4 ms·
It might be too little too late. Twitter made a lot of sense in 2010 when you only saw messages from the people you followed, and only discovered new people whe
by kirillzubovsky 6y ago
It might be too little too late. Twitter made a lot of sense in 2010 when you only saw messages from the people you followed, and only discovered new people when those people specifically put their words into your stream. It was valuable, even if people were tweeting about their vacations and sandwiches.
App.net tried to launch a paid version of Twitter and it almost succeeded, but they weren't able to get a critical mass to convert all at once. Twitter was still more valuable as 8 out of 10 of your friends were still on Twitter, not App.
It has changed though. With more users and the new algorithmic timeline, If you are just starting out, you have no chance at getting good value out. Twitter decides what you see, whether you like it or not. It's like the TV from 1984 that is continuously bringing junk food into your living room. You don't get a remote, you only get a choice to have it, or not to have it in your house.
Twitter used to deliver unprecedented value 1:1, and now it extracts value from 100,000 and delivers it to 1. Unless you are the one, there is no sense in paying for it.
- jsnell 6y ago> App.net tried to launch a paid version of Twitter and it almost succeeded, but they weren't able to get a critical mass to convert all at once. Did it really almost succeed? Looking at the numbers from the Wikipedia page, it got a decent number of people joining right at the start. Then it seems to have gotten very limited user engagement from those people, and after the first year the subscriptions could only barely cover the hosting costs.
- chanmad29 6y agoI started out loving Twitter. But not getting much interaction as opposed to reddit, and the force feeding of the algorithmic timeline, just put me off.
- xhkkffbf 6y agoLately all I see are tech ads and a few tweets from some people I follow who really should get back to work. There are a few that have a good signal to noise ratio, but it seems like the algorithms hide them. I'm not sure why. I've switched over to some private blogs and better curated news feeds. It's just becoming dominated by the haters.