Hacker News Reader

Comment by Swizec | Hacker News Reader

If I was Medium, I would use machine learning to curate a feed for each individual subscriber.

Their personalized curated feed is already really good. I had to stop looking at their weekly email because my clickthrough rate was too high and it was a time sink.

View on HN

"If I was Medium, I would use machine learning to curate a feed for each individual subscriber."

If I was Medium I'd hire a team (three or four) experienced newspaper editors [0] and use their expertise to bring back some discipline back to the story process. Do you remember newspapers? There were hundreds of competing stories but the news was carefully selected on a daily basis. This is what is lacking.

The machine is being gamed. Maybe the editorial teams' suggestions should be used to train some aspects of story and category selection? Machine learning is a tool and if Facebook with all it's propellor heads have a problem, what hope have Medium? A different approach is needed.

reference

[0] Lots of newspaper people are out of work or under utilised and as such their availability is high. For example this tweet highlighting guardian journalists getting the sack: https://twitter.com/Hadas_Gold/status/844580929226072065

> Do you remember newspapers?

Honestly? No.

They're that thing my dad would buy on the weekends, read one page of, and then we'd use it for potato peels, lining the floor when painting rooms, and stuff like that.

By the time I was old enough to even consider reading newspapers, they had devolved into clickbait sensationalist garbage. Designed primarily to find whatever boogeyman would make each newspaper's audience most likely to buy the issue. For left newspapers the top story was always about a businessman doing something bad, for right newspapers the top story was about how this or that nationality is threatening our economy. Sometimes both sides had the top story as this or that natural disaster or war that happened so far away that it's irrelevant.

At least the trashy magazines were always honest about being trashy. Their top story was this or that random celebrity doing some outrageous act that nobody cares about.

Honestly, the problem is tying remuneration to readership. The other problems are emergent.

Edit: the intro of this System of a Down video sums up my thoughts excellently --> https://www.youtube.com/watch?v=5vBGOrI6yBk

Whatever your feelings about newspapers are, they have existed for hundreds of years and learnt some incredibly valuable lessons about news, journalism, reader psychology, etc. The collective wisdom of newspapers of record could fuel countless of business ventures and avoid lots of costly mistakes.

Totally. I just don't think I've ever experienced that. For as long as I can remember news has been about sensationalism at all cost.

The question was do I personally remember newspapers and the answer is no. I remember them as a thing that exists, yes, but not as a thing that provides value.

As one of the other commenters suggested, I encourage you to pick up a copy of the Economist.

I dearly love The Economist (and, indeed, have a small stack of them about a foot from me right now), but I do sometimes get the feeling that reading a newspaper with which I agree so strongly on almost all issues (hell, even the advertising feels like it's targeted at people just like me) is probably a little unhelpful in terms of giving me a balanced perspective on the world. I'm not sure what one can do to balance this, though - I picked up a copy of Nexus (https://en.wikipedia.org/wiki/Nexus_(magazine)) recently and that didn't really much help either, though it was pretty damn entertaining.

I'm not sure what one can do to balance this, though

Try reading magazines like The Spectator, The National Review and The American Conservative. Even if you don't agree with everything they write (or especially if you don't agree with everything they write), they're on the whole well written and present their view in a pretty well argued manner.

I have a friend who is a intelligent, die hard progressive who makes it a point to read the most extreme right wing news sites, just to glean how a person on the other side might view and construct the world.

Maybe that is a bit extreme, but I think Economists readers (I think the publication is center right?) can pick up a center left or far left publications from time to time.

It's a bit tricky to pin The Economist to either the left or the right; they describe themselves as being "radically centrist". This Quora answer from an Economist writer is good: https://www.quora.com/Is-The-Economist-left-or-right

  news has been about sensationalism at all cost.

Sadly, newspapers in the 19th and early 20th centuries were even worse, and unlike today, you had no way to independently research if you were being lied to. (Hearst was a classic example.)

The "collective wisdom of newspapers of record" hasn't managed to save them from their own costly mistakes, generally speaking, when you look at their current financial state.

That they are having trouble navigating the new world of the web is orthogonal to the knowledge of how to digest the world's happenings in a concise, informative manner, and (IMO) tossing all of what newspapers have accomplished out due to financial issues is a clear baby-with-the-bathwater situation.

> By the time I was old enough to even consider reading newspapers, they had devolved into clickbait sensationalist garbage.

Talk about generalization. The funny thing about poor catch-all statements like this is that I only need one counter example to make your argument silly. The Economist is one of the finest publications in the world (yes, it's a newspaper). WSJ and NYTimes additionally are totally great papers (although I only read the Economist regularly).

Maybe give these publications another try? It seems like you've limited your exposure to potentially great works.

My English wasn't good enough to understand any of those when I became interested in the topics they cover.

I pay for NYT and some stories are amazing. I finish maybe 5% of the ones I start reading and I probably attempt less than 5% of what they publish.

But, let's see. NYT headline right now: "Terror attack kills 4 in Heart of London". WSJ: "UK parliament attacker leaves 4 dead, 40 injured". The Economist: "A Brazilian meat scandal damages the country's two global producers"

Meh, sensationalism. How is any of this pertinent to me as a 29 year old engineer in SF? Other than making me afraid of my own shadow and riling up negative emotions.

Maybe the problem is that I am fundamentally just not interested in news. That seems increasingly likely. My fav NYT stories, for instance, are the researched not-news article. The Avalanche was one of the best things I ever read.

> I pay for NYT and some stories are amazing. I finish maybe 5% of the ones I start reading and I probably attempt less than 5% of what they publish.

Well, hopefully you finish more than 5% of my reply...

> Meh, sensationalism. How is any of this pertinent to me as a 29 year old engineer in SF? Other than making me afraid of my own shadow and riling up negative emotions.

Seriously? Those are the only thoughts that an article like that (it's finely written, just read it) can evoke on you?

Makes me sad to read your statements, I can only hope that you're not representative of a "29 year old engineer in SF"

I happen to be 31 years old, an engineer and also live in SF... a terrorist attack in London, does happen to be pertinent to me.

> Seriously? Those are the only thoughts that an article like that (it's finely written, just read it) can evoke on you?

It's interesting yes. It's sad that some people died. It's inspiring that life continues, that people get on with their lives and don't let themselves get bogged down because some people have stupid ideas. All of that is great.

The article even teaches me some practical things about how terrorist attacks work. How to thwart them, how to carry them out. What happens after.

Fundamentally though, London has a 0.24% death rate per year. With 8.6 million residents that's 56 dead people per day.

I'm sure those 4 people that died were very important to their families and loved ones. And the other 50-ish that died on the same day I'm sure were just as important to their loved ones.

And what am i to do with this information? How does it affect my life? I went to work today, I'll go tomorrow. I'll kiss my girlfriend goodnight and play with my parrot. I'll waste some time online. Maybe write or make a video.

I'm just as likely to die in a terrorist attack tomorrow as I was yesterday. It is overall not something worth worrying about.

Even the people I know who live in London have all said "Meh, this affects nothing. I got shit to do"

My point is that following the news is kind of toxic. Every day they find ten things you should be upset about. Why? To what end? What's the point? When one crisis blows over, they'll find the next one.

If you can't meaningfully differentiate between people dying of illness or old age and people dying from someone driving a car deliberately into them and then running amok with a knife... well, I think the problem's with you rather than with the news content you're consuming.

The people I know in London (I'm about 150 miles west of there) most certainly were not saying, "Meh, this affects nothing".

Have you been able to discern why an attack in London that kills 4 people is considered more pertinent by these bastions of journalism and not the numerous that happen every single day in Iraq and Syria, a place plunged into this power vacuum by Westminister and Washington?

Lets be honest, the pertinence of an attack in London is being established not because people dieing is pertinent, its that the people are dieing in a city that looks like yours and mine.

I don't really know what to say. It is unfortunately self-evident that when something that is unusual occurs, it is newsworthy, and when something happens every day, it becomes less newsworthy (I guess that's the "news" bit of "newsworthy").

If there were a terror attack in London of a similar magnitude every Wednesday for the next 16 weeks, I imagine the headlines in July would be somewhat more muted (though of course there would then be op-ed pieces wondering what we could do to stop terror attacks happening every Wednesday, just like there are pieces in the newspaper periodically wondering what can be done to improve the situation in Iraq or Syria).

>"Meh, this affects nothing".

As someone who lives in London (but no longer works by westminster) - this is my unpopular opinion. I'm pretty sure a bunch of people have been stabbed this year in random/gang violence but nobody is terrified by that.

Yes, it sucks that a guy in a car can intentionally run a bunch of people over. I'm sad for the police officer killed in the line of duty. It's a bad day. Literally nothing has changed.

Yeah agree.

I lived through IRA bombings although I can't recall much, I was near a pub that a neo-nazi nail-bombed in '99 and I was on a train one station along in '05. So I have trouble working out why I'm so not-scared.

I work in London and have done for 20+ years. Westminster is close by to where I currently work. I still echo this though - it affects nothing.

For those of us old enough to have lived through the IRA bombings of the early 90s, you've gotta really step up your terrorist game[1] if you actually want Londoners to react in a meaningful way.

Despite the public 'shock' shown in the media, most Londoners typically do just go 'meh', and get on with their day. London is a BIG city, 8+ million people in it every day. Unless tragedy affects you or someone you know, it's hard to get upset about it.

[1] Absolutely not condoning terrorism, just making a point.

I read the article about the Brazilian meat manufacturers. Here's how it affects you. If you eat meat, you can start to wonder where it's sourced from and what standards it adheres to. The article mentioned that Brazilian exports to the US started last year. You would also know that the vast majority of such factories are well run, unlike the bad apples that this investigation uncovered. Basically it would make you think more about the food you eat.

I don't know why the parent is being down-voted, it raises some valid points.

"Honestly? No. They're that thing my dad would buy on the weekends, read one page of, and then we'd use it for potato peels, lining the floor when painting rooms, and stuff like that."

I see no problem with not living in the time of centralised news. A quick history lesson. Newspapers are old forms of information dissemination, they do however offer a glimpse of a solution to some problems you describe, Not all.

News in the form of facts used to be collected by reporters, and journalists who would write up what they saw. The raw information would be passed to editors (subeditor in newspapers) to correct, then transcribe news into a house style fit for publishing. This is what we used to see in news. On the news-stand, the radio and to a lesser extent television.

The WWW removed this technical hurdle to publish news. Think about that. Some of the processes were technical requirements, others were there for accuracy. This resulted in (degrees) of quality news.

"By the time I was old enough to even consider reading newspapers, they had devolved into clickbait sensationalist garbage."

The money part is one problem. There are hints of what companies are trying. [0] The rivers of revenue from car sales, houses, jobs are gone. They had a monopoly and have now lost it. Journalism and news was always subsidised from these revenue streams. I don't know the answer to this.

Then there is the problem of where we get our information. The real problem now, is serious attempts to generate dis-information. Disinformation, distortion of news is not new. [2] At it's mildest it's the annoying advertising passed as news. What is going on in the Whitehouse is at the extreme end. [3] For most of us reading online we have to wade through the middle ground trying to work out what is worth reading. This is a job that could be tackled by humans trained in understanding and curating news and information, aided by smart algorythms. AI won't solve this problem (yet). It is too easy to game.

In the mean time, as Dave Winer suggests, invest carefully in your own personal news flows. [4]

reference

[0] "Ads, he said, “[are causing] increasing amounts of misinformation…and pressure to put out more content more cheaply — depth, originality, or quality be damned. It’s unsustainable and unsatisfying for producers and consumers alike….We need a new model.”"

https://thenextweb.com/opinion/2017/03/23/ev-williams-lost-g...

[1] "Since January I've been adding to a thread on how Bannon and Trump are building a disinformation validation network."

https://twitter.com/justinhendrix/status/844718202349371393

[2] If you read through the CIA reading room, search for Soviet disinformation and be surprised at the attempts to destabilise western democracy through fake news: https://www.cia.gov/library/readingroom/search/site/soviet%2...

[3] "This is a thread about #RussiaGate and Paul Manafort's $10M/year contract to further the interests of Putin's government:" https://twitter.com/AndreaChalupa/status/844710804985393153

[4] "One of the most patriotic things you can do is to upgrade the quality and breadth of the news you read. Invest in your personal news flow." https://twitter.com/davewiner/status/844691778724859904

Consider, wouldn't your NN most likely be trained on what gets clicks and inevitably lead to a clickbait feed that you're dreading?

It's a shame you got heavily downvoted, because in England the newspapers really are this fucking awful.

The guardian, telegraph, independent, and times are sort of exceptions. The FT is a definite exception.

Two UK newspapers were highlighted by the EU for providing such inaccurate coverage of migration they were likely leading to hate crime.

Just one example: Here's what Statins do, according to the front page of one newspaper. https://pbs.twimg.com/media/C6GuXQhWUAAN5F5.jpg

It's an incoherent mess of terrible science reporting blasted across the front page.

PROOF STATINS SAVE MILLIONS

STATINS IN NEW HEALTH ALERT

STATINS REALLY DO SAVE YOUR LIFE

HEALTH CHIEF SLAMS STATINS

HIGH DOSE OF STATINS CAN BEAT DEMENTIA

DOCTORS BAN ON STATINS

TAKE STATINS TO SAVE YOUR LIFE

NEW STATINS BOMBSHELL

OFFICIAL: STATINS ARE SAFE

STATINS: NEW SAFETY CHECKS

PROOF STATINS BEAT DEMENTIA

HOW STATINS CAN CAUSE DIABETES

STATINS FIGHT CANCER

STATINS SLASH RISK OF STROKE BY 30%

STATINS INCREASE RISK OF DIABETES

STATINS ADD A MERE 3 DAYS TO LIFE

STATINS DOUBLE RISK OF DIABETES

STATINS AGE YOU FASTER

NEW STATINS SAFETY ALERT

STATINS LINKED TO 227 DEATHS

PROOF AT LAST STATINS ARE SAFE

To be honest, your newspapers (primarily the tabloids) are famous/infamous for being junk. It would be hard almost anywhere in the world to print such trash and not have to suffer consequences (libel lawsuits, etc.).

This might work for most people in five to ten years.

Today, Spotify's algorithms couldn't figure out that I might be interested in a new album released by one of my most-listened-to bands last year. I didn't know about it until seeing it on Wikipedia, of all places, since I rely (over-rely) on Spotify for discovery.

Netflix is great at telling me I might be interested in stuff I've heard of but intentionally not watched cause I'm not interested in it despite superficial similarities with stuff I do like. They seem rather useless at detecting that I like a certain type/style/"quality" (this is too pretentious, but I can't come up with a better word at the moment) of movie/show regardless of genre rather than a certain set of genre tags.

And so on.

What I wouldn't do to get most machine-learning curated recommendation engines with human-curated ones.

Think of it as the difference between reading The Wirecutter and buying based on whatever Amazon.com's homepage pops up.

Or comparing a daily digest of The New Yorker with whatever shit articles Facebook bubbles up to me.

You're implying that it was a time sink because your clickthrough rate was too high.

If reading a news source isn't worth doing once a week it probably isn't worth doing once a month, once a year, or ever.

I don't think the issue is your clickthrough rate. I think the issue is that once you click through to the articles they don't contain anything of value. Your email sounds like a curated list of clickbait articles without any substance.

It sounds like they've used machine learning to curate the most clickable titles (short-term gains) while ignoring quality of content (long-term retention).

In a high-growth environment, for a startup that's less than five years old, that isn't surprising.

Reply on news.ycombinator.com