[jameshart]

It's not immediately clear to me that they can't.

Twitter makes NO guarantees that they will deliver every tweet to all consumers simultaneously, or even at all. They don't guarantee they won't hold the tweet up while they decide how to process it - they certainly analyze tweets for content and automatically hold some for manual review, for example. So why on earth couldn't they accept the tweet, process the information in it, and make corporate decisions about how to, say, adjust their investment portfolio... before they put the tweet's contents out on the wire for the world? Is there anything in the ToS that says they can't act on the contents of tweets you send them? You shared the information with Twitter with the intent of publishing it to your followers - but the agent you handed the information to first was Twitter.

There's a sci-fi black mirror kind of scenario in there, too... It's quite possible that if the president of the US were to declare war, the first place that information would reach outside the White House and Pentagon would be a Twitter HTTP service.

Think about the responsibility the software that has to handle, route, and process that request bears. Think about the threat model the systems that software is part of need to consider, if that really is potentially one way it could be used.

Worse still, it might not even be the 'send tweet' handler that gets notice first; the twitter client into which the message is being typed, sending back analytic data to help optimize its autocomplete suggestions or pushing data to the server so it will be ready to offer an appropriate gif to accompany the tweet, might be sending back the draft text of the announcement before the president hits 'Send'.

So here's your story writing prompt:

A Twitter engineer reviewing logs trying to track down a bug in the gif autosuggest algorithm chances across a series of requests from the president's iPhone in the last two minutes, containing text from a tweet being composed, that look for all the world like the drafting and redrafting of an announcement that the country is at war...

What does Twitter do?

[maest]

IANAL, but it strikes me as 100% illegal for Twitter to trade on their tweets before they are released (at least for the class of tweets which contain corporate announcements - tweets from e.g. the persident might be treated differently).

Some of those tweets can be resonably be construed as being MNPI, so trading on it before it's published on the platform is most likely illegal.

Can newspapers/journalists trade on a piece of news before they break it?

[rdruxn]

This is remarkable. I would love to see this in black mirror

[arthurcolle]

Block the tweet? It's like the USSR guy who decided that the early warning signal was incorrect, and prevented WW3.

This idea of unilateral "freedom of speech" is beyond goofy - some actions that deliberately cause massive amounts of distress, even in the short-term, must be attenuated in some manner.

[gpm]

If Trump announces the US is at war, for most purposes in the short term (until congress steps in to assert their power) the US is at war. It is not a false alarm.

If twitter had some reason to believe that the tweet was a false alarm about going to war, and not an official announcement by the US military's commander in chief, that might be a reasonable to take. Perhaps if they also believed the account had been hacked or that a staffer had been duped into thinking that they were supposed to announce a war when they weren't. That's not the hypothetical situation here though.

[jameshart]

Sounds like a good cliffhanger leading into the third act, though.... :)