My AI policy

When going over the paperwork for a new client recently, I saw something I hadn’t seen before. I had to sign a document guaranteeing that I wouldn’t use AI to write anything for the site.

For the avoidance of doubt, I had no problem at all with this clause. I just found it depressing that this is something they needed to put in writing, rather than it being blindingly obvious that this would be a serious breach of trust. 

Oddly for somebody who mostly writes about technology for a living, I’m a bit of an AI Luddite. I don’t really like the way people are queuing up to outsource tasks that they could do themselves with a little effort. It feels intellectually lazy, and I don't care if that gives me "old man yells at cloud" energy. 

As a writer, there’s simply no way I’d get AI to write anything to publish in my name. That’s not just because it feels like ushering in the end of a career that's been pretty darned good to me, but because it’s just not that good. AI hallucinations are a thing prompting much  misinformation, with no real safeguards in place to mitigate the damage, but beyond that, there’s no personality or wit. I wouldn’t want my name on that.   

On top of that, there’s the huge environmental impact, which doesn’t feel worth the cost for frivolous activities, like creating an action figure of yourself or whatever the social media-driven trend of the month is (yes, again, I'm yelling at clouds). Frankly, I’d be very happy if AI use was legally limited to things where it could be transformational for humankind: medical research or coming up with climate-saving ideas that could help dig us out of the terrifying hole we find ourselves in.

But as it’s not, I can see a number of limited uses for AI in journalism, which I recently outlined in a piece for the Media Diversity Institute. Efficiently digging through datasets for research, automated transcription, coming up with headline alternatives, or sourcing ideas are all reasonable uses. Automated article writing or editing is not.

In other words, I’m not quite an AI vegan, but I’m certainly an AI flexatarian. If you commission me, you can rest assured that all the words will be written by me. I’m unlikely to use AI in any capacity, but if I do, it’ll be to augment my own work rather than to outsource it to a soulless chat bot.

What Does The Future Hold For the Roller Coaster?

This was originally published on Gizmodo UK back in 2018. I was reminded of its existence after getting in touch with Brendon again for comment on a news piece I did on the Evening Standard. As the original feature is lost to the internet since the Gizmodo UK closed in 2020, I figured there was no harm in republishing it here for posterity.

If you’ve ever witnessed innocent digital thrill seekers thrown off a poorly-conceived ride on RollerCoaster Tycoon, you’ll know that the design of thrill rides is best left to the professionals. Brendan Walker is one such professional, and at New Scientist Live this weekend, he revealed some of the physics, mathematics, psychology and biology you need to have a grip on to ensure people are left scared, but not scared.

Let’s take the loop-the-loop, for example. The first instance of this, from 1901 in Coney Island, was just that: a single loop-the-loop. And it worked well… if you didn’t mind a chance of whiplash and head trauma. That’s always a risk if you face 10Gs of force in an instant.

Over a century on, and Walker can speak fluently about what you need to achieve this feat not just with vehicles, but with people on a slide – as a TV crew once asked him to try. “Firstly we need potential energy - height at the top of the slide,” he explains. “As we go down the slide, that's converted into kinetic energy which gives us velocity and the speed we need. As we go into the loop we need centripetal force, which sticks you to corners. The centripetal force still needs to be large enough when you get to the top that the effects of gravity don't overweigh it and you drop out.”

The problem with that? You can’t throw people down a slide at 10Gs for obvious safety reasons. Ride designers get around this with something called the clothoid loop: two circles of different diameter. You hit a looser loop first with a low centripetal force, minimising g-force levels and, as velocity drops, you can safely go into a second tighter loop. That’s why you no longer have nurses station at the end of rides to treat for head trauma. Despite the warning precedents, Walker managed a single-loop vehicle-free loop-the-loop, but it never made it onto our screens as it was still viewed as too dangerous.

Thrill factor

But there’s more to making an engaging thrill ride than just physics and avoiding injury, and much of it is down to psychology. Working off questionnaires, interviews and biological responses including neurological monitoring, heart-rate variability, skin sweat and pupil dilation, Walker has his own equation for this which squares arousal against pleasure:

“Our levels of arousal track quite closely the changing forces on a roller coaster, but that is only half the picture because thrill also relies on the element of pleasure which is a far more subjective part of human experience, and that's where the artistry comes into it,” Walker explains to me on the phone a few weeks before his talk.

The main thing, you see, is that optimal thrill factor comes from the contrasts inherent in ride design: you need the lows to appreciate the highs, in other words. “We experience thrill when there's a rapid and large increase of pleasure and arousal together,” he explains. “You get this rush of dopamine and adrenaline - that's starting from a base point. If you can pull them lower and make them feel fearful, create displeasure, then not only will the release from this displeasure cause thrill, but if you can give them something that's exciting, fun and cathartic and will make them laugh afterwards, then the move in that thrilling trajectory is even greater.”

To that end, there’s a theoretical maximum that the thrill level a roller coaster can physically achieve. Against this scale theme park rides routinely achieve a score of around 20% on the Walker Thrill-Factor scale, something that theme park management often find disappointing. Walker thinks they’re beating themselves up over nothing: “When you think of thermal efficiency of combustion engines, they're around 20% too. So I think theme parks aren't doing too badly turning potential energy into emotional energy.”

Interestingly, much of the psychology kicks in long before you actually sit down in the chair. “Part of the success of any ride relies on marketing. The story that is told which can go on for days and weeks in advance from buying the ticket to strapping in,” Walker explains.

“You've probably got more chance of dying from falling off a donkey at Blackpool beach than you have of falling off a roller coaster, but people perceive it to be dangerous and that's part of the marketing.”

But the very peak of thrill, according to Walker’s own research, actually comes before the ride kicks in. “The moment the restraint bar is pulled down and locked in – that's the peak moment in arousal. After that, you only achieve 80% of that.

“There's something about the fear of the unknown and what might happen to you, compared to the actual reality of what does happen to you. That's the moment that really intrigues me – when you're locked in and there's no escape.”

A thrill without spills?

This is why the restraint bar is authentically replicated in Walker’s pet project: a VR roller coaster, designed and built by a team of engineers from the University of Middlesex. A chair on a moving platform is supported by air muscles that flex and contract in time to what’s happening on the VR headset, but as graduate assistant Denis Tsvetkov tells me after I’ve tried the thing, the bar is purely a nod to realism: “It's completely not a safety thing - the platform only tilts around 20 degrees, so there's no real danger of you falling out.”

That’s not how it feels at the time, as I am briefly taken away from hubbub of the show floor to an authentic steel roller coaster. The only thing missing is the wind in the face. “We've been consistently been getting that as feedback,” Tsvetkov tells me when I bring this up. “The handlebars come from the top, so we're considering building something that can attach to that.”

You may feel this virtual experience isn’t really in the spirit of roller coasters, but it’s not without precedent. In his talk, Walker points to a 19th-century ride called “The Haunted Swing.” Guests would enter a room and sit in a giant baby’s pram that would rock back and forth, getting faster and faster until they were actually upside down. They weren’t, of course: unbeknown to them, the room was also on an axis swinging in the opposite direction creating a powerful optical illusion. Appropriately, this ride attracted the same kind of scientific wonderment that Walker spreads now, getting this write up in the 1895 periodical Psychological Reviewcurious and interesting feature however, was, that even though the action was fully understood, as it was in my case, it was impossible to quench the sensations of 'goneness within'."

The text implies that many late 19th-century American gentleman subsequently lost their lunches, but in the 123 years since The Haunted Swing debuted the world has come a long way, and VR tricks the brain without literally needing the room to spin. The drawback – and it’s a pretty major one – is the cost: the VR roller coaster chair will cost you between £8,000 and £10,000 to build. And you can if you want: the guides are all open-source, in order to encourage schools and universities to work on their own.

Despite this, Walker believes that this kind of VR experience is the future of thrill. “I'm not just interested in VR because I think it can be bolted onto roller coasters, but because it can be distributed globally,” he explains. “The concept of a theme park doesn't need to be geographically located, it can be an idea, and experiences can be created and distributed in very different ways,” he adds imagining a world where a fantastical VR scene can be projected onto skydivers or bungee jumpers. He already has a VR experience that can augment the standard playground swing to make you feel as if you’re a jellyfish under the sea, or speeding through a cityscape. He doesn’t anticipate buy-in from the big theme parks though – not at first anyway. “They won't make the giant strides that are necessary at this stage."

The here and now

Still, they have his number: he’s designed and consulted on thrill-ride design for years, and actually has a story of saving Alton Towers thousands of pounds. The ride is Thirteen, and if you’re not familiar – SPOILERS – the ‘coaster doesn’t end when it seems to. Instead, the floor gives way and you fall into a pitch-black room, and proceed to accelerate backwards. “They asked me how far they needed to drop members of the public for them to be thrilled,” Walker explains. “Every foot of steelwork was going to cost them tens of thousands of pounds to produce so they needed to get it right, but they wouldn't let me build a test track.”

How do you get around that problem? As ever with Walker, the answer is by turning to science: he looked at studies examining the average human response time to unexpected stimulus. “The rule I set myself: as soon as the hands grip the side of the ride, your brain has had time to process the information and create a response... something like 500ms for the 95th percentile." That meant 1.2 seconds of freefall, or around 2.4 metres to recognise the danger, and seven metres to respond to it. In the end, he managed a workaround that only required five: there’s an initial 30cm jolt that doesn’t actually go anywhere, before the five metre drop kicks in – a cunning psychological trick to save thousands of pounds worth of steel.

But perhaps Walker’s main strength is not as a designer, but as a critic. So which rides thrill the man himself? “It's not exhaustive,” he five-metre laughs. “I keep thinking I should create a Michelin Guide to all the rides.” Ultimately, he names two: the first is the Saw ride at Thorpe Park. “It may not have provided the peak highest thrilling experience - like the singular moment at Oblivion at Alton Towers – but what it did do over the graduation of three minutes was, like a well-choreographed piece of music, it kept delivering a new thrilling experience,” he explains. And if you can’t get to Thorpe Park, you’ll find a family-friendly identikit version called Rage sat at Southend’s Adventure Island. Something Walker discovered when he couldn’t get the adult themes of the Saw ride onto a segment of Blue Peter.

The second is a bit of a wildcard: Spinball Whizzer at Alton Towers. “It's a coaster that doesn't invert, and because the car spins, you don't know which way you'll be facing,” Walker explains. “And because you're always looking at other riders, it doesn't fails to make people howl with laughter. It really bucks the trend when people think about what makes a successful thrill ride.”

Whether it’s screams or laughter, then, social bonding remains a key part of the thrill-ride experience. Walker’s vision for a VR-based thrill-seeking future may have the basic mechanics right, but until the medium becomes a truly social experience, family days out to Alton Towers should have a good few years left to dish out the screams.

Eight months' worth of weekly tech columns

Back in July, I was approached by an old friend to write a weekly tech column for his New Zealand startup site, theBit.

David was half of the duo who interviewed me for my job at Dennis Publishing which, aside from being the best job I've ever had, literally changed my life. That may sound like hyperbole, but it's true: there's simply no way I'd be sustainably freelance without the connections and portfolio that came from that opportunity.

So I was delighted to work with him again, especially as he actively encouraged my slightly unorthadox writing style in the early days of Alphr when the site was still finding its editorial feet (before having them stamped on by its future owners... but I somewhat bitterly digress).

While I'm still chipping in news stories for the site, for now I've written my last column, so I thought I'd catalogue them all here for posterity.

Coming up with an idea each week would always take more time than the actual writing, and I'd sometimes spend entire mornings digging around the internet for topical inspiration, worrying that I'd never come up with anything. Each time, I'd eventually finish, file, breathe a sigh of relief and then repeat the cycle all over again.

So here they are, all 28 columns, covering the full gamut of silly to serious and flippant to thoughtful. Some were syndicated to Stuff.co.nz, so to mix things up a bit I've included those versions (with suitably frothy comments sections) where available.

If you like what you see and want a columnist to write regular opinion pieces for your site - well, you know how to reach me.

When Yahoo almost saved the world from Facebook - Yahoo once wanted to buy Facebook, and it's hard to imagine it ending in anything other than failure, given the company's history of ruining good brands.

Is my cryptocurrency investment really just gambling under a more professional name? - I compare my modest wins in political betting founded on knowing my stuff with my (comparitive) runaway success in crypto based on pig ignorance.

Tragedy bring out the worst of Twitter (or: you say it best when you tweet nothing at all) - An angry piece about the partisan unkindness of Twitter when faced with tragedy.

I’ve never had a VR injury, but I do have one embarrassing tech tale - You want to hear about the time I ended up nearly naked in Tooting lake for my job, right? Of course you do.

Apple’s vision of wireless-charging convenience sounds awful - In which I moan about the enormous environmental cost of making things a tiny bit more convenient.

If you're over 30, it's safer not to use emoji - A piece on the ambiguity of emoji, and how people of a certain age should just stay out (myself very much included).

Is Twitch streaming a dream career? Not for me - A look at the utterly miserable lifestyle that always-on video game streamers seem to endure.

AR-gh: Why nobody should welcome back AR glasses - I learned about the dangers of AR headsets by interviewing, of all people, a magician. Here's a recap.

Smart glasses won’t be ubiquitous until they carry an Apple logo - On a related note, here's my theory that only Apple's unique ability to normalise stupid looking things (hello AirPods) can make smart glasses mainstream.

Samsung wants to make foldables mainstream, but can a folding phone beat my apathy? - I get the point of folding phones, but I can't imagine ever actually buying one.

The Joe Rogan controversy and how Spotify may regret moving beyond music - All about Spotify and what economists call the 'sunk cost fallacy'.

If you want to change Spotify policy, first sell 120 million albums - On Adele's very minor victory over Spotify and the strange relationship modern musicians have with their art in the age of streaming.

Sorry Xiaomi: 4K smartphones are still utterly pointless - On tech's occasional persistence of chasing 'magic beans' innovations for innovation's sake.

Ron’s Gone Wrong can’t decide whether big tech is a problem or solution - In which I overthink an entertaining family-friendly movie.

Facial recognition ad tech flunks the creepiness test, but someone will make it work - It's not the companies bragging about facial recognition tech you need to worry about - it's the ones that can afford to experiment on the downlow.

What’s with all the weird tech collaborations in 2021? - In which I roast some truly ugly (and expensive) tech special editions.

Microsoft’s $100b purchase of Activision looks like crazy money - Questioning the wisdom of spending 17 times more than Disney spent on Lucasfilm on the creators of Call of Duty.

Boston Dynamics’ viral appeal is undeniable, but where will the money come from? - On the weak long-term prospects of Boston Dynamics.

Companies: stop trying to normalise mid-game snacking, please - It's just a recipe for greasy control pads, right?

Facebook is too big to tame or kill - On Facebook's continuing scandals and the strange legislative place it finds itself in.

I’m starting to hate true wireless earbuds - This is probably one of those rare occasions where my headline actually negates the need for a summary.

Why Apple should care about silly superstitions - My theory that releasing an iPhone 13 was an unnessesary risk, given peoples' superstitions around the number. Doesn't look to have hurt Apple, as it turned out, but I wasn't to know that at the time!

The iPhone 13 may be a dull update, but boring will prove to be brilliant - Still on the iPhone 13: why anybody predicting failure based on a lack of innovation hasn't been paying attention to Apple's history.

5 ways Amazon Astro is better than a real butler (and 5 ways it’s far worse) - A listacle based on Amazon's mad little Wall-E wannabe.

Why we should all rely a little less on Googling - A reliance on Google gives us a false sense of our own knowledge, researchers say, and that's just the beginning of the problems of a search addiction.

Sorry, small-handed folk – it doesn’t look like mini phones are coming back - How the world fell out of love with small phones, and even Apple can't turn it around.

The Samsung Galaxy S21 FE is still coming, but is late really better than never? - How one phone's ridiculous delays killed its chances of making a splash. Sometimes 'never' is better than 'late'.

Samsung to throw everything at the wall with Galaxy Unpacked, but will anything stick? - Another slightly dated piece given 'yes, it did stick thankyouverymuch.'