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.

One of the joys of reading the news is the occasional opportunity for disparate articles to end up explaining one another.

Ernie Smith’s Tedium yesterday was about Facebook’s new project to ask users to consent to a new feature called ‘link history’. This appears to be a feature built solely to give users a unobjectionable justification for Facebook to continue to collect data which is primarily used for advertising.

Ernie linked to a Gizmodo article by Thomas Germain which said:

When you click on a link in the Facebook or Instagram apps, the website loads in a special browser built into the app, rather than your phone’s default browser. In 2022, privacy researcher Felix Krause found that Meta injects special “keylogging” JavaScript onto the website you’re visiting that allows the company to monitor everything you type and tap on, including passwords. Other apps including TikTok do the same thing.

I find it astonishing that Facebook is harvesting people’s passwords for other services, and yet this is neither major news nor has prompted a mass exodus from the platform.

Yet, I often hear people discussing with certainty the conspiracy theory that Facebook covertly analyses continuous audio recorded from people’s phones to target advertising. Confirmation bias provides ‘evidence’ for people. It’s essentially nonsense, but it’s accepted as fact by many people.

So… what’s going on here? Why would people choose to keep using a service that violates their security and which they believe to spy on them? Why are so many people still active users?

In Platformer, Casey Newton shared a link to an article by Hannah Devlin in The Guardian which answers those questions:

Almost half of British teenagers say they feel addicted to social media … The finding, from the Millennium Cohort study, adds to evidence that many people feel they have lost control over their use of digital interactive media.

People feeling addicted to products which don’t have their best interests at heart is a depressing situation, though I guess it’s a common one.

I quit social media in 2020, for no better reason than noticing that my mood after opening the apps was typically worse than my mood when logging on. I can’t claim that I’ve become a new person, reclaimed hours of time, or cast off any psychological shackles. But I can say, without a scintilla of doubt, that I don’t miss it at all.

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

When I was about 14, I wrote some software in QBasic to support the running of the school library. Hitherto, the library catalogue and loans had been stored in a spreadsheet. I replaced it with bespoke software.

Over the next couple of years, I iteratively tweaked the software based on feedback until I finished my GCSEs and left. I think I learned more from that process than from any computing course I’ve ever taken: I also came to love Greg Perry’s book QBasic by Example, which—ironically—wasn’t to be found in the library’s collection.

In my A-level computing course, I was required to write in Visual Basic instead. I remember finding this irritating, deriving much less satisfaction from developing software using a graphical user interface rather than a blank screen with an expectant blinking cursor. It felt like a little bit of power and control had been taken away, as though the final product wasn’t completely mine in the same way.

I’ve been thinking about that a lot lately. These days, there’s nothing I create from scratch as I did in QBasic. The closest thing is writing for this blog: it’s only here that I start with a blank screen and a blinking cursor. Only here do I exercise complete control over every word and every pixel.

I’m sometimes surprised at my own tenacity in keeping this blog up for more than two decades. It’s not for anyone else’s benefit that it continues to exist. I’m sure no one else would think twice if it disappeared tomorrow. Yet, I’ve published over 760,000 words here in nearly 3,000 posts over more than twenty years. I suppose it must scratch a creative itch, even if I struggle to explain precisely where that itch is located.

Maybe I’m just a control freak at heart.

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

In 1999, I registered the domain name simonhoward.co.uk, and I’ve owned it ever since.

There was a certain amount of foresight in that decision. I was only fourteen years old, and so I’m probably lucky that I didn’t register something wackier.

There are a lot of Simon Howards, so I’m lucky to be the one who snapped that address up early—I wasn’t early enough to get the .com, which these days hosts the website of one of many models who share my name. Thank goodness it wasn’t taken by the aristocratic child abuser who already dominates search engine results.

These days, I own 23 domain names. This is patently absurd when I only have one main website. I separate my blog (sjhoward.co.uk) from my CV (simonhoward.co.uk) from the bookshop (simonsbooks.com), which accounts for a few.

Four others are domains I registered for specific temporary projects this year, which I have no intention of renewing.

A few are handy: sjh.bz allows me a very short email address, which is occasionally handy when my main one is too long.

Others are obvious variations on my main domains to avoid confusion (sjhoward.uk and sjhoward.com).

Some are addresses I just like, though with no practical use. There’s one that would make an excellent public health company name. I recently bought simon.how because it’s the first eight letters of my name, but what will I ever do with that?

This all leads back to the same point: most people in the world have no use for a domain name, and it’s clearly ridiculous to have 23 of them, no matter how I try to justify it!

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

You’ll no doubt have seen that Spotify cancelled the podcast Heavyweight this week, alongside a tranche of redundancies. Ironically, I’d introduced Wendy to the show hosted by Jonathan Goldstein only a fortnight ago. I’ve listened since the start, though in a pretty on-and-off fashion. The episodes were so well-crafted and moving that it felt weirdly disrespectful to listen to them as background audio, which ironically meant that I didn’t get around to hearing them.

The best thing I’ve read on the topic is PJ Vogt’s insightful piece on the cancellation of Heavyweight and the broader state of the podcast industry. Of his Search Engine podcast, which I also enjoy, Vogt says:

We have a budget that’s good until July. After July, we’ll see.

One of the comforts of working as a doctor is the confidence in future employment. Even in public health, which is undoubtedly the rockiest of medical specialties in those terms, it’s easy to have a high degree of confidence that doctors will always be required in the system somewhere, even if the system is dismantled every five years or so.

Podcasting, by contrast, sounds terrifying: most of my job is about dealing with uncertainty, but I couldn’t sleep if I had no idea if my job would still exist in a few months. The creative drive of people like Vogt and Goldstein, their willingness to sacrifice certainty to make journalistic and artistic products, is truly something to behold.

I’m reminded of how fortunate I am to have a job that allows me to pursue my interests while maintaining a degree of personal assurance. Most people aren’t so lucky.

James O’Malley opens his latest Substack post:

Twitter’s most underrated function is that it is a tool for social mobility.

After graduating into the financial crisis, I moved to London and I didn’t know anyone. My parents are not super wealthy, I went to a normal comprehensive school in the Midlands – and then to a post-1992 university. So when it came to getting a job and building a professional network, I was essentially starting from scratch.

But the one asset I did have was that I was a Twitter early adopter. Through the platform I discovered an entire professional world that would historically have been closed off to me, and as I followed journalists, politicos and other high-flyers, I was able to glimpse into their world, learn the industry gossip, identify the behind-the-scenes power-players and master the terms of art. Scrolling my timeline was like having a permanent spot at the water-cooler of London’s political and media elite.

I’ve never previously considered Twitter from the perspective of social mobility, but now that James has pointed it out, it seems obvious. Indeed, when I was active on Twitter in the early stages of my career, it helped me to make professional connections in a similar way. Public health in England is a small professional world, and as a result, it is disproportionately London-centric. Twitter helped me to make professional connections even from afar.

Yet, looking back, many of the people I made professional connections to were the—for want of a better phrase—‘extremely online’ colleagues, who were often not as influential, switched-on or expert as they thought they were.

Last week, we heard complaints at the COVID inquiry about politicians’ inability to distinguish between absolute and relative risk. Not many years ago, I did my best to explain the same to an uncomprehending prominent professional by direct message on Twitter. They may have been rarely off the telly and hugely popular on Twitter, but the conversation revealed their grasp of the basics to be a little loose. Another told me that a study they shared was significant because of a wide confidence interval, which they believed to demonstrate a very high level of confidence: an almost perfect inversion of the true meaning.

I came to learn that the people with deep knowledge and real, lasting influence tended not to be active members of the Twitterati. In my field, at least, when ‘extremely online’ people reach positions of real-world seniority, it can become somewhat fraught. Public health is a naturally complex, subtle and political field which perhaps doesn’t reward talent for brief, blunt, off-the-cuff responses. There are exceptions, of course, and I’m not inured to the irony of writing all of this in a blog post.

In politics and journalism, public engagement is the job, so naturally, excellent practitioners will often do well on social media. In these fields, Twitter is a perfect tool for social mobility, precisely as James describes. In medicine, I’m not so sure.

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

Three news stories have played on my mind recently:

• In his latest newsletter, Benedict Evans updates on the seemingly never-ending Horizon scandal. The Post Office rolled out a buggy computer system which wrongly calculated cash balances held by branches. The assumption that money had been stolen led to almost a thousand people being wrongly prosecuted. Some people ended their lives.
• As reported in the Guardian, one of my local hospitals failed to send 24,000 letters over five years ‘due to a new computer system’.
• The FT reported that an incorrectly weighted computer algorithm for allocating liver transplants led to specific groups of patients waiting longer than others.

All three of these stories are similar: they are presented as though the problem is with computers when it is actually with management. In all three cases, managers knew about concerns with the computer systems, yet the organisations involved did not take timely corrective action.

I don’t know the inside story behind any of these incidents, but I’m willing to bet that all involved managers saying something like ‘I’m not good with computers’—and failing to engage as a result. In some organisations, it’s a daily occurrence to hear senior leaders say things like, ‘Ooh, I’m not sure how to share my slides; is it working?’ That discomfort writ large means that they subject IT processes to less scrutiny.¹

The irony is that the consequences of all three failings were in the real world, not inside a computer. In all three cases, it seems likely that a reasonable and non-technical workaround would have been to stop using the computer system until it was fixed. The real-life resolution required no knowledge of how the systems worked.

One can’t imagine the same managerial non-response to a problem if, say, the underlying issue was secretaries who weren’t bothering to type letters or an accountancy team consistently failing to produce reports.

And therein lies the issue. Even in the third decade of the twenty-first century, it is still acceptable in many organisations for senior leaders to freeze up over any problem involving computers. Nobody expects them to be IT experts, but they ought to be experts in leadership and management and not simply give up when the situation involves a computer.

The greater development of artificial intelligence means that ever more processes will be completed in an automated fashion in the coming years. This will result in progressively less oversight of day-to-day work within organisations whose senior staff seize up over this stuff. For them, it’s a disaster waiting to happen.

1. Unfamiliarity with the tools required for daily work is, in itself, deeply problematic. It ought to be no more acceptable to playfully boast about one’s lack of ability to use basic computer functions than to boast about being unable to chair a meeting or manage people. It’s your job! ↩

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

On Search Engine recently, Ezra Klein told PJ Vogt:

This is an argument I’m always having inside my own industry. I’m someone who’s done a lot of hiring in my industry, so I think I have some credibility on it. People’s social media accounts are typically a reason why they don’t get hired, not a reason they do, in my experience.

The reason is that if they’re doing really well on social media, it’s for that exact reason they’re not doing as much of the actual work people are looking for.

I’ve been involved in many hiring decisions in my line of work, and I’ve never checked candidates’ social media profiles… but my job is in the Civil Service rather than the real world.

However, Klein’s observation chimed with my perspective on the experiences of a couple of colleagues during the pandemic. They both used different forms of social media to ‘subvert’ the usual routes of communication processes of their organisation. They both became ‘known’ (at least within the public health world) as ‘personalities’ who communicated clearly—that is, they were both successful at social media.

Both inevitably ended up in hot water for giving messages which did not align with the organisational view. The problem was that they had become ‘big name’ representatives of their organisation but were operating entirely outside the normal processes. Crucially, they both also mistook their success in the game of social media as expertise in public communications.

In other words, they did well on social media by demonstrating a dismissiveness towards due process and the expertise of others. The errors that undid them would not have been made by people with expertise (but may have been made by—say—me, in other circumstances). They became popular, but possibly undermined their professional reputations more than they bolstered them.

In contrast, while I know many people who have fostered professional connections via social media, I can’t think of anyone I know in my field for whom success on social media has led to genuine professional progression.

But, in fairness, this is only my personal experience: while it aligns with Klein’s view, others think differently. According to Alex Heath in Command Line,

Musk thinks X can build a viable LinkedIn competitor. “Historically, I’ve done a lot of recruiting on Twitter,” he said, adding that he sees someone’s posts on the platform as the “single biggest indicator for whether they are excellent or someone you would want to hire.”

Your mileage, as they say, may vary.

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

Google’s parent company, Alphabet, will publish its financial results for the third quarter today.

A couple of weeks ago, I was on a course for work about something utterly unrelated to technology.¹ During one presentation, one of the tutors said something along the lines of:

The report of that public inquiry is really easy to find on G—

I mean, it’s easy to find if you search online. I was going to say ‘on Google’, but that feels a bit creepy nowadays, doesn’t it?

It’s the first time I’ve heard someone express that idea in real life, which felt like a significant moment to me. I’m not suggesting that a perception of creepiness is about to make a huge dent in Alphabet’s profits, but it feels like a far cry from the social cachet that a brand like Google once held. It doesn’t feel like that’s a positive place for a consumer-facing brand to find itself, and certainly not for the long term.

1. It was about commanding the response to major incidents, which is the sort of course which is both incredibly useful and a bit frightening. ↩

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

There’s been a lot written about the death of the small smartphone, including two Verge articles that resonated with me: this from Allison Johnson back in April and this from Sean Hollister yesterday.

A lot of the coverage of small phones is imbued with an underlying assumption that it is people with small hands or pockets that particularly lament them: it’s often presented as an example of how the views of women are under-considered in the tech world.

I’m neither a woman nor a person with especially small hands¹ or pockets, but I am the owner of an iPhone 12 Mini. I mourn the passing of the iPhone Mini series. If there were an iPhone 14 or 15 Mini, there’s a good chance that I’d have bought one. Instead, all I’ve done is replaced the 12 Mini’s battery: far better for the environment, but less positive for Apple’s profits.

I prefer a phone with a smaller screen because it feels ‘handier’ than a larger phone and because it feels less immersive. It feels like using a tool, not like being sucked into a portal to a world of internet nonsense.

1. While writing this, I realise that I’ve finally forgotten my glove size. In surgery, sterile gloves are sized by number, much like shoes, and knowing which size gloves to pull off the shelf was an everyday essential. I reckon it’s fifteen years since I was last in an operating theatre, so perhaps it’s no surprise that I can remember that detail any more. ↩