Twitter's Fucked

I actually think none of these things are Twitter's biggest problems and that while they are big issues Twitter could gradually, painstakingly fix them through thoughtful policies coupled with careful engineering. Their biggest problem is Facebook rapidly building Twitter's future inside Facebook.

What I mean is this: I agree with Jack's vision of Twitter being a "window to the world", a place where you can instantly get a sense of what your compatriots are thinking and how they are reacting to news. They have two features which accomplish this pretty well: hashtags and live video. When crazy things are happening Twitter is one of the best ways to "live" the event through those two features. Jack appears to have taken this to heart and recently said that Twitter's future will heavily involve breaking news. This also addresses one of the big complaints in this post - you can't have a thoughtful debate in 140 characters. But you can react! And you can absorb little pieces of information and media.

Unfortunately for Twitter, Facebook has also seen the writing and the wall and has built superior versions of hashtags and live video directly into their experience. I have seen a noticeable improvement in the specificity and granularity of their trending topics lately and we've all seen over the past couple weeks how effective and raw (maybe too raw) their live video is.

So I don't think Twitter is fucked - I think their future still, after all these laggard years, has a lot of potential. The problem is they have to stare down an 800,000 pound blue and white gorilla to win.

The likely end result of that is just coded communication.

As a similar example, Facebook's "no gun sales" policy now has people selling a pair of socks for $650 using a picture that just happens to have a gun in the background. (Examples: https://twitter.com/ecined/status/756940120826097664 https://twitter.com/monteiro/status/756639900796088320)

While I agree with most of the author's points, I'm not convinced at how possible it would be to have a product that "adopts the market-orientation and ad hoc network of Twitter" in such a way that "harassment would be impossible." While deep learning and similar technologies can certainly make it much more difficult to harass someone, you can't rely on the content of someones message[0] nor can you rely on the network status of an individual, as spam accounts won't be readily differentiable from new accounts, and blocking new accounts from contacting people defeats the "ad hoc network" requirement.

[0]: People are great at communicating meaning non-literally (see: definition of literally) and one of the ways that manifests itself is through slang, which is going to break anything but a so-close-it-might-as-well-be-AI system. If you ban a word or phrase, it's meaning will be communicated through another word or phrase. The insulting adversary has an advantage here, as they can use other real words and phrases that would otherwise be innocuous to insult you, it would be especially advantageous to use popular/common words/phrases as replacements (see: euphemisms).

It's certainly possible to train a system in hindsight, but I highly doubt you could do so in real-time and any delay in reacting to new slang means that's when antagonists will strike. Hell, you could even do asynchronously defined insults, send many messages containing various words phrases and then announce, shortly after their delivery, what they mean (heh...).

I'm pretty sure they allow it because the government asked them to. It can provide useful intelligence. I doubt detecting and removing that stuff is beyond the Twitter team's technical capabilities.

I agree with a lot of this article, but want to call out this point for discussion:

>This means if you follow someone for what they have to say about professional improv comedy, you also get to hear them trash your religion, berate the intelligence of people who vote the way you do, and otherwise rant about and retweet topics you don’t want to hear about.

This sounds like the author is asking for Twitter to build an echo chamber for him? I would rather not have that for my account tbh and I would like to discourage Twitter from heading in this direction (though it has already started to do some of that).