A year ago, I wrote about using Kagi as my search engine. I wrote:

Do you remember the feeling when you first tried Google instead of Alta Vista, Excite, or Yahoo? The pages were clean, clear, and amazingly fast to load. The results were uncannily accurate. It was just obviously better, by leaps and bounds. The alternatives felt like they belonged to a different era.

That’s the feeling I get with Kagi.

A year on, I’m still delighted with the experience. Using Kagi has become such a routine, ingrained default that I really don’t think about it much anymore. I certainly can’t remember the last time I had to re-run a query on a different search engine.

I use Kagi’s Orion browser on my personal devices and have Kagi set as my default search engine in Microsoft Edge at work.

Kagi features which I particularly enjoy include:

• The ability to ‘raise’ or ‘lower’ domains in search results (or even ‘block’ them). For example, Kagi up-rates anything on a gov.uk or nhs.uk domain for me, which usually helps the thing I’m looking for bubble up to the top of the results.
• The ability to use Kagi Assistant with ‘bangs’. I have this set up so that I can just type the letter ‘c’ followed by text to begin a chat with one of Kagi’s supported large language model AIs, and ‘r’ followed by text to use one of the large language models in ‘research’ mode.
• The ability to use ‘lenses’ to append custom search parameters to a query in a single click.
  

Most of all, though, I enjoy the fact that Kagi is ad-free: its results are ranked solely by their likelihood to be the correct answer to my query, not by how much anyone is bidding for a specific keyword.

Looking at my Kagi stats, I average around 750-1,000 searches per month. I’m clearly reliant on searching as part of my day-to-day personal and professional work. Perhaps if I searched less, I would balk at paying a subscription fee to search the web, but Kagi is such a huge quality-of-life improvement over any other search engine I have found, that I’m happy to pay for it.

I’ve also been a happy Fastmail customer for quite a number of years. It strikes me that the combination of these two customer-focused ad-free products, which are among my most-used web tools, results in a calm, sane, and humane web browsing experience. I could cope without them… but I’d really rather not.

Give them a go!

I’m not using any referral links here—this is just my personal reflection, and while your mileage may vary, I’m happy to share what works for me. Who knows, I might feel differently in five years (cough).

The picture at the top of the post is a cartoon of Kagi’s mascot ‘Doggo’—which is actually one of the few things I dislike about Kagi. It reminds me a bit of Lycos—who knew that was still going?

Twenty years ago, I posted a link to this classic new story by Lester Haines of The Register regarding a courier mix-up at the BBC. This morning, it’s made me chuckle all over again: it’s the audio file which really makes it. There aren’t many posts that old where the links still work.

Sadly, Lester Haines died in 2016. The Register has a really lovely obituary for a man who was clearly much-loved by a great many people.

Amelia Tait has a great article in the July/August issue of Wired—not online as far as I can tell. She profiles some people who worked for 118 118, a UK phone service which once offered to ’answer any question’, and AQA 63336, a UK SMS service with essentially the same aim.

As a student, I did part-time work with AQA 63336 and very much enjoyed it, so this brought the memories flooding back. Working for the service meant logging onto an online system which would show all the questions flowing into the service. Click on one, and you’d get a feed of previous answers to similar questions which could be sent or tweaked. Or—more fun to my mind—one might have to write an entirely original answer.

This was made much more fun by the brand’s ‘voice’—they wanted workers to write in a wry style, and most importantly, to always have an opinion. If someone asked a yes/no question, they had to receive a yes/no answer, preferably with a dash of humour. And, of course, it had to fit in the length of a text message.

As I remember, there was a set payment per question answered, though the payment varied according to demand so that more workers would log on when the service was busy. As someone who liked the challenge of taking a few minutes to write new witty answers rather than picking pre-written items from a list, this served me quite poorly… but I nevertheless loved the job.

A few months after Steve Jobs announced the iPhone, AQA 63336 led an April Fool’s Day prank announcing their own phone. I’ve never quite worked out whether this was gallows humour from a company who understood that smartphones would kill their business, or just a simple attempt at promotion.

Wired isn’t where I typically turn to reminisce, but Tait’s article provided exactly that.

In 2015, along with over 1,000 other people from over 100 countries, I attended an international conference of health professionals in Geneva. One delegate gave a fascinating presentation on their country’s response to the 2014 outbreak of Ebola in West Africa.

Early in her presentation, she spoke self-deprecatingly about how her country previously hadn’t needed to develop public health guidance for high consequence infectious diseases. Instead, her agency simply looked at the guidance on the USA’s Centers for Disease Control (CDC) website, looked at the guidance on the UK’s Health Protection Agency website, and split the difference.

“There was one problem with that approach this time,” she said.

After a dramatic pause, she flashed up a screenshot of the gov.uk homepage, quipping “Good luck finding anything useful on there!”

It was a joke that brought the house down. The UK was a laughingstock; gov.uk was the punchline.

My friend James O’Malley wrote an excellent Substack post a few days ago, in which he called gov.uk “genius” and “the best thing the Tories did in office”. As always, he makes brilliant points. He knows more about technology and the web than I could ever hope to. His post is much better informed than mine and is definitely worth more of your time. Yet, my perspective differs a little from his.

I agree with many of James’s arguments: renewing a passport is miraculously easy; the site’s layout is brilliantly clear; the design principles and style guide are first-rate (though too rarely followed in practice).

But there are problems—or, if we want to be optimistic, opportunities for future development. These were evident when the service was being designed, but making passport renewal easier and putting bunting on the bank holiday page were, it seems, greater strategic priorities.

The first issue was the decision to sweep away many professional-facing websites and to collect everything on one giant gov.uk system. James links to some other people who have issues with this, but it has been a particular concern in the health field.

One major problem is that specialist-facing content is very difficult to find. I don’t see the advantage to a primarily public-facing website hosting, say, the request form that microbiology labs must use to refer gastrointestinal bacterial isolates to the national reference lab. What member of the public needs to see the form to request intrathecal antibody testing for Hepatitis C?

I know from frequent complaints that people often find it difficult to find that sort of thing on the site, even all these years after it launched. The problem is compounded where animal and human health cross: hosting content about testing birds and testing humans for influenza on a single platform is a recipe for confusion. This wasn’t true when the content was hosted on a previous website targeted quite narrowly at health professionals.

The mixing of public- and professional-facing information and the resulting unnavigability of the site became a particular problem during the COVID-19 pandemic, as already cited in evidence given to the COVID-19 enquiry. It is quite likely that professionals were, at times, taking the wrong action because they couldn’t find the right guidance among the morass of COVID-19 content that was being published daily. That’s a pretty major failing.

There are also two other significant effects of this approach.

One is that government agencies increasingly publish resources outside of gov.uk, usually partnering with non-governmental bodies to host content that would not be allowed—or be easily added to—gov.uk. I’m not going to point to specific examples for obvious reasons—if you know, you know.

Another is that a lot of content is simply not published any more, as it was deemed inappropriate to sit on a public-facing website—much of this sort of thing is no longer published online, and is therefore inaccessible to professionals and the public alike. The decision to sweep away professional-facing websites has had a chilling effect on government transparency—even while people celebrate gov.uk as a transparency success.

In 2010, when the gov.uk strategy was being developed, it was entirely foreseeable—and foreseen—that the spread of unreliable information would be an ever greater challenge.

Perhaps the single most baffling decision for gov.uk has been the choice to seemingly ignore that. This presents itself in three ways.

First: When government agencies need to give emergency messages—“there’s a big fire so close your doors and windows”, for example—these are routinely distributed via social media. If there were ever messages you’d want to present first on a stable, reliable, verified site and then syndicate elsewhere, those about public safety would seem to be at the top of the list. Instead, presumably because the gov.uk strategy didn’t account for this “user need”, the government has de facto outsourced low-level emergency messaging to Twitter. This sort of thing used to be routinely published, but doesn’t appear on gov.uk—we don’t give the public the opportunity to check facts they’ve seen on social media.

To take a really minor example: instead of this quote being published on a government domain, as it would have been before the launch of gov.uk, it’s now on a Twitter account that could have been set up by anyone. Without gov.uk, it’s likely that the quote would have been on a verifiable departmental website and linked from Twitter.

Second: Accuracy doesn’t seem to be prioritised. In the early days of the COVID-19 pandemic, there was a page about public transport on gov.uk that gave blatantly incorrect advice, suggesting that COVID-19 persisted longer on surfaces than human hands. This resulted in a deluge of questions for hospital infection specialists—and despite a chorus of complaints, the page wasn’t corrected for—literally—weeks. There is a guidance document on another topic live on gov.uk now with a persistent error in it—years after the error was first reported. If the information on gov.uk can’t be trusted to be reliable, then it undermines the whole endeavour.

Third: gov.uk doesn’t follow its own advice. In his Substack post, James celebrates the success of gov.uk Notify—a service which is routinely used to send unsolicited messages to members of the public exhorting them to click embedded links. This is not a shining example of thought-through technological prowess—it’s a tool that trains people to fall victim to scams. And, just as James says, this vision of insecurity is actively pushed across all government departments—because a failure to think in a cyber-secure way isn’t limited to a single area of government but is automatically propagated to all.

In his post, James says

To give you a sense of where it was starting from back when GDS was created, the government didn’t know how many websites it had.

In that narrow sense, I think the situation is now worse: I don’t think the government knows how many sites it ‘officially’ posts on. We’ve gone from a situation where official government content is spread over an uncountable number of accountable government websites, to one where it’s spread across countless websites which aren’t even in the government’s control. Many of them aren’t even based in the UK, and there’s no mechanism for professionals or the public to verify that the content is genuine.

Some professional content on the ‘official’ website is so difficult to find that it is likely to have caused people to come to physical harm. Factual errors in content persist for years, meaning that it’s difficult to trust the content that is on there. And some things are just not published any more, making government less transparent than it used to be.

The bits of gov.uk that work are, as James says, brilliant. But some of the strategic and cultural change that has happened around the introduction of gov.uk, partly driven by the limitations of the system, has been damaging in ways that are under-recognised. If you ask me—and there’s no reason why you should—it all needs a bit of a rethink.

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

Last year, I reflected a little on Civil Servants being encouraged to bring their ‘whole self’ to work. I wouldn’t have guessed that the very next year, under the same political leadership, the same workforce would be told to ensure that ‘your beliefs remain at the front door’ (because heaven forfend that someone should wear a non-standard lanyard).

Yesterday, I was interested to read Zoë Schiffer’s piece on Platformer about a similar—but altogether more thoughtful—change in the culture of technology firms.

It’s intriguing to watch the pendulum swing.

I recently read a corporate document that promised the creation of ‘a user-friendly front door.’

I’m part of the intended audience, but I can’t explain what the sentence was trying to communicate. I don’t know whether the ‘front door’ is a website, a phone line, a physical location, a team of people, some combination of the above, or something different altogether. It is, apparently, to be an automatic door: it will be ‘using automation to make processes more efficient’.

I enjoyed the delicious irony of the authors failing to communicate while, at the same time, promising to be ‘user-friendly.’ I enjoyed the mixed metaphor of ‘automation’ making ‘a door’ ‘more efficient’. And, most of all, I enjoyed the fundamental absurdity of a ‘user-friendly’ ‘door’.

I was reminded—as I often am—of this from Jeanette Winterson’s 12 Bytes:

When institutional content tries to be more user-friendly, we get marketing-speak clichés like: stakeholders, bad actors, road maps, blue-sky thinking, low-hanging fruit, facilitators, roll-out … Conferences are the worst. I have been to some of them. By the afternoon I am sweating under the mental pressure of translating non-language. We need writers involved – and we need language that speaks to people. This isn’t about dumbing down, it’s about doing what writers do well – finding a clear, precise, everyday language that goes beyond utility, without jargon, with beauty.

The image at the top of this post was generated by DALL·E 3. I note with wry amusement that the AI conception of a ‘user-friendly front door’ has a knob on the left and a handle on the right, making it entirely unclear as to which way the door opens.

Apple recently released an advert for a new iPad, and it seems it’s like a Rorschach test for our times.

The first I heard of this advert was when I saw this article in the FT, reporting that Apple had apologised for it. And so I sought out the advert. The message I got from it? Apple has managed to fit a load of different tools and functions into an extremely thin device. I wasn’t offended by it, but thought I could see why others would be: wanton destruction of perfectly good instruments, tools etc. In a world of limited resources, and from a company that preaches about sustainability, it’s not a good look, even if it’s all just visual effects.

But it turns out that I was wrong. The controversy was related to a different metaphorical interpretation of the advert. As Tedium explained:

Apple’s infamous “Crush” ad deeply misunderstands the role of the hydraulic press in meme culture.

I’m completely ignorant of the role of hydraulic press in meme culture. It turns out that there’s a whole industry around videos showing hydraulic presses crushing things. I did not know this existed. I’ve heard of Will It Blend—but I’m clearly behind the times when it comes to online video culture.

The ad doesn’t connect because the message it’s trying to promote is essentially completely at odds with our understanding of the hydraulic press, which we only understand as a device that breaks things in the most brutal way possible. There’s no intelligence at all, artficial or otherwise. It just crushes things.

Clearly, many people had viscerally negative reactions. TechCrunch called the advert ‘disgusting’. Where I saw a neat metaphor for packing functions into a device, others saw an enforced digital transformation:

Does your child like music? They don’t need a harp; throw it in the dump. An iPad is good enough. Do they like to paint? Here, Apple Pencil, just as good as pens, watercolors, oils! Books? Don’t make us laugh! Destroy them. Paper is worthless. Use another screen. In fact, why not read in Apple Vision Pro, with even faker paper?

Our social context can completely change the way we interpret the same piece of footage… and perhaps I’m getting old.

Yesterday, I mentioned the unprecedented pace of social change and its effect on Westminster politics. Today, I’m thinking about something similar, albeit on a very small scale.

A decade ago, Facebook boldly announced a ‘shift to video’, predicting that most posts on Facebook’s ‘newsfeed’ would be videos within five years.

Much of the technology commentariat was unconvinced. They argued that people don’t like using headphones—and wouldn’t put up with subtitles to watch silently. Perhaps they were right: video didn’t take off in the way that Facebook expected, and lawsuits followed.

Fast-forward a decade, and today you’ll find me sitting in a coffee shop, with headphones in: not because I’m watching videos, but to block out the noise from those who are. In just a few years, it seems to have become socially acceptable for people to watch videos without headphones in public places. Shorts have become acceptable in restaurants when they’re the YouTube version; the Reels on the bus go round and round; the prevailing sound on trains has become TikTok rather than clackety-clack.

I don’t especially begrudge this: social norms are in constant flux, and there’s little point being curmudgeonly. But blimey, it’s been a quick change from being unviable to being unavoidable.

The image at the top of the post was generated by Playground 2.5.

Seventeen years ago, I highlighted a BBC News article in which Dave Gilbert reflected on thirteen years of online news.

So much of what has come to pass since is unwittingly signalled in that article.

The shrewdest comments came from the newspaper advertising executives who wondered where the revenue would come from. It’s a question some are still asking.

In 2024, this question is posed more frequently than in 2007. Waves of news organisations have collapsed, having transitioned to a business model that siphoned their income to Google and Facebook.

We clocked very early on that it was about a global audience, audience response and encouraging feedback – Web 2.0 in fact.

The idea of ‘audience response’ feels so quaint in 2024, when many news articles are built around embedded Twitter quotes. Often, it feels like the commentary has become the news, in a way that I certainly never foresaw.

People buy newspapers for a host of reasons but reporters never know how many read their own stories. From our real-time statistics, I know exactly what the audience is reading, and the feedback is almost instantaneous.

I don’t think any of us quite understood the extent to which those statistics would come to drive the news agenda, ‘public interest’ playing second fiddle to ‘things that interest the public,’ even at the BBC. ‘Beyoncé’s country album: the verdict’ would not have been one of BBC News’s top stories in 2007, though it was yesterday.

The pace of cultural change is difficult to believe.

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

Nineteen years ago this week, I blogged about a curious story on the now long-defunct silicon.com website site.

A burglar is today behind bars after picking the wrong house to burgle. His crime was caught in full by a webcam, which the hapless thief stole along with the computer, but not before it had sent pictures of him to a website.

A couple of decades on, it’s remarkable how twee this story seems. These days, there’s nothing remarkable about burglars being caught on webcams: Ring, Blink, Arlo and more have made it their entire business model. There are many millions of webcams set up for this exact purpose.

People often mention the pace of the change of technology, but the pace of change in how we use technology is equally astounding.

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