Hacker News Reader

Comment by rootusrootus | Hacker News Reader

Great project, but this seems like a classic example of a programmer naively assuming a technical solution can solve a social problem.

View on HN

This is definitely worth trying, however.

I'm a bit jealous because I had nearly the same idea back when I worked in public radio, but didn't have the time to implement it. Basically it was about having facts and quotes exist as embedded units within articles, all with their own individual records of verification and relationships to all the articles in which they were used. I'm really glad someone is taking a shot at this, even though I have my doubts as to whether journalists will take it seriously.

YazIAm: If you want any further inspiration, here's my old github issue where I wrote about this concept: https://github.com/SCPR/SCPRv4/issues/718

I don't know what all the criteria are, but you should consider applying for a Knight Foundation grant: https://knightfoundation.org/apply/

Just read through some of your repo, ravenstine. Very cool! I'll be sending you an email :)

But it might be worth a try. Wikipedia does pretty well imho but if someone would told me the idea I would never thought that it would work.

Wikipedia has flaws,too. 1. Your account need to have good history or social repo for your edits to be seriously considered by Wikipedia admins. 2. Admin can have their own set of bias and can decide to ignore edits.

Remember Wikipedia is a centralized and under control of single entity, fact that they rely on donation doesn't change anything.

Just because something has flaws doesn’t mean it’s not a good idea to do it.

There wasn't anything said about Wikipedia being a bad idea, but rather that it shouldn't be your sole source for fact-checking. I've learned that people only trust Wikipedia to the point at which it confirms their biases.

Part of the problem is that Wikipedia is generally pretty darn accurate. Everybody always talks about funny vandalism stories, but generally it's pretty good. The issue of course is that there's parts of it that definitely aren't correct.

This makes it easy to say "Well if it's accurate 99% of the time then it must confirm what I thought!".

I have found that Wikipedia is pretty great, with the sole exception for anything that is remotely political. I would look elsewhere for such content.

The flow of the conversation was:

- OP made an open platform

- Someone said it could be a good idea because Wikipedia

- Someone else said that Wikipedia has flaws

Hence my comment.

Wikipedia, for a few high-profile current events each week, serves a function similar to the one proposed for sourcedfact; and it regularly results in deeper + more contextualized coverage than even the best outlets. However, WP's commitment to being a tertiary source means that _some_ secondary source has to properly interpret the primary sources, before that interpretation can be included.

This is more a communication standard than a complicated tech solution. A communication standard for fact checks from public primary sources.

Coding this wasn't actually that complicated, figuring out a design that enables readers to review the fact checking as easily as possible and enables fact checkers to collaborate easily took a lot of experimentation though. Although even that is still very much a work in progress.

Exactly. I would just appreciate footnotes in the articles I read that gave a detailed explanation of how much fact checking was done.

That way if a story needs to be expanded upon at a later date, the next journalist can dig deeper easily picking up the trail where the previous journalist left off.

I think to solve the problem, you need to do it in steps, kind of like how chained logic works. You start with something simple: "Can we verify what someone says with public record?" Yes, here's how we do it.

As one problem is solved, it starts hinting at solutions to subsequent problems. This is not a problem with a silver bullet. It's going to take a fundamental shift in how we understand and verify truth.

Agreed, it is also a question of goals. Journalists can bend "facts" to support the argument they want. Title are 99% click-bait anymore. Sure there are some people who want to present fair and balanced view points but that takes a ton of work and time. Most "journalism" these days is get likes and move on.

Right. Many prominent newsrooms seem to regard fact checking as an obstacle to avoid.

Still, it seems valuable to make it easy for readers to view the supporting or contradictory evidence. That might’ve stymied many notable media hit-pieces. Of course, many newsrooms wouldn’t want this (at least I can’t imagine The Guardian and it’s ilk would want to draw attention to their own impoverished fact:BS ratios), so it would have to be a browser plugin or similar.

People post contradictory evidence to journalism they don't like all the time. Their "evidence" is usually conspiracy theory crap that can't pass even logic checking, much less fact checking.

Sure, but it's not easy to find all of the debate about a given assertion in a particular article, much less contribute to that debate (given that it's probably distributed across a bunch of social media venues).

One can easily imagine hovering over a given assertion (e.g., "Damore posted an anti-diversity screed") and see the fact-checks, perhaps ranked by fact-checker's credibility in order to suppress conspiracy theory crap as well as contribute to the debate.

Anyway, that there is lots of noise among the criticism doesn't mean that there isn't a lot of valid criticism. It's only been a few months since nearly the whole American newsmedia willfully mislead the nation, sending an angry mob after innocent kids, after all.

> Many prominent newsrooms seem to regard fact checking as an obstacle to avoid.

Or they simply can't afford to employ fact checkers anymore, now that people refuse to pay for news.

> now that people refuse to pay for news

People (the audience, that is) have historically refused to pay a significant share of the cost for news, what has changed is that advertisers have stopped paying as much for news (because the share of people's attention debited to news, and therefore which news outlets have to sell access to, has dropped.)

Paid circulation used to often be important not so much because of direct revenue but because advertisers looked to it as an important measure of reach.

That's true as well. The effect is the same either way, though -- news organizations aren't exactly swimming in money these days. They have trouble just paying journalists, much less support staff for those journalists like fact-checkers.

Maybe that's part of it, but there are plenty of examples of the media deliberately misleading their audience. "Can't afford fact checkers" doesn't really explain that.

"These days" suggests that this is something new. Tabloids and magazines have been doing bunk journalism for decades. Now they are motivated by getting clicks to inflate ad costs instead of getting sales to inflate ad costs. The only thing that has changed is the delivery method. The reliable sources of journalism are unchanged as well. There are plenty of new online tabloids and "e-zines" but journalism still comes from the same places.

Fact checking was very effective for publications like Readers' Digest that routinely invited guest authors. The publication wanted a variety of authors to contribute, and they also had a reputation to defend.

If you've ever been interviewed by a modern journo, you know you just get the same question reworded a dozen different ways. The reason is that they decide what story they want to write up front, and then just go shopping for facts to support it. That story is usually dictated by market pressures, which presently reward sensationalism to feed the addiction of a dwindling number of listeners in an echo chambers.

And current fact checkers are typically checking material put out by others, which is completely broken. They have all the same incentives as anyone writing a hit / puff piece, except they put a phony "truth number" at the top and other sites will tally those up as though it means something.

Isn't that just setting a high bar? It might not solve an issue, but it could help.

I think there is a technical solution to this problem though. A structured graph of supporting facts (including an analog of identity for the sources and corroborators, pointing to the edges) is an excellent datastructure for maintaining a record of reasoning and evidence for a statement of fact.

There is, of course, the social issue of convincing people that this is important, but the underlying issue is in part a technical one. I don't think it's fair to disqualify a solution based on the fact that involves technical work.

I'm reminded of the oft-quoted phrase "you cannot reason someone out of a position they did not reason themselves into."

I am not trying to disqualify the solution, more like observing that the existence of provable facts has not yet been any sort of antidote for 'fake news' (whether that refers to the real deal, or the position many people take that any news they don't like is fake news).

I don't think this sort of effort has to be ironclad to be meaningful. The margins could be a big deal, here.

Some sort of attestation toolchain might enable people who care about evidence to produce more strongly-attested cases, and to get a better handle on what is and isn't sourced from others. It seems likely that using good attestation toolchains could create feedback loops for researchers, journalists, archivists, authors, film-makers, educators, students, and other people involved in knowledge production.

These feedback loops might change how they research, where they start, how often they have to re-invent wheels, and raise standards by enabling these people to compete on metrics that might drive more social value than clicks, engagement, shares, or ad revenue.

These effects could be important, even if you never budge the kind of person inclined to consume the information equivalent of junk-food.

Technical solutions solve social problems all the time. A lock is a technical solution to the social problem of theft, and much superior to a social solution.

Why does it seem like that to you? Do you assume that the project should solve the problem entirely in order for it to be of any help or have any value?

Reply on news.ycombinator.com