There really needs to be an option that users can turn on to block all @ from people they don't follow. Even temporarily, say for a period of a month at a time.

I'm also a fan of the "reddit gold" style idea of cheap paid accounts that don't necessarily have to be paid by the account holder. That would make it easy to allow people to filter out unpaid unknown accounts and raise the cost of abuse.

The underlying problem is that comment moderation is a kind of emotional labour that can be fairly gruelling at times, and it needs suitably paid staff to do. Attempts at automatically dealing with it are very much like the early days of spam fighting: not very accurate and constantly shifting. But whereas spam is broadcast, abuse is targeted.

> There really needs to be an option that users can turn on to block all @ from people they don't follow. Even temporarily, say for a period of a month at a time.

Get a blue tick :)

I have the worst source (I think I heard it on a podcast...) but I'm pretty sure that verified Twitter accounts get an extra notification option to prevent notifications or messages from non-verified users from appearing. It basically turns Twitter into a verified-users-only service where you never see the regular plebs.

That's the behaviour, yeah. Verified accounts live in their own rareified world. The idea being that high-follower count accounts are probably more interested in chatting with their peers (most often, celebrities) than with deal with the deluge of plebeian notifications.

They already have this filter - https://twitter.com/i/notifications?filter=following

404

Only works if you're logged in.

I can't think of a reasonable way of blocking people with bad intentions from creating Twitter accounts. But there are other things they could do.

One approach would be to expand their verified user program to people like me (ie the proletariat) and then have an option to switch the blacklist model of blocking people to a whitelist model of showing only tweets from verified users.

Or they could make a whitelist based on people I follow or that friends follow (or friends-of-friends etc...).

Or maybe just get serious with analytics and sentiment analysis and don't show me stuff that is likely a troll.

For a service that's trying to grow, that would force them to put all new users into an "unverified" bucket, which will probably be a turn-off to those users.

> that would force them to put all new users into an "unverified" bucket

But that's exactly how it is now.

I personally like the verified user option but I can understand how some wouldn't for reasons like the one you give. Frankly, I think Twitter's focus on growth is a big part of what got it into this problem in the first place. I'd rather see them focus on quality than growth and I think expanding the verified user program does that.

The measure-of-connectedness option is a little gentler and would probably fix a pretty big piece of the problem.

> I'd rather see them focus on quality than growth...

Quality of what? Discourse? Network connectedness? Ad-accuracy? (I'm nitpicking, but I'm curious what you mean.)

That's a good question. From my perspective as a user and fan of Twitter, I would say quality of discourse. That's a funny thing though because reading a tweet can be like trying to decipher a vanity license plate. Still, that's enough to communicate some pretty important stuff (like during the Arab Spring events a few years ago).

So the logical next question is how does Twitter shape how it is used? I think the answer to that is in part, let other people take care of that. In some ways, Twitter-the-platform is much more interesting that Twitter-the-product. My current favorite example is Nuzzle.

Unfortunately, Twitter is a $13 billion (and falling!) company. Investors think Twitter is Facebook Jr. and it really isn't.

Here are some common answers:

- ban the email address associated with an account from creating other accounts, or in the softer form make someone blocking a user on Twitter block every address using that address. This would run into problems with things like + aliases and people running custom domains who have an effectively unlimited number of addresses, all of which can be reduced but not eliminated if you invest time in it.

- require every account to be linked to a cell phone number using SMS (“mandatory two-factor for everyone!”). This would be more effective than emails but would cost some users real money, screw or delay others if they have unreliable cell carriers, and completely block anyone who doesn't have a cell phone.

- require every account to be verified using a credit card, blocking anyone without a card and no doubt discouraging others. This works great for metafilter.com but there's also a reason why MetaFilter is a site loved by thousands of users rather than millions.

The problem is that Twitter was a VC-funded company which had a big IPO based on highly optimistic growth estimates and their corporate thinking appears to be focused on maintaining those unsustainable numbers. Anything which slows people creating accounts goes directly against the goal of having ever-larger numbers to report.

- require every account to be linked to a cell phone number using SMS

They currently do this to "problem accounts". I have no interest in giving Twitter my phone number and would ideally like to not even have a phone number in the next few years.

That's a perfect example of the problem: it'd make it harder but far from impossible to cheat (cheap SIMs, spoofing through insecure telcos, etc) and there are a non-zero number of legitimate users who don't want to give Twitter a phone number.

You'd probably need some sort of hybrid approach to do this for real – SMS validation counts as x points, y months of interaction with other users without a complaint counts as z points, etc. It's similar to the email anti-spam measures but a less daunting problem since they don't interoperate with many thousands of other peers (for better and for worse).

Not without making it much harder to onboard new users. User growth (Monthly active users) is a business goal Twitter reports to investors so that's a non-starter.

I suspect that it's now a solvable problem to determine if the first few messages from an account are, if not "harassment", at the least... argumentative. Angry. Uncharitable. Sentiment analysis is pretty well-established, as is figuring out what people are talking about for the purpose of advertising at them.

If a new account joins and immediately starts expressing strong opinions on, say, a certain Breitbart reporter and how much ethics is in his journalism, whatever those opinions are, those don't need to be sent as notifications or show up as a reply to other tweets. It's not like Twitter guarantees reliable delivery, anyway. If they've been around on Twitter for a bit, sure.

But it's not as clearly in the interest of profit as identifying advertising interests (or even identifying spammers).

"The worst messages we see on Twitter"

The key might be the verb "see". If a troll trolls in the forest and no one sees it, did it actually troll?

This eliminates trustworthy delivery and predictability, you'll have to send tweets and just kinda hope they appear rather than expecting them to always appear, but it might be worth it. Or make it optional.

They already force people in some ip ranges (for example tor exits) to do mobile # auth.