What Is The Internet Hiding?

Remove all ads!

I thought this was really interesting about what is happening to the internet:

http://www.ted.com/talks/eli_pariser_beware_online_filter_bubbles.html

I'm not surprised, not in the least.

 

My hearing's screwed up at the minute, can you give a summary of what's going on please?

 

I think it's a little scary - in a "1984" kind of way...

 

Silvery:

 

Oddly enough, I was thinking about this today. I searched for a recipe on my lap top at home and then went to my mums, searched for the same thing, got a different result!

 

@Silv, all TED videos have subtitles. Look for the dropdown box below the video.

 

Me too!

 

With an endless supply of data at your fingertips, answering which is the "right" endless supply for the right people is no easy question to answer. I lean toward minimizing this sort of filtering so that we all are seeing essentially the same information for any given topic we choose to look into. This at least keeps us on a more level playing field in the information age.

Filtering based on type of browser? That's a new one for me. What... IE users get funny pictures when they look up a serious political issue, while Firefox users get highly academic, well researched articles? What in the hell does browser choice have to do with what a person should see?

 

Well, it's their data mining algorithms. It's not that someone at Google actively discriminates against types of browsers, it's the algorithms that they use that automatically make these decisions for them because, apparently, browser selection is good factor for predicting surf behavior.

 

The algorithms are written by people, so whether it’s active or programmatic, it's a form of decision making by them.

I'm sure it starts at a strategy meeting, then requirements are created, then technical specs, then a design, then testing, then perhaps some tweaking, then more testing, then, eventually, eventually, it gets implemented on their live website and is used whenever someone launches it.

All those layers of separation don't eliminate the fact that humans initially got together and discussed how they wanted things to work.

 

Depends if it was done by mathematicaly principals based on the spread of searches and site access by browser. If on a search with the similar results are accessed by 50 IE browsers and 50 Firefox browsers and 40 IE users open the top link and 10 open the second and on Firefox 40 users open the second link and 10 open the first the application will sort it in that fashion.

It's not necessarily elimination of results - it's sorting codes by popularity to try and match as many factors as possible, it's card pairs - not poker.

 

I do not see this is a solely bad thing. If everyone had access to the exact same information the people deciding what that information would be extremely powerful. With this filtering there is some variation in the information which hopefully means that it is slightly more difficult to get a monopoly on it.

 

It's not elimination of results. They aren't "removing" content from the Internet. That's not the question, or issue.

What I'm questioning is why an Internet Explorer search and a Firefox search on the exact same topic in the same part of the world would produce different results.

Google could design their site to work about a million and one different ways. There's no requirement that results are served based on what's popular in IE vs. Firefox vs. Chrome or even vs. Safari. If they filter based on this, I am wondering what the reasoning behind it is -- we can't just blame "mathematical principals." People decide how the website will function, and use those mathematical principals to accomplish their goals.

 

Most data mining algorithms train themselves. It is up to the original designer to specify which factors (such as browser) to use in the calculation, but the algorithm eventually learns the appropriate weight for each factor. If browser selection didn't matter, then the browser used wouldn't be a factor in the calculation at all since the algorithm would assign it an insignificant weight. But the browser used does affect the results and we can thus conclude that it is an important input factor and is assigned a significant weight value.

The amount of weight given to this factor isn't decided by the engineer who made the algorithm, but is determined by the algorithm itself when it adapts itself to the gathered data. There is no one at Google saying "oh let's give the browser at weight of 1.32 and see what happens". No human being is clever enough to find the correct weights, and assuming that user preferences were static, it would the machines at Google months to gather and analyze the data.

 

Who's to say that the weights are "correct" if it's all being decided by computers and people can no longer keep up with the decisions being made?

Case in point, I'm a human being who happens to think providing different search results to two people in the same town based on their choice of browser is, at best, bizzare. Further, I assert that Google should not filter based on this particular criterion. I don't understand why these results should be different... but the algorithms powering Google's search engine do? And I can't blame Google for this, because "those whacky algorithms just do as they see fit?"

It was Google that deployed the algorithms and made them function the way that they do. So the way things are working are still all part of their strategic decisions, whether directly or indirectly. If I don't like the end result, I sure as hell won't be calling an algorithm to complain.

 

This is all over my head.

 

I'm pretty sure these are the voices of programmers talking, right? It seems like a lot of "but this is how it works" type responses. Pretty circular.

My question, and my point is maybe it shouldn't work that way, but some other way. Maybe search results should be more standardized so that people are receiving more consistent results from their searches.