Category: Writing

A Writing Retreat

I spent the last week at a writing retreat, which felt indulgent but has been very useful. Firstly, the place I went was fully-catered, which meant I was able to focus on my writing throughout.

I’d expected to produce lots of new pieces. Instead I deleted tens of thousands of words of old notes. There are ideas for stories that I’ve been carrying around for years without ever finishing – particularly ideas for longer projects. I’d sometimes work on these without getting anywhere, so they were a distraction. I also took the time to finally clear the stacked documents on my Desktop.

By the last day, my head felt clearer, and I was able to focus better. I’m hopeful that the retreat will produce a permanent change – working with how I write rather than how I wished I wrote.

Too often, I come up with concepts I’d like to work on, but never manage to fill in any actual sentences. I also tend to work best quickly – trying lots of things and seeing which produces a viable story. That’s difficult, as you’re always starting from nothing, but it always eventually works out.

By the last day of the retreat I felt as if I was taking advantage of my new clarity. I wrote a couple of blog posts in a single sitting, and they as good as the ones I’ve tinkered with for weeks (some of which would get discarded as they went stale).

Spending a week on a writing retreat has taken up a lot of time, but I think it’s been good for me. The question is whether writing will be less frustrating in future.

The True Clown Stories Kickstarter is now live!

On Thursday morning, Dan from Peakrill press kicked off the True Clown Stories kickstarter. We’re looking for pledges of £900 to support the publication of a book of my clown stories. It also features work from Chris Parkinson, Louise Halvardsson and Michael Somerset Ward.

These clown stories are not straightforward tales about evil clowns. Rather, they’re about talented people who’ve devoted themselves to an art yet struggle to make ends meet. Sometimes this causes them to be angry, other times they despair. Some of the pieces date back to the noughties and were read at spoken word nights. Others have been written over the last few years.

This is a book I’ve been meaning to publish for an embarrassingly long time – we’re talking over a decade. Dan has pushed this project towards being a reality. We just need £500 more in pledges and it will happen.

Writers Notebook: Do you need to be an influencer to be a writer?

Jason Pargin is a successful writer who emerged from Cracked and has gone on to publish six novels. He also has a substack, and his recent post Celebrity Worship is Weird and Will Only Get Weirder discussed the strange situations that he’s in as an artist.

While Pargin has a big readership, it’s not quite large enough for bookstores and publishers to do his marketing for him. In order to let people know he has a new book out, he needs to keep an audience around ready for that announcement. Which means he has a newsletter, but that needs to be entertaining rather than just containing adverts for new books. He’s also had some success on TikTok, but that platform requires regular updates. A significant amount of Pargin’s time is spent maintaining an audience that he can market his books to.

As Pargin writes: “I literally cannot write novels as a full time job unless I turn myself into a multimedia influencer that posts daily to a large, loyal, highly-engaged audience.”

Writing a novel takes a substantial amount of work but, at the end of it, you then need to take it to readers. I’m guessing this was never an easy thing to do. But in the past you were dealing with human gatekeepers, but now the gatekeepers are algorithmic. You either need to build an audience, or be amplified by someone who already has one.

The obvious case of the latter is how a twitter review of This is How You Lose the Time War by someone called Bigolas Dickolas boosted the book to the top of the charts. It’s a lovely story. But Robin Sloan compared this to ‘the breath of the gods’, pointing out this is not a very efficient system. You either spend a substantial amount of time cultivating your own audience, or hope to be picked up by someone else’s.

You can complain about this state of affairs, but you still need to think about it. Writing a novel takes a large amount of time. Finding people have want to read a novel requires either hard work as an influencer or trusting to luck.

Mycelium Parish News 2023

I’ve just published a new zine, The Mycelium Parish News 2023, in association with Dan Sumption of Peakrill Press. The Mycelium Parish News is a collection of things that have happened in our particular corner of UK counter-culture over the last year and features a list of podcasts, books, websites, events and more. It’s available on Etsy for just £2.30 including postage (£5 for overseas).

What is the mycelium? It’s the word coined by Daisy Campbell for a loose network of creative people, across the UK and beyond, which emerged from a knot of discordians, KLF fans, Robert Anton Wilson afficionados, arts labs and magicians. We’ve also included lots of links that will be interesting to people we know. One inspiration for this is the counter-culture readers in the 90s, where you could pick one up and find half-a-dozen links that might change your life.

We print the zine as a hard copy as we think it’s important to have a physical artefact of the year, and we work to keep it as cheap as possible. Once we’ve distributed the print copies we will put the PDF on the Internet archive (you can currently read last year’s edition).

And, yes, we are already collecting things for the 2024 edition.

Links from my AI and Creative Writing workshop

Towards the end of June, I gave a small workshop about writing with AI. We looked at some techniques used to generate creative work with ChatGPT. During the session, I referred to a number of resources, which are collected below:

AI Book review: Death of An Author by Aidan Marchine

Death of An Author is an ebook novella produced by Stephen Marche using AI tools, which was commissioned by Pushkin Industries. I read a couple of good interviews about the book and, while I’d found the excerpts underwhelming as prose, I was curious to read the whole thing, particularly Marche’s afterword.

The book is about 95% AI text, using three tools. The original text was generated using ChatGPT, with Sudowrite used to tidy the text, making it “more active” or “more conversational”, and finally Cohere used to generate figurative language.

Death of An Author is a sly book, one that understands its place within a larger debate, and making allusions and interventions based upon this. There are references to the act of reading, the meaning of copyright and the book’s literary context – this is very much a book produced by someone with an English PhD, and the text itself is aware of what is at stake. The pseudonymous collaborative writer (Aidan Machine) even refers to an imaginary article by Stephen Marche, the human collaborator in the text. The writing itself is brisk if superficial, but the mystery had an interesting resolution – at least it felt so to me, as someone who rarely enjoys mysteries.

An example of the book’s commentary is in the description a dream the main character has:

That night, Gus had a terrible nightmare. He was taking an oral exam in front of his mother and ex-wife. Each time he tried to answer, a different writer’s voice came out: William Faulkner, Ernest Hemingway, Jane Austen.

Hemingway is the name of another AI writing tool. The dream questions the nature of imitation for human writers, as well as referring to one of ChatGPT’s finest tricks, that of imitating well-known writers.

The book also occasionally digs into specific descriptions, as in this outline of a meal.

That night, Gus made himself a meal of fried mushrooms in a cream sauce on toast. He started by heating a pan on the stove, adding butter to it. He then sliced the mushrooms and added them to the pan, cooking until they were browned. He added cream to the pan, stirring until the sauce thickened. He placed slices of toast on a plate and spooned the mushroom and cream sauce on top.

That passage is obviously AI, right? But I can’t be sure – maybe this is part of the 5% Marche wrote by hand. In a Guardian interview, M John Harrison spoke of an ambition he has: “I want to be the first human being to imitate ChatGPT perfectly. I bet you it’s already got mimickable traits”. Either way, it’s interesting that Marche chose to leave this prosaic description in the text.

Given the way in which the book’s plot takes it to such interesting places, it’s likely that Marche gave the AI some fairly clear leads about the overall story. It would be interesting to know the actual prompts used, although Marche talks about how he generated the book’s style:

What you need is to have it write something about a murder scene in the style of Chinese nature poetry, then make it active, then make it conversational, then Select All and put it in the style of Ernest Hemingway. That gets you something interesting. Raymond Chandler, after all, was not trying to write like Raymond Chandler.

The prose is often workman-like, but some flourishes and philosophical asides stick out. One sentence I particularly liked was “The policeman at the door of Gus’s office was a tall, thin, cadaverous man wearing a dove-gray suit that did not fit him well”. It is simple but has a lovely rhythm, and fits the hardboiled style. The choice of ‘dove-grey’ is interesting when we’d normally think of doves as white. And it makes me wonder how much work was done to get this just right. Marche has spoken about how the best results come from very precise prompts that are specific about substance and style. How much work was this, compared with just writing the sentence one wants?

Marche says that he did a significant amount of sifting of the produced text. This is some way from the dream of giving an AI a short prompt and having it produce perfect, entertaining prose. However, it’s notable that when Burroughs worked on his cut-ups, a great deal of time was spent exploring the results for interesting lines. Again, I’m curious about how much work was required to get the best lines (“He wondered why there was no good English word for slimy in a good way.”).

Marche talks about how good the AI is at ’heteroglossia’, with an uncanny ability to turn its hand to specific modes of speech such as “a paragraph from a Lacanian literary critic”. Marche says that he “[defies] any writer to improve on AI at that particular skill”. While the AI struggles to produce crisp narrative prose, it is an excellent mimic. In a podcast interview , Marche made the exciting suggestion that an AI might be very good at turning out a book like Dracula, which consists of different forms of documentation.

(A discussion in Wired Magazine placed Death of an Author alongside books including Moby Dick, which used extensive found text, and Graham Rawle’s Woman’s World, which sampled 50s women’s magazines)

Discussing the debate around AI, Marche writes:

So little of how we talk about AI actually comes from the experience of using it… Like the camera, the full consequences of this technology will be worked out over a great deal of time by a great number of talents responding to a great number of developments. But at the time of writing, almost all of the conversation surrounding generative AI is imaginary, not rooted in the use of the tool but in extrapolated visions.

It is important to actually play with these tools in good faith to see what they can do. The first book I bought made using ChatGPT was shardcore’s remix of John Higgs’ writing, The Future has Already Begun. Shardcore has spoken about how having the book as a physical object changes things. Reading Aidan Machine’s book makes a compelling argument better than the thousands of thinkpieces and opinion chatter. I only wish it had been made available as an actual paperback.

Marche concludes his afterword by saying that “What makes a good painting and what makes a good photograph are different. That transition required a complete reevaluation of the nature of visual creativity” The best AI art will not come from reproducing writing that humans can do better, but from finding new forms for this medium.

Is AI/ChatGPT as exciting as people say?

On this page, I’m collecting together some links about AI that I want to refer back to. Those are below; but first I wanted to discuss what I find interesting about AI right now.

I’ve been suspicious of the hype around AI for a long time, but I keep reading articles saying that AI models will revolutionise the economy, replacing millions of jobs. I’m even seeing serious articles saying that these new AIs are approaching sentience and are an existential threat to humanity.

One theory I’ve read is that the AI hype is so loud because the crypto grifters have moved onto this: there is always a profit to be made from hyping the next big thing. It was also an amazing marketing move from OpenAI to claim that their early models were too dangerous for people to have open access.

At the same time, very smart people I know, people I trust, are telling me there is something important here. And I’ve had several dreams about prompting AIs – so I should not be dismissing this too easily.

I’ve been playing a little with ChatGPT recently. Showing a friend how it could generate an email was an enlightening moment. He doesn’t like writing formal messages and having a detailed text produced from a prompt seemed revolutionary. My own experiments show that ChatGPT is good at certain types of output, but its grasp on facts is hazy. Asking it about things I know well, like hiking routes, it returns plausible information, but lacks telling details and gets significant facts wrong. This is beguiling, miraculous technology, but it (currently) has very clear limits.

Links on AI

The following are some good pieces that I have read on AI:

My current view

I find it hard to see how these huge statistical models are related to ‘true intelligence’, even as they raise questions by doing things that we once thought relied on intelligence. One notable thing is that (as with machine translation) these models are entirely reliant on human-produced work. This has led to the ethical questions around the model incorporating copyrighted works – and I note (via the Washington Post) that this blog is one of the sources for Google’s C4 data set:

I also wonder how much further these models can go. It won’t be long before the deluge of AI content begins to be absorbed into the models, which may undermine their effectiveness.

I also suspect there is a limit to the effectiveness of LLMs for problem solving. Matt Webb’s Braggoscope is the most compelling experiment I’ve seen, where ChatGPT was used to classify the thousands of In Our Time podcasts into the dewey decimal system. It’s a task where small inaccuracies will cause little harm, and Webb estimates that the automation of this was 108x faster than doing it manually.

But for tasks like programming, much of the art is not in producing the code, but figuring out what code needs to be written. It’s possible that a new paradigm of programming emerges from AI, but for any form of programming as we currently understand it, the trick is not writing the code but defining what code we want written, and making sure that we have achieved our aims.

The difficulty with AI is in producing very specific text. Producing remarkable sonnets about odd subjects is breath-taking, but getting an AI to write Allen Ginsberg’s Howl or Pierre Menard’s Selections from Quixote would be a different matter.

Book review: literally show me a healthy person

Darcie Wilder’s 2017 book, literally show me a healthy person is very much a twitter novel, consisting mostly of aphoristic sentences. It bears obvious comparisons to Patricia Lockwood’s No-one is Talking About This (2021), particularly since both juxtapose the ephemerality of twitter with the realities of grief.

Wilder’s book is both shorter and scrappier. It buries the story about trauma under glib, often funny phrases (one that particularly stood out: “saying ‘awesome’ on work calls is just another way to stay punk” – I feel slighted). The lines about nihilistic partying come to stand in reaction to the narrator’s descriptions of childhood trauma.

I think this aphoristic style works well, and captures one of the strangest feelings of the social media age – what the Content Mines podcast referred to as ‘structural dissonance’ – the way in which social media platforms blur together trivia, marketing messages and horrific news. One example that the Content Mines used was when the SweetMiniDollsHouse Instagram account interrupted its posts about dolls house miniatures to document the account owner’s pictures of the Ukraine invasion.

This is an extreme example, but social media is full of such examples. Pictures of people partying rub alongside political messages. When Wilder’s narrator fails to focus on their trauma (and any chance/attempt of healing) it reflects the way that we bury things in favour of surface entertainments.

The idea that trivia is a distraction from the world’s issues is a common criticism of capitalism but social media provides a constant distraction from our own lives (there are values to these tools, but they are easily swamped by commercial needs). The world we live in very much reflects the one described in Guy Debord’s Society of the Spectacle, which predicted the current media landscape from 1967.

David Shields’ 2010 book Reality Hunger predicted more texts like Wilder’s novel, with its aphorisms and blurring of fact with fiction (many of the narrator’s lines were originally posted on Wilder’s twitter account). literally show me a healthy person is a good example of the sort of novel Shields was talking about.

Looking for something in my blog this week, I found a link from 2010 to Russell Davies’ review of a Douglas Couplan novel. Davies wrote:

I’m sure I’ll read more of Mr Coupland’s books but I’d almost rather read his lists or his notes. It seems like he’s the perfect novelist to write the something-that’s-not-a-novel that must be just around the corner.

I’ve been waiting for this “something-that’s-not-a-novel” promised by Reality Hunger for a long time now. Wilder’s novel is a good example of the form.

Why ARGs never worked

Back in November 2022 I gave a talk to students at Leeds Trinity University about transmedia storytelling, specifically around alternate reality games and electronic folklore. Researching the talk allowed me to explore my long-held fascination with ARGs.

The first ARG I heard about was The Beast, which was beautifully dissected by 50 Years of Text Adventures. The ARG initially looked like a new art form, an Internet-native way to tell stories: “the team began to see their scheme as a Gesamtkunstwerk, an all-encompassing art form for the new century, just as opera and cinema had been for earlier eras“. The form generated fanatical audiences – one puzzle from Perplex City was solved recently after 15 years of effort.

The most exciting thing for me was how ARGs told stories in the world. They merged reality and fiction, promising to make normal life a little more magical. I loved the idea of hosting stories within the world, of walking through them; how technology allowed ARGs to reach out to people through email or even phone calls.

But ARGs never took off, and I personally never got excited enough to chase them. One issue was the actual mechanics. I liked the idea of seeing symbols in the real world that led to an ongoing story, but the actual ARGs did not have the magic I longed for. I’m not a huge fan of puzzles so that is perhaps a part of it – I like the joy of solving something, but pitching the difficulty is tricky – too easy and it’s trivial, too hard and it’s frustrating.

This led to one of the problems with ARGs, that some of them had puzzles so hard that they could only be solved once. In a 2008 talk about ARGs, Everything you know about ARGs is WRONG, Dan Hon talked about the need for ‘casual’ ARGs with “short, snappy, fun gameplay”. The sort of thing that did not require people pooling obscure knowledge. As Hon said, “What the fuck kind of “normal person”… wants to de-steg a jpeg, or write a distributed brute force attack against military grade encryption? Jesus. Whatever happened to the kind of people who like playing Singstar? Or, you know, Snake”. Puzzles based on lute tablature are great, but they are not going to bring in many people. Without active communities, these games are basically unplayable.

The other problem with the games was scaling up to massive audiences. Some of the ARGs made use of personalised e-mails to players. The Dark Knight promotion famously baked cakes containing clues. Quoting Hon again, “It’s fine sending a cake to twenty people. You just try sending a cake to twenty thousand people though. Or twenty million”. This is “by intention an experience that can never scale to become a mass-market event“. Admittedly, the cake stunt brought a lot of vicarious delight, but then we’re back to the games relying on active community.

The games also quickly developed cliches, which were being clued out as past their sell-by date in 2008. Hon was harsh on codebreaking, secret societies, intrusive puzzles and treasure hunts (“I’ll smack the next person who proposes a treasure hunt and thinks it will make people play”). Another problem was hopping between platforms just for the sake of it.

A major point of contention was the boundary between the games and real life. I was fascinated by the idea of blurring these boundaries, but many ARG designers became obsessed with the idea of pretending the games were entirely real, and hiding all the seams – a technique known as TINAG, ‘this is not a game’. Dan Hon pointed out in a talk that ‘Alternate Reality’ is just a roundabout way of saying ‘fictional’. “No more “alternate reality” bullshit. We can use the word “fiction” or “story” instead, so normal people can understand us.” Later in the same talk, he said that ”ARGs are just things that you can play and that tell stories. ARG doesn’t mean anything.” Expending so much energy into pretending these things are entirely real is a waste of time.

I think there are many lessons that can be learned from ARGs. As Andrea Phillips said in an article on Atlas Obscura, “A lot of [the ARG] energy has gone into lightly interactive web series, room escape games, narratives-in-a-box. Things that use a few of the ARG tools (tangible artifacts, in-story websites, email) but don’t use the full-fledged ARG formula.”

The collaboration and investigation aspects of the games seemed to work better for explainers/investigations of TV shows like Lost where people would blog and discuss clues in the episodes – while most viewers ignored the actual ARGs that had been built around the game. People love looking for clues in serialised drama, and the ‘replayability’ of TV shows is much higher than for ARGs.

Writing and ChatGPT

As far back as 2008, Kenneth Goldsmith was saying that, through the Internet, writing had ‘met its photography’ – referring to the supposed crisis painting faced once realistic images could easily be produced. ChatGPT is another part of this long-running crisis, rather than something new.

Like most people, I find the output of tools like Midjourney, Stable Diffusion and ChatGPT miraculous. Being able to put a few words into a system and receive a picture matching that description is incredible. I keep reading claims that ChatGPT can pass the bar exam, or can think at the age of a small child, or generate computer code.

While ChatGPT can produce very good undergraduate essays on certain themes, it is not able to generate spontaneous writing about obscure texts. And while it might be able to create specific examples of code, that is not the main problem in programming. (Describing what a programme should do, and seeing whether it works are far more time-consuming). These tools are remarkable but they cannot easily synthesise new things.

I’ve had a lot of debate with a friend about whether these tools are creative. They definitely do some tasks that would be described as genuinely creative. However, this is a brute-force approach to only one type of ‘creativity’. These models are huge statistical analyses of existing content, a huge multi-dimensional data table. They are not artificially intelligent in the way we normally understand that term, rather they are reliant on a huge pool of imported data.

ChatGPT is very good at is producing styles of writing seen on the Internet. It can automatically generate the sort of text that provoke reactions, but it has trouble producing sustained and detailed texts. This tool will be able to flood the Internet with the sort of writing that already appears on the Internet. It is wickedly good at listicles, short blog posts that seem to say something, and arguments about major franchises.

This sort of language was already being crafted for the Internet. People were writing web copy to fit in with SEO. Buzzfeed was producing headlines that would be popular, and then crafting the stories to fit them. Twitter was promoting a particular style of discourse. The algorithmic ranking of text was a problem long before, because it was shaping the sorts of writing being produced.

ChatGPT arrives at an significant time. More people are reading than ever before, but they have changed what they are reading, moving on from novels and newspapers to smaller pieces of text. This is an fascinating time to be writing stories. ChatGPT is going to make certain types of content worthless (it’s a bad time to be producing small blog posts to increase engagement). It’s time to leave basic writing to the machines and move on to more interesting things.