No big stories this week, but lots and lots of little ones – including a subscriber suggestion. If you ever come across anything you think other subscribers might find interesting, please do send them my way! Just hit reply to this email, or message me on Twitter.
Betsy O’Donovan and Melody Kramer have put a name to the practice of maximising the use you get from content: zero-waste journalism. It’s an idea I’m very fond of: so much evergreen content is wasted and never resurfaced, when it could still generate tons of utility for readers. Their article talks about how women are particularly suited to creating it because they’re better at thinking about resources – would be good for more men to start thinking this way too.
Condé Nast has got a new CEO, and he doesn’t have any publishing experience. I think that’s probably for the best: Roger Lynch has a great background in transforming businesses at Pandora and Dish Networks, and a lack of preconceptions about magazines will hopefully help him to see all the very obvious places for budget cuts that Condé is otherwise blind to. Article about him here: sounds like a great guy.
Nieman have published a long report about smart speakers and audio news. It seems obvious, but right now the best UX on speakers is traditional live radio. I think that’s partly because it’s a constant feed, rather than a staccato series of briefings which requires you to re-interact with it after a few minutes. Makes me think Google’s personalised audio feed will be really successful: all publishers should be getting involved with that if they’re not already. Also some interesting points about conversational interfaces (e.g. users asking speakers “What’s going on with Brexit?”), specifically regarding tone. From NPR’s Tamar Charney: ‘"What we were aiming for is a more genuine answer to a question.” Charney says the goal is to have updates sound more like the “engaging, compelling, human-to-human answer” that an actual person would give.’
Stylist have launched a fitness studio, which is such an obviously great idea you wonder why no one else has done it already.
Entrepreneur First had their Demo Day the other week and there are two companies you need to know about. Speech AI have got human quality voice generation (they’re using it for video games, which is v smart), and Presscast have developed a system that lets marketers add paragraphs of texts to news articles as an alternative to advertising. Watch the latter’s demo: it’s very ethically dubious and potentially illegal because there’s no disclosure that the text is an ad? Very gross.
On the topic of games: this very clever software from RADICAL can make an animated 3D model from a video (none of those weird pingpong ball suits required). Could revolutionise all industries that use motion capture.
Reuters made their own deepfake video to try and train reporters on how to spot them. This is only going to get harder: thankfully companies like Deeptrace are working on automated detection tools, and for text it sounds like there might be hope too (the same AIs that generate the fake text can be used to detect it).
Automated text generation just won’t stop getting better though: Springer have just published the first machine-generated research book (about batteries). It’s…not bad! (PDF here)
The BBC have also started using automated news to generate local data-driven stories. They’re using Arria’s platform (massive get for them), and though it’s not groundbreaking, the Beeb using a new technology always feels like a coming-of-age moment.
It sounds like DMGT (parent company of the Daily Mail) is about to invest in a new used-car sale website, which is smart (they previously made £640m from their investment in Zoopla, which was founded by the same guy as this new platform).
Interview with the FT’s CEO suggests international growth is about to become a huge focus for them: particularly India and China. They’ve just hit 1 million subscribers – staff weren’t impressed with Google’s congratulatory gift, which was “the wrong shade of pink”.
The Guardian are very cleverly using metadata to automatically include the publication year of an article on social media preview images, to help stop old content being misleadingly shared as if it’s new. Nieman Lab have some great ideas for what else you could do with this tech: I particularly like the idea of telling Twitter users when a liveblog was last updated, or using the images to factcheck old stories.
The UX of Apple News Plus is sorely lacking, it seems – The Verge have called it “messy and inconsistent” (though they do say it’s a great deal if you want to read all those mags). The faults sound fixable, though Apple tends to be slow at fixing things. Still not available outside America so I can’t try it…
Clearbanc have developed what’s effectively a well-structured loan for retail startups that want to increase their ad-spend: they’re calling it a “20 minute term sheet” which is a bit of a stretch, but it’s a good idea nonetheless.
Great piece on Netflix’s plans for kids content. Seems their strategy is to give complete creative freedom, which should differentiate them from Disney – the question is whether being creatively better will make up for weaker IP (I don’t think it will, kids care more about characters than artistic brilliance).
Shared with me by subscriber B.S., this is an incredible piece by someone who took a job in an Amazon warehouse to see what it’s like. The stuff about gamification is so dystopian it’s hard to believe it’s real.
Final Thought
This is a long digression about the influence of media on politics, feel free to ignore if you’re just here for the fun tech and innovation stuff.
Everyone has been congratulating the massive New York Times investigation into the Murdoch family. It’s decent and interesting, but absolutely not worth the praise it’s receiving. There are massive inaccuracies (citing It Was The Sun Wot Won It splash as evidence of Murdoch revelling in his power, when it’s known that he was furious with the editor for publishing it; claiming over a million people were on the anti-Brexit march the other week when that number has been pretty conclusively rubbished), which is pretty sloppy for a big investigation.
But more importantly, the piece exaggerates Murdoch’s influence. Ignorance of academic research into media has always been a problem in media reporting, and this is a perfect example. The writers attribute every major political change in the US, UK and Australia to Murdoch. Admittedly that’s a fairly common view: the media loves to think that it has influence.
But it’s not true! So much research has been done by political sociologists into the effect of the media on election results, and the studies consistently find that the media’s impact on elections is very, very minimal (if you want copious notes/sources on this, let me know and I can email you some stuff).
The article cites a 2007 study that says Fox News “pushed local voters to the right”: but the much more widely-accepted 2013 study from Levendusky found that Fox doesn’t make people right-wing, it just makes already very right-wing voters even more right-wing. That’s a subtle distinction, but an important one. In the same vein, the article says Murdoch was “a sense responsible for unleashing the forces that were now propelling Trump’s rise” – but has no explanation for why the same nationalist/racist ideologies have flourished in countries where Murdoch has no influence.
The direction of causality matters. The academic evidence suggests media doesn’t influence views, it reflects them. The horrible views that Fox espouses came from somewhere – the public. And if we don’t understand what caused that, we can’t tackle it. Attributing everything to Murdoch is a lazy explanation, because it frees us from taking responsibility – from considering that maybe it was a generation of bad politics and an exclusionary economic system that led to Trump.
In short: can media reporters please start reading academic literature rather than wildly speculating! Thank you! Have a lovely week!