About me
Bookshop

Get new posts by email.

About me

In summary, no thanks

One of the commonly promoted function of generative artificial intelligence tools like ChatGPT is summarising long pieces of text. Ryan Broderick recently wrote in Garbage Day:

The assumption that people want summaries of information when they receive news, is also a funny one. It seems to come around every four-to-eight years. Typically when Democrats are in the White House, I’ve noticed. This was the impetus behind Vox, for instance, with its big initial claims of inventing “explainer journalism,” which quickly just devolved into blogging, again. My own assumption here is that this is a byproduct of CEO brain. “I can’t possibly read all of the information I need to pretend to care about to run my company, so other people must treat information as a nuisance to be fixed, as well.” But, once again, that is not really the case. The internet has turned the consumption of information into its own form of entertainment — or in the context of conspiracy theories, madness.

This is something that I’ve often thought about, too.

There are some scenarios where an AI summary can be useful. Occasionally, I’m copied into long email chains with a vague subject line and have to spend time scrolling up and down to orientate myself as to what the conversation is about. I occasionally skimp on that step and end up missing the point. A couple of auto-generated sentences saying ‘this is a series of emails discussing x, with the goal of producing y, looking for input on aspect z’ can be a godsend.1

And yet, the products which advertising tends to push at me most frequently are services which offer summaries of things which I don’t think benefit from summarisation. The commonest one is books. Short books distilling the ‘key messages’ of longer books are clearly popular, and pre-date the web, let alone generative AI, but I’ve never really understood the point. The format assumes that books are about imparting a series of facts; in my experience, most are actually about encouraging readers to think differently about subjects. Even in the simplest airport bookshop management paperbacks, the identification of key messages is highly subjective.

With human-authored summaries, we can at least have a sense of whether we trust the subjective judgements of the summariser, but this becomes much trickier with the black box of artificial intelligence. Summarisation usually involves value judgements, and they are not easily ‘outsourced’ to AI. This is a problem when summarising books, but even more so in summarising news.

Alan Rusbridger recently commented on his podcast that an experiment using AI summaries to generate key points to draw people into reading Prospect articles had impressed both him and the magazine’s writers. But that’s a different goal from relying on the summary instead of reading the article. It would be a bit like relying on a headline rather than reading the full story… which seems to be common behaviour, not necessarily a behaviour for which proliferation benefits humanity.

It strikes me as unfortunate that we’re building tools—or at least promoting the ability of tools—to allow people to engage more superficially with subjects than they already do. It’s not like humanity is short of examples of the downsides of people engaging only with the headlines and glossing over the detail.

Fortunately, like Broderick, I doubt that’s what people are seeking in practice, I think people are more into ‘deep dives’. Broderick attributes the miscalculation to ‘CEO brain’; I’d attribute it, at least equally, to ‘social media brain’. The ‘BREAKING’ and ‘HUGE IF TRUE’ style of sharing information in bitesize chunk on social media might suggest that people like consuming information in vastly abbreviated forms, but I don’t buy it. I think those interactions are much more about socialisation than about assessing information.

Fortunately, generative AI can work perfectly well in the other direction, too, recommending books and sources that can help people to explore a topic more deeply. I think this would make the more interesting tool: not ‘summarise this webpage’, but ‘recommend another three web pages which explore this subject in more detail’.

The obvious difficulty in making such a tool work is the rabbit-hole phenomenon, much-discussed in the context of the YouTube algorithm. How do you imbue such as system with the sensitivity and awareness to avoid pulling people into ever-more extreme versions of conspiracy theories, for example?

It’s a difficult problem, but one that equally needs solving to make summarisation engines work in a reliable and trustworthy way. Let’s hope someone can tackle it.


  1. Experience over many years had taught me to start my own replies to email chains like that with my own couple of sentences, starting ’I understand from the below that…’. It’s a technique which can nip misunderstandings in the bud.

The image at the top of this post was generated by DALL·E 3.

This post was filed under: Technology, , .

Snowflake says the world is ending

In Garbage Day recently, Ryan Broderick wrote:

About three years ago, Atlantic writer Charlie Warzel tweeted something that has always stuck with me. In regards to the then still-emerging coronavirus outbreak, Warzel wrote, “The coronavirus scenario I can’t stop thinking about is the one where we simply get used to all the dying. There’s a national precedent: America’s response to gun violence.”

He was, of course, right about that. But I think he inadvertently summed up America’s national response to pretty much every large-scale systemic crisis we’re bound to face going forward. The weather’s going to keep getting worse and the right-wing media ecosystem will downplay or outright deny it, and, more often than not, they’ll find ways to tie the acknowledgement of it to definitions of masculinity. And just like the folks who thought they could ride out the pandemic without a vaccine or masks because they were tough, only to end up on ventilators, so too will a lot of folks get hurt trying to man up and ignore the rising temperatures.

This is one of those sets of observations that ties together a lot of disparate strands of thought, making me see something slightly differently—and while Broderick (and Warzel) were writing about America, I think their ideas are more generally applicable.

Of course I’ve heard people dismiss the impacts of climate change, saying that the temperatures aren’t as high as they were in 197X. Of course I’ve seen how gun violence is seemingly just accepted in the States, where other considerations and vested interests have come to count for more than human life. Of course, I’ve witnessed the rapid spread of harmful anti-scientific views during the COVID-19 pandemic. Of course I’ve seen the proliferation of toxic ideas of masculinity and the dismissive branding of those expressing genuine concerns as ‘snowflakes’.

But I’ve never before seen so clearly how the combination of those things could conspire in the context of climate change. It’s sobering.


The image at the top of this post was generated by Midjourney.

This post was filed under: News and Comment, Post-a-day 2023, , , , .

Pulling threads

I’m predisposed to dislike Meta’s new social media app Threads. It’s years since I left Instagram and Twitter, and I haven’t missed either of them. I’m therefore unlikely to be convinced by something which seems to be a combination of the two. But equally, my lack of engagement means that I’m not well-informed, and my opinions ought to be taken with a pinch of salt.

Yet, Threads has been unavoidable in recent days, with acres of press coverage. In this post, I want to reflect on some of the things I’ve read.


In Tedium, Ernie Smith wrote:

I don’t feel particularly motivated to write about Threads, the Instagram-in-Twitter-form social network that Meta launched this week. It feels like a social network that exists to check a box for a company that owns a lot of social networks. But I do think it represents something about the moment that we’re in, where people are so desperate for a certain kind of experience that they will go from the hands of one billionaire to another, in hopes that will give them what they don’t feel like they were getting before.

It would be a bit rich of me to claim that I don’t feel motivated to write about Threads at the start of a blog post about Threads. But I do feel removed from it. I can’t imagine I’m ever going to join the service, and—in all honesty—I think the moment for this sort of short-text-based social media service has probably passed.


In his Platformer newsletter, Casey Newton interviewed Instagram’s Adam Mosseri about the new product. This bit stood out to me:

The thing that makes Twitter distinctive, Mosseri said, is that replies are given the same visual priority as the original posts. In a world where every other social network buries comments underneath posts, Twitter elevates them. And that encourages people to participate in discussions.

“The post-and-comment model is great,” Mosseri said. “But it really does not support public discourse nearly as well as the tweet-and-reply model. Elevating the reply to the same level as the original post allows for much more robust, diverse discourse.”

I fundamentally disagree with this: I think giving replies equal prominence to original posts fuels terrible discourse. It pits experts on the same level as conspiracy theorists. It encourages people to seize on the tiniest error, the most abstracted perceived slight, in any Tweet in order to ramp up engagement with the reply. It encourages attack, not discussion, and pile-ons (a natural by-product of giving responses the same prominence as original posts) squeeze out diversity rather than encouraging it.


In The Atlantic Daily, David Graham wrote about the impact of the death of Twitter on the ability of journalists to build personal brands. He said:

What comes after Twitter is a much more fragmented landscape. Many social-media sites command significant audiences, but no single platform can do what Twitter once did. A journalist can make a big bet on one platform, or they can try to hedge and be active on Reddit, YouTube, TikTok, Substack, and, as of this week, Meta’s Threads—give or take a dozen more. But who has the time? And besides, you don’t get the same reach. TikTok and YouTube command enormous but typically niche audiences.

I’ve written with boring frequency on this site about the BBC’s promotion of Twitter. This is, in my view, the wrong approach for myriad reasons, but the more the social media landscape fragments, the more unsustainable that approach becomes.

Imagine a 2024 Presidential election fought between Joe Biden and Donald Trump. Suppose Biden continues to campaign on Twitter, but that Trump restricts himself to Truth Social, safe in the knowledge that he won’t be banned or fact-checked and that his messages will find their way beyond the network anyway. Would the BBC continue to promote only the platform used by only one of the candidates? Self-evidently not. Would it promote Truth Social? No chance.

And so, the era of free promotion of a particular network ends… surely. We presumably revert to “Campaign X said Y”, rather than “Candidate A posted on social network B”. The latter has never made sense for the BBC anyway: it’s no more logical than saying “Party M has posted a press released on media repository N”, which has never been a thing.

Threads won’t benefit (in the long-term) from the free promotion that built Twitter.


FullFact recently posted about false news circulating on Facebook suggesting that animals had been released from Paris Zoo in recent riots.

This stood out to me because I well remember, in the days I was on Facebook, the frequency with which “escaped zoo animals” were reported in connection with any number of news stories. I remember this because I briefly parodied it in an email to partners after arranging rabies vaccinations for an individual bitten by a wild animal abroad.

This is a long-standing and easily disproved item of fake news which could reasonably cause unnecessarily fear among readers. Even all these years on, Meta hasn’t got a grip of this simple stuff. The idea that Threads will be a reliable source of anything, or a place for informed discussion, is clearly not rooted in reality.


In Garbage Day, Ryan Broderick wrote:

My verdict: Threads sucks shit. It has no purpose. It is for no one. It launched as a content graveyard and will assuredly only become more of one over time. It’s iFunny for people who miss The Ellen Show. It has a distinct celebrities-making-videos-during-COVID-lockdown vibe. It feels like a 90s-themed office party organized by a human resources department. And my theory, after staring into its dark heart for several days, is that it was never meant to “beat” Twitter — regardless of what Zuckerberg has been tweeting. Threads’ true purpose was to act as a fresh coat of paint for Instagram’s code in the hopes it might make the network relevant again. And Threads is also proof that Meta, even after all these years, still has no other ambition aside from scale.

If you look back at every era of the internet, towards the end of each, you’ll see a whole pile-up of forgotten apps that tried to swoop in and bring back or replace something that was dying or dead. Myspace tried to bounce back with a redesign. Vine was going to be resurrected as Byte. BeReal was marketed as a way to recapture the simple fun of Snapchat, etc. But it doesn’t really work that way. Our tastes change. We move on. And then suddenly we can’t imagine ever going back.

I’ve wrongly predicted the death of Twitter an embarrassing number of times. I even did it earlier this month. I think Broderick might be right that it’s the class of services that is fracturing and dying, not just Twitter specifically. And that bodes ill for Threads.


In The Atlantic Daily, Charlie Warzel is interviewed about the impact of Meta’s appalling privacy record on the potential success of Threads:

People are never going to be as concerned with privacy stuff as they are with I want to be where my friends are. I want to try something new and interesting and see if it works. And if that initial experience is easy, fun, and intriguing, the potential to hook a new user and turn them into a quality repeat customer is very high.

It’s fair to say that Warzel’s overall impression of the prospects for Threads is less positive than this brief quotation implies. But this quotation stood out for me because I think Warzel is right.

It was the ‘creepy line’ that drove me away from Google services. Like most people, I think I projected my views onto others, and assumed that everyone else is increasingly concerned about privacy too. Except, objectively, they are not.

Warzel’s words reminded me that I often overhear people talking nonsense about Facebook spying on them through their phone’s microphone and inserting related adverts into their feeds. Yet, none of them have deleted the app or left the service… even when they genuinely believe they are being covertly surveilled.

Privacy is not a big deal to most people, which probably works in favour of Threads.


It’s probably not in their favour that when I think of ‘Threads’, I think of this:


The image at the top of this post was generated by Midjourney.

This post was filed under: Post-a-day 2023, , , , , , , , , , , , , , , .




The content of this site is copyright protected by a Creative Commons License, with some rights reserved. All trademarks, images and logos remain the property of their respective owners. The accuracy of information on this site is in no way guaranteed. Opinions expressed are solely those of the author. No responsibility can be accepted for any loss or damage caused by reliance on the information provided by this site. Information about cookies and the handling of emails submitted for the 'new posts by email' service can be found in the privacy policy. This site uses affiliate links: if you buy something via a link on this site, I might get a small percentage in commission. Here's hoping.