Peter Gasston

A note on Twitter’s latest feature

Twitter made a change to their algorithmic timeline recently, and have started showing tweets from strangers, that are liked by the people you follow. I don’t know why, or what benefit they offer, or even what criteria is used; I presumed at first that they’re showing tweets that have a good number of replies, retweets, or likes, in an effort to surface quality conversations.

But there are many which are replies to specific tweets, telling me nothing about the conversation or context they were used in. (Note that I’m not criticising the tweets themselves, just why Twitter thinks they’re valuable enough to show me.)

Some are so wildly out of context as to appear nonsensical, kind of like lines from a Dadaist poem.

Some are quite revealing of the tweeter’s psyche.

A few seem so personal that, although they’ve been posted on a public channel, the tweeter may not have thought they’d be seen by a wider audience.

But what it seems to massively over-index for is people liking tweets that have thanked them or praised them.

To be fair, they’re not all totally without some amusement value; every now and then you get something that’s funny because of the context in which it appears.

But mostly, they’re of little to no worth. There are occasional — once, maybe twice, a week — interesting or useful tweets that get surfaced, but they’re heavily in the minority. I can see what Twitter are trying to do with this feature, but at the moment it’s just unwelcome noise in my timeline.

Anyway, I don’t like to simply criticise without being constructive, so I’d like to offer a solution to fix it. Here’s a mockup of a simple toggle to let people choose whether or not they want to see these tweets:

You’re welcome, Twitter.

Trends in digital media for 2017

Alright, stand back everyone: I’m about to have some opinions about technology in 2017. Because obviously there’s been a shortage of those.

As part of my Technologist role at +rehabstudio I put together internal briefings about digital media, consumer technology, where the digital marketing industry could go in the near future, and what we should be communicating to our clients. Not trying to make predictions, but to follow trends.

This article is based on my latest briefing. It’s somewhat informed, purposely skimpy on detail, and very incomplete: I have some thoughts on advertising and publishing that I can’t quite distil yet, and machine learning is a vast surface that I can barely scratch.

If for nothing more than press coverage, 2016 was the year of messaging, and the explosion of the messaging bot. The biggest player in the game, Facebook’s Messenger, launched their bot platform in April, and by November some 33,000 bots had been released. Recent tools added to the platform include embedded webviews, HTML5 games, and in-app payments.

The first six months of bots were largely the ‘fart app’ stage, but there are signs that brands and services are finally starting to see the real opportunities in messaging: removing friction from their users’ interactions with them. Friction in app management and UI complexity, for example.

The same removal of friction is also a key driver behind the growth of home assistants and voice interaction, like Alexa. Removing the UI abstraction between users and tasks is a clear trend. As an illustration, compare two user flows for watching Stranger Things on Netflix on your TV; first using a smartphone:

Unlock phone.
Find and open Netflix app.
Press the ‘cast’ button.
Find ‘Stranger Things’.
Play.

Now using Google Home:

“OK Google, play Stranger Things from Netflix on My TV.”

Home assistants make the smart home easier to manage. No more separate apps for Wemo, Hue, Nest, etc; a single voice interface (perhaps glued together with a cloud service like IFTT) controls all the different devices in your home.

Messaging and voice are visible aspects of the trend towards the interface on demand:

The app only appears in a particular context when necessary and in the format which is most convenient for the user.

While native mobile apps are still a growth area, it’s becoming much harder to get users to download and engage with apps outside of a small popular core. This is especially true for retail, where consumers are more omnivorous and like to browse widely.

Improvements in the capabilities of web apps (especially on Chrome for Android) suggest an alternative to native apps in some cases. This has been demonstrated by the success of new web apps from major retail brands like Flipkart and Ali Baba in developing economies where an official app store may not be available, or network costs may make app downloads undesirable.

Web apps require no installation, avoiding the app store problem. They’re starting to get important features like push notifications and payment APIs. And messaging platforms, with their large installed user base, provide the web with a social and distribution layer that the browser never did:

Messaging apps and social networks [are] wrappers for the mobile web. They’re actually browsers… [and] give us the social context and connections we crave, something traditional browsers do not.

So it may be that for some brands, a website optimised for performance, engagement, and sharing, along with a decent messaging and social strategy, will offer a better investment than native apps and app store marketing. Patagonia already closed their native app. Gartner predict that some 20% of brands will follow by 2019:

Many brands are finding that their mobile apps are not paying off.

The most important app on your phone could be the camera, which will be increasingly important this year. First, by revealing the ‘dark matter’ of the internet: images, video and sound. So much of this data is uploaded every day, but without the semantic value of text, it’s meaning is lost to non-humans — like search engines, for example. But machine learning is becoming very good at understanding the content of this opaque data, meaning the role of the camera changes:

It’s not really a camera, taking pictures; it’s an eye, that can see.

It can see faces, landmarks, logos, objects; hear background chat and music. That’s understanding context, location, purchase history, and behaviour, without being explicitly told anything. This is why Facebook, through Messenger and Instagram, are furiously copying Snapchat’s best features: they want their young audience and the data they bring.

Will it be intrusive? Yes. Will it happen? Yes. I’ve tried to avoid making hard predictions in this piece, but I am as confident as I can be that our image and video history will be used for marketing data.

Cameras will also be important in altering the images that are shown to the users. Augmented reality is an exciting technology, although good-enough dedicated hardware is still a while away. But there’s a definite market drift in that direction, and leading it is Snapchat: they’re stealthily introducing AR through modifying the base layer of reality—first, by altering faces using their lenses. This isn’t frivolous; it’s expanding the range of digital communication, like emoji do for text.

If people are talking in pictures, they need those pictures to be capable of expressing the whole range of human emotion.

Recent Snapchat lenses have started altering voices, and your environment. They’ve recently bought a company that specialises in adding 3D objects into real environments. With Spectacles they’re not only removing friction from the process of taking a photo, they’re prototyping hardware at scale. This is the road to AR. Snap Inc. want to be the camera company — not in the way that Nikon was, but in the way that Facebook is the social company.

The companion to an augmented reality is a virtual one, but I don’t believe we’ll see VR going mainstream in 2017—and I say that as a proponent. It’s static, isolating, and it requires people to form a new behaviour. It’s interesting to see creators experiment with the form, and I’ve no doubt that we’ll see some very interesting experiences launched this year. But domestic sales aren’t huge, and high-end units are too expensive, and low-end not quite up to scratch yet. Still think it will be big for gamers, though.

I have more. A lot more. But I think it will all be better explained in a series of subsequent blog posts, so I’ll aim to do that. In the meantime, would love to hear your thoughts, arguments, objections, and conclusions.

My Favourite Books I Read in 2016

I tend to have at least two books on the go at any one time: one fiction, one non-fiction. I read fiction when I go to bed, since I read somewhere that fiction encourages present-state attention, which makes you feel sleepy. It works for me. I generally read non-fiction (or, more often, Pocket articles) when I’m commuting.

My Goodreads Year in Review tells me I read 33 books last year. These are the highlights.

The best book I read was John Higgs’ Stranger Than We Can Imagine, an attempt to explain the 20th century through philosophy, art and science, rather than geopolitics. I wrote a post about it which you can read if you want more detail; but if you’re willing to take my word, it comes with a strong recommendation from me.

The Inevitable, by Wired founder Kevin Kelly, looks at technologies which will shape the near future. Not specific implementations, but more general trends: sharing, remixing, tracking, etc. If you keep up to date on tech trends some of this can seem like it’s just reinforcing what you already know; even so there are enough interesting points of view and insights to make this a good and compelling read.

Time Travel, by James Gleick, is an exhaustive (and occasionally exhausting) exploration of its subject in fiction, philosophy, and physics. I didn’t enjoy it as much as his previous book, The Information, but it’s still worth your time.

Andrew Hosken’s Empire of Fear is a history and investigation of the so-called Islamic State (as the BBC put it). It really helped me better understand the complicated situation in the Middle East, and the shameful decisions by foreign powers that made it all happen.

You Could Do Something Amazing With Your Life (You Are Raoul Moat), by Andrew Hankinson, is a semi-fictionalised first-person account (it uses real dialog, social services documents, and police reports) of the last days of the man hunted by police in 2010. Phenomenal true-crime writing.

In comics, Steffen Kverneland’s Munch is both an incredible biography of the Norwegian artist and his relationship with the author August Strindberg, and a fourth-wall-breaking story of how the book was written. And that barely scratches the surface. It apparently took seven years to create, and that’s apparent in the breadth and detail.

Mary Wept Over The Feet of Jesus, by Chester Brown, retells (with copious footnotes and reference) Bible stories that feature prostitutes. It’s part of the author’s ongoing attempts to contextualise and justify his own use of paid sex, and is quite fascinating.

The novel I enjoyed most was Don Winslow’s The Cartel, a story of the drug wars in South and Central Americas (and sequel to The Power of the Dog). It’s a robust thriller that only occasionally slips into cliche.

Finally, Beast, by Paul Kingsnorth, and Pig Iron, by Benjamin Myers, are very different stories but both are first-person, and use the language of the narrator, and their landscape and environment, to create a feeling of deep immersion. Both authors are poets, which shows.

I’ve got three books on the go right now which didn’t quite make it into this roundup, and another ten in my to-read list. Exciting and daunting.

Making Nature: how we see animals

A visit to the Wellcome Collection this week, for the exhibition Making Nature. It explores human interaction with animals; how we classify them, display them, observe them, and change them. From Walter Potter’s taxidermy tableaux, to tigers in Manhattan apartments, to BioSteel™ goats that lactate spider silk, it’s a well-curated, thoughtful, and eventually unsettling experience.

It starts with Linnaeus’ taxonomies, the desire to impose order on the natural world. Perhaps well-intentioned, but his racist human stereotypes (‘indolent and capricious Africans’) indicate that order is as much about opinion as fact.

Making Nature shows how our view of animals changed from asset to resource to commodity, to entertainment and decoration. This observation struck me:

Humans soon discovered they could train captive birds to sing songs that may not attract mates in the wild, but would captivate human listeners.

The songbird became a programmable musical device.

Perhaps the saddest part of the exhibit was the video installation The Great Silence, by artists Allora & Calzadilla with author Ted Chiang. It contrasts shots of two locations in Puerto Rico—the Arecibo observatory for monitoring signals of alien life, and a sanctuary for endangered parrots—simultaneously with Chiang’s eponymous short story, written from the point of view of one of those parrots.

Humans have lived alongside parrots for thousands of years, and only recently have they considered the possibility that we might be intelligent.

But parrots are more similar to humans than any extraterrestrial species will be, and humans can observe us up close; they can look us in the eye. How do they expect to recognize an alien intelligence if all they can do is eavesdrop from a hundred light years away?

The exhibition is, like everything in the Wellcome Collection, free to visit. It runs until 21st May, and is the first in a year-long series exploring our relationship with nature.