Basically the economics are insanely good if you're using this as an open source tool to create NYT-style articles, less so if you're Circa.
The people comparing this to the semantic web don't understand Zipf's Law, and just how slowly things actually change. E.g. we only get new data on adult literacy every 10 years. And the last data we have on antibiotic resistance for some bacteria/drugs is from the early 90s, and that's more the rule than the exception. Pretty much every single article about the U.S. is using the same set of a couple thousand facts, and most of those only get updated every ten years or so. There are a few exceptions like with federal arrest data that gets updated yearly, but that's pretty rare.
Basically if you're trying to do this using machine learning or any sort of algorithms, you're completely wasting your time and going down the wrong path. This is way easier to implement well than you think. (But again, not necessarily super profitable unless you own the NYT.)
Let's go over some of the points brought up with that in mind:
Would you as a reader pay for more for a service that had "enhanced tools for journalists"? No.
Would you as a reader pay more for "summarization and synthesis"? Probably not, Wikipedia and Google already do this really well.
Would you as a reader pay more for "adaptive content"? No.
Newspapers (including the New York Times) are in serious trouble, and instead of playing catch up to look like a mix of every tech service, they should focus on delivering an experience that people actual want to pay for.
There is no future of news without figuring out the next model for news.
- More AMA (and not just with the famous people; I loved the "I'm a Joe Random Starbucks employee, AMA" threads before AMAs went mainstream).
- More ELI5.
- Supporting HN and (parts of) Reddit so that every article is analyzed and dissected by people in the know, who point out all the bullshit the news station put in, and point towards relevant resources.
Oh, and at this point in my life I'm willing to pay quite a bit for a news service that can prove to me they a) don't blatantly lie, and b) have some minimum competence on-board to cover the topics they're writing about. An information source that I could trust to incorporate to my daily decision-making process is something worth paying money for.
Looking for such a single source will only ever result in frustration. Whether a given news source is assembled by a huge corporation with political affiliation, or a single intellectual who answers to nobody and has complete artistic freedom, it's irrelevant - the source has its own incentives, biases and ultimately, an agenda of some sort.
Even though it obviously happens in some instances, they don't even have to 'blatantly lie', just select and present facts in a way which supports their narrative, and as such even if you could somehow cut out all sources that simply lie, you'd still have a similar problem. How can you routinely trust a single source for your decision-making when you know that source must have its own independent agenda that may not always align with yours?
The answer is you have to do the hard work of reading several sources with different agendas and incentives, getting different perspectives, and making up your own mind about what's really going on with a particular issue based on that. There's no way around it, no shortcut, in fact a source which would charge to provide willing customers with such a shortcut is probably the type of source more inclined to shape their stories to suit what the most well-paying customers, or at least the section of the market they've cornered off, want to hear.
And when I don't have friends to ask about a particular subject, I might end up going to an online community like HN or a subreddit to ask individuals or groups of individuals those questions.
I tend to trust people who don't have a profit motive more than I do a big company which has an incentive on pushing their product.
This is what I've wanted to build for a while; some kind of annotation site, like RapGenius for normal webpages.
Tangential, but my favorite annotation ever is Asimov's Gilbert and Sullivan, which I'm fortunate enough to own.
The New York Times is already a large corporation, with lots of employees, and tied to revenue streams from advertisers that are waning. I don't seehow it's like going back to Facebook or Google in that regard, because their head counts were minuscule.
That said I think you're undervaluing the potential of the OP. You say at  that one of the key things you've learned is that readers fund journalism because they care about having an impact. If successful, the approaches they're discussing could significantly increase the impact of journalists' work -- reusing the same Particles in different contexts, being able to provide multiple tailored versions of the same post that could resonate with different audiences, etc. And it also may point to ways for publishers to participate in and leverage the Buzzfeed/Facebook ecosystems while still providing additional value on their own sites/apps/publications.
So, agreed that it by itself isn't a business model solution, but don't write it off so quickly ...
And it still works.
People are willing to pay enough for quality content or content about local news that it works.
Really these guys are trying to save a sinking ship.
I'll be an average user for the site. I look at the frontpage/feed and see recent news about "The Syrian Crisis" the link would aim towards an article about the event specifically ("Violence in Syria Spurs a Huge Surge in Civilian Flight"). The article itself would focus on the specific event without having to rehash what the Syrian crisis is. There could be a very clear link (Maybe the subheader of "Syrian Conflict") to what topic this event relates to.
I could click on this link and be taken to the broad topic of the Syrian Crisis, which would feature a well-written summary of the topic as it is currently, a ticker of the most recent events (weighted for more significant events), an easy way to filter by country ("Germany and the Refugee Crisis", which would also include a well-written summary and Germany-specific statistics), along with any article tagged with that topic. There's still a focus on well written articles, but augmented by live data and focusing on how specific countries or global leaders/groups are handling events (complete with interviews and smaller local stories). If the article contains important events to the global stage, they can be pushed up to higher topics in the tree (From "Refugee Crisis in Germany" to "Refugee Crisis" for example) Following those topics can give you notifications on recent updates with them. I could glance through each topic, featuring longer/higher-quality articles towards the top (interviews, op-ed, etc.), while it just takes a bit of scrolling down to get to easier-to-consume pictures, videos, quick quotes, or graphs below.
It's all about linking great journalism together into cohesive topics. Articles and op-ed is still created and promoted on the "front page", but photographs, videos, statistics, audio interviews, quotes/tweets, and smaller interest pieces can all be provided in a sort of live stream for somebody passing the time or augmenting the topic's page.
Actually they don't. Wikipedia's long article format is very poor at creating a news archive that can be filtered, sorted and searched like a database. OTOH, Google has a lot of tools for searching, sorting and filtering, but they can only do those actions on existing articles, which are part of the problem. Simply curating existing articles doesn't create a news database. You end up with Google news which has 2000 versions of the same news. To make a news database work, the articles themselves have to be in a format that can work in a database (shorter, fact based etc).
News is not just about « delivering experience » or « a business model worth paying for », everything is. It has solid social objectives and uf people, if our society, deems newspapers or news not worth anything then it will disappear and it'll be sad but it'll be fine because it will regrow somewhere else.
What used to be exclusive due to scarcity, labour costs and lack of relevant technology is being made more an more accessible.
To apply this process to "news" as a means of getting relevant information (as opposed to such journalism which is at least equal part entertainment), you need to look at executive briefings, the news delivered to the president, a congressman's morning file prepped by assistants.
Summly and Circa have tried, but picking 5-10 of THE things you need to see out of a ~million stories or even ~100 main news stories is very difficult.
But I bet Facebook, Google and Apple are very well positioned to solve this, eventually making media an ever-lower margin business.
A lot of news organizations online are doing fine, and many are making a bunch of money. Granted, they're not making what they used to (because they're the text equivalent of record labels) and the Internet leveled the playing field, but news + ads definitely makes money.
Over is the time where people would pay for content. The new model is for service providers to provide content whose call to action is in their best interest.
Perhaps my understanding of business models is flawed, but this year my company will pay out millions dollars to journalists. In that regard perhaps my flawed understanding is an asset!
Because the news summary is treated like data, we can sort it to show different views of the data. for example, reversing the sort gives a "biography view" . Compare this to a Wikipedia page, or even a newspaper article, which, because they text-based, cannot be sorted. By adding meta data we an then filter the data. For example, we use "Event Types" such as births, deaths, arrests, and many more to let the users take control over what they want to see. For example you can see all the apologies on the site.
A big advantage of this way of creating pages is that it results in far less bias than in a traditional news article. If you're interested in more, I wrote a follow up to the NYT article .
That is exactly what Wikipedia does. It works well because most readers are not looking in an encyclopedia for information on yesterday's events.
Likewise inverted pyramid works well because it simultaneously satisfies the needs of new readers who need the most important details at the top and repeat readers who can quickly scan the short paragraphs for new information. I despise new-style live streaming because it is so awkward to read; I have to read backwards, bottom-to-top, and the most important details are often in the middle.
The Tribune Company is about to sell its Michigan Ave building since its bleeding cash. They established a venture fund a few years ago but last I heard hadn't invested yet since they hadn't found anything worth an investment.
In an industry full of failing companies it's delightful to see one Titan try to stay relevant.
I really like the idea of Particles. I had a similar thought this past year about building a news MVP that simply consisted of atomic facts (no more than a sentence) with an attached probability and discussion. Kind of like those you see in the IPCC (international panel on climate change) reports, where they have their claim and degree of certainty.
We were getting there with the whole Wikidata initiative, but it's still in its primitive stages with respect to content being statically authored and updated (AFAIK). It would be very interesting to see where we go with GraphQL - I could see that becoming the dominant machine-to-machine protocol for Semantic Web 3.0. The whole idea of taking what are usually REST resources and turning them into hierarchical JSON documents which you can query very flexibly is immediately appealing.
If this Particle concept catches on like cards have in UI design, I'd expect Twitter to be the first to ride the wave.
As for Twitter, I doubt they can do it either. Their "Moments" initiative shows how little they understand curation.
Could you elaborate on this? Specifically, what do you think is the better alternative?
Newspapers don't print feeds, there's no room. But if you have the archive to hand, as a database, then you don't need to make a perfect article. You just need to add updates as they happen.
Event 1: Plane crashes 10 dead
Event 2: 100 dead confirmed
Event 3: Actually 99 dead
This leaves the original events as a record of what was reported at the time, but the reader can also see the update. This is important to readers because while information can be added to articles, it can also be taken out, to try to conform to a narrative. Wikipedia is particularly prone to this kind of selection bias dressed up as fact-checking.
In many case, fact checking is often a way to hide bias under a veneer of authority, by proclaiming a selected set of selection bias facts as "the fact-checked truth". You only need to follow the major fact checking sites for a short time to see this in action.
The question is how can you effectively present conflicting information. I believe the article format is biased from the outset, whereas a more data-driven approach leads to less bias.
I want to make a slightly bolder statement than the headline of the post:
The future of news is not written, and it's not one way. It's conversational and spoken.
We've been migrating away from the printed word at least since the advent of radio, and TV only accelerated that trend.
What people want is the interaction and surprise and meaning created in a conversation. Papers like the NYT aspire to "drive the conversation." But they are not engaged in "the conversation" on an individual level. They are largely confined to the one-to-many schema of the old news flow, where publications speak and readers listen.
The future of news, imho, is chatbots personalized to the user, that bring up the daily news like small talk on a long commute, based on the AI's knowledge of the news consumer. And it'll have through a voice UX just as much as through print.
BTW. the way you described it - choose-your-own-adventure, seamlessly explorable news - sounds really cool. I'd definitely love to try something like this out.
Sure, there's a subset of information that's ephemeral and must expire at some point. The kind of information with a call to action. But that shouldn't be the whole picture.
We need a new kind of information engines that both knows what you know, and knows what you want to know. Something that both teach you old and timeless facts, and keeps you up to date with new discoveries.
New shiny stories shouldn't distract me from what I planned to read yesterday. We need a cure to novelty.
I like this New York Times initiative.
New information is often more entertaining than old information, particularly if the old information is already known to the reader.
For better or for worse, I think news is often used for entertainment.
The only reason I don't order morning paper these days is that it's physically so big. I only read about 5% of it. If I forget about it, it's going to explode my mailbox. And I would have to take the papers to garbage twice a week to fend off chaos.
Could I please subscribe only to politics, science, actual news and opinions. Curated by local major news outlet. It would be the best possible way to kill the time needed to drink two cups of coffee.
I think people are overthinking this. By "what would average person want?". Maybe they want what you want?
> The information a newspaper-reading citizen receives about economic developments or the spread of epidemics, for example, lacks both continuity and constant correctives. Information comes in isolated fragments. We can assume that those conditions make it considerably more difficult to develop an adequate picture of developments over time.
Maybe this is the answer?
Agility is key. The world of big content will change everything as we begin to ask questions of content and deliver adaptive content focused on individual needs.
My two cents. My crystal ball is cracked, but seems to have been working fine lately. ;)
We're now in online publishing (obviously) and just finished tagging several hundred thousand pieces of content with metadata that we'd used a 3rd party service to extract. We'd reached the exact same conclusions about the process of organizing and categorizing our ginormous back catalog of content, and having humans perform the tasks of summarization and categorization just "doesn't scale" temporally. Interesting topics come and go, editors come and go.
Anyway, I'm really proud that we already have (in production) a system for tagging our content in this exact same way and have already built a v0 recommendation engine out of it.
I've been spreading the merits of the semantic web for years. People just ignore it. I've never been able to tell whether the idea is flawed, or if people simply don't get it.
I regularly go as far as to claim that the lack of semantic web is the cause of all the world's problems.
Semantic Web expects people to do extra work, with no benefit. Create a benefit for those people, and you'll see it get done.
Imagine Facebook being replaced by the semantic web.
Yes, for the readers, if there were enough structured content available, that would be a benefit. They aren't the ones that need convincing.
User can then filter based on the reporter they want to follow on that news item, the hashtag that is interesting to them, or on the timeline of the news item they're following.
In order to leverage the knowledge that is inside every article published, we need to first encode it in a way that makes it searchable and extractable...While news organizations have adapted to new media through the creative use of interactivity, video, and audio, even the most innovative formats are still conceived of as dispatches: items that get published once and don 't evolve or accumulate knowledge over time...
Auto Extracted Ranked Tags
(Algorithm: Tuatara GS1)
article, future, apple, media, published, reporting, facebook, print, knowledge, constraint, publisher, shape, app, discussion, rethink, assumption, accumulate, propose, story, reading, website, rhythm, focu, piece, day, specially, publish, searchable, possibility, accumulative, conceive