Why I prefer RSS for reading News

Rss

Lately the german chancellor Angela Merkel said: ‘Machine learning without data is like a cow without food.’. My first thought was not if that is right or wrong, there are self learning systems etc.. But I thought: ‘Now even the chancellor talks about machine learning?!’. Machine learning and artificial intelligence are topics that everybody talks about these days. From the self driving car to a movie that you might like, all machine learning, got it.

I started to get into machine learning therory a while ago, but I am by no means an expert. You do not have to be an expert to use machine learning though. Those terms immediateley sound like future. The algorithms are around for quite some time, but the technology first started to take off with ‘big data’ and the ‘internet’.

While it sounds pretty futuristic, machine learning is still based on - what else: numbers crunching. To be more precise, minimize a cost function. Remember school when you had to get the derivative of some function? Exactly, just more complicated. It is basically an optimisation of a function. Still sounding futuristic?

Why am I telling you that? What does that have to do with RSS? Well, I wanted to get this ‘intelligence’, ‘learning’ thing a little bit out of the way, as I think those terms are a little misleading, while not wrong. What happens on facebook and twitter is, they optimize your feed so that in the end their revenue is optimised. Not only that but sponsored posts are shown more likely. A funny cat video probably gets more social buzz than some news article.

‘Filter bubble’, a widely used term for that you are only shown what you will probably like anyway, is a direct consequence of the introduction. The system is optimised by numbers and metrics, stuff you like, stuff your facebook friends like. It is still numbers crunching.

Another case is Spotify, where machine learning tends to result in a support of main stream artists and catchy hooks. So those algorithms even effect culture. Pop songs are one thing, but what about news?

The news have been in the news a lot lately. Trump with his fake news, now Silicon Trump complaining about bad news about his car. I have to admit, that I also have the feeling, that the internet made the news quality worse. That again has to do with money and metrics.

In print it is hard to measure which news perform well. And advertisers only care about circulation and probably roughly the target group. Online you can measure and target nearly everything, and if more readers click on ‘Android is better than iOS’ or ‘You won’t believe what happens next’ articles, it is logical that these article are written.

Now it is also logical, that these articles appear more often in your social media feed, as they are clicked more. What I concluded for myself is that news and social media are two seperate things. What works in social media or netflix is probably not that good for news. And news is a pretty serious topic.

I am pretty sure that there are some important decisions and developments that have been shifted by social media lately. It just seems to tend to extremes, something that could be explained with the optimisation theory I wrote about in the beginning. If you lean towards one direction, this direction will probably be emphasized.

When RSS was invented, the internet was a different place of course. While most social networks have APIs you can use, RSS is a standard that everyone can use to share information. Your blog, a feed reader, an app on your phone etc. A common standard without some big company with interest behind it. Please comment if you find one today..

With RSS you can select your news sources, big media companies or small blogs, youtube channels and what not, and get them chronologically. Even twitter used to support RSS but does not anymore. It might sound boring against ‘artificial intelligence’ and a giant robot that learns while reading a book, but in that case it is about the information that you read, and that you select for yourself what you find interesting and what not.

My little intro speech about machine learning might have sounded a little negative. It was not meant that way. But machine learning is a technique, it is maths. You can do great things with it. But while machine learning is great, I do not think it is a good case for news filtering or recommendations.

That all and the fact that I was and still am interested in text clustering lead me to develop a RSS reader, Nooshub (all good domains were already taken). It clusters similar entries, so that it gets easier to see different positions on the same topic. It is early alpha and free while it is, but I am using it and it works, check it out and let me know what you think @nooshub.

When I started it some weeks ago I did not check if there are still people using RSS, but after some time I realized that RSS is very much alive and kicking. Hopefully it will stay that way! Cheers.


Comments