more common for underspecified questions to have issues and be ambiguous.

I bought my original yes shares April 3 before clarification anyway. Just feels strange that we ignore sense of the title and leading paragraph.

Given the populist turn mm has taken lately with market ratings, something doesn't square with this kind of legal rules parsing.

@StrayClimb The first paragraph is "context" introducing betters to reasons why people might be betting against Twitter. "There's a lot of uncertainty in the air, here's some examples of things that could cause NO". But you're correct that it is does not directly affect resolution.

My usual template is:

• Non-binding introductory context

• Resolution criteria

• Definitions/fine print used in the criteria

• Edit history

• Comments(if they haven't made their way into the description yet)

@Mira thanks, that's useful. Looking back, nobody is super deep in this market so hopefully it will be smooth no matter what happens!