Hacker News Reader

Comment by lsanger | Hacker News Reader

I started Wikipedia back in 2001 and after a long string of nonprofits have started Infobitt.com, which is, sort of, Wikipedia for news. We scour the web for facts about a given story, aggregate hand-written summaries of the facts, rank-order them, and rank-order whole stories (collections of facts = bitts) as well. At scale, we will make news less noisy and more efficient to catch up on. My claim is that only a giant online community could do this--traditional news orgs aren't big enough, and algorithms aren't sophisticated enough (any more than they're sophisticated enough to write Wikipedia).

View on HN

What about Wikinews? It is a Wikimedia project and does collaborative community written articles. http://www.wikinews.org/

It is bothering me that I have to even sign up to view anything on Infobitt.com. That already is sending me away since I refuse to blindly sign up for any service.

One of the main problems right now is that Wikipedia is being used for news, but the format is very poor.

Actually I prefer it to many news formats. If there is a big story (e.g. https://en.wikipedia.org/wiki/2011_England_riots ) and you turn on the news or go to the front page of reuters.com in the middle of it they are usually focused on some up-to-the-second minutiae or the latest sub-story to have occurred in the larger story.

This is fine if you have been paying attention to everything up until that point.

This is not so great if you came in a bit late and want an overview of everything about the story.

We handle this problem by creating the page as a news stream, based on each individual news event. We don't have UK riots yet, but here's an equivalent. http://newslines.org/michael-brown/ Unlike Wikipedia's static page-based format, because each news item is discrete, readers can sort the line from oldest news to first, or vice versa. We also use video and other embedded media to tell the story of events, unlike Wikipedia's textbook-like pages.

I actually agree. This is one area in which Wikipedia shines. But when it comes to breaking news, or tracking the actual order of events as they were reported, not so much.

I have of course wondered what it would be like to do Infobitt with a wiki, and I considered setting a wiki up for that purpose. The bottom line is that wikis lack the potential reasons for using the Infobitt format in the first place--making it easier to compete as well as collaborate, making it possible to vote on small pieces of content (as well as the ordering of the content), etc.

In other words, wiki forces users to collaborate on the same extended piece of content. This has all sorts of great effects, if enough people are participating. But it makes it harder to make short fungible pieces of content, rearrange them by vote, and do contests to discover the best version of each type.

Exactly.

How do you aim to tackle news about Russia & Ukraine?

How do you aim to tackle lengthy bike-shedding over trivial nonsense? (for example, there are at least a million words about hyphens, dashes, endashes and em dashes).

It's democratic, like Hacker News. We'll ask people not to include value judgments/obvious bias, but when it comes to relative ranking of facts, and how facts are worded (a matter for contests), we'll have to trust in the moderating influence of a wide variety of contributors.

In the fullness of time, at scale, we'll have features permitting us to filter/weight the news presented to us according to our preferences. Breaking news into small pieces makes this possible.

Interesting. How do you plan to keep coverage unbiased? A lot of bias manages to sneak into Wikipedia articles on current events in spite of their community guidelines and moderation.

The fact that bitt and fact order, as well as which fact is displayed to the reader, is (or will soon be) subject to vote will provide a moderating influence, I hope.

The reason Wikipedia articles are often biased is that people in authority in the community simply do not respect the neutrality policy. Even if most Wikipedians and most readers would prefer things to be more balanced, those in de facto control of an article determine the article's degree of bias. If the wording and order of facts is determined by open public vote, my hypothesis is that there will be more of a moderating effect.

However it would depend on who is doing the voting. If you, for example, used Twitter to gauge political sentiment, you'd be representing x% of the majority position, which would introduce a selection bias that would imply that x% represent the views of everyone. The problems with Wikipedia for controversial topics Ali's that typically both sides are highly motivated to "win" whereas those that aren't so motivated probably have the correct answer.

I second this one. Traditional news have little problem with their reputation. In general it's known if the newspaper/magazine describes itself as leftleaning, conservative etc..

I can imagine this to be much harder in a community with a vague, biased crowd not identifiable due to the hight amount of uncoordinated accounts participating.

I'd love to hear more thoughts on this bias-problem.

Battling organized partisanship is a problem for down the road. My hope is that, by the time we deal with that, we'll have the funding and the personnel to code up a system that enables us to test out some technical solutions to this problem. There are lots of ideas...

Cool idea worth testing.

Reply on news.ycombinator.com