A First Quarter Update on My Tech and Creative Projects
Prompt Engineering, Fiction Writing, Editing and more
It has been a quiet few months for this newsletter. Top of mind have been way too many events in my personal life: house renovations and moving! contractors! Keeping up with the kids! our first dog!
Work has not been easy (the hustle of seeking consulting clients), either. The combo has kept me away from properly journaling my work at the intersection of publishing and AI.
That work, however, continues apace, at least as much as possible in the time spared by all the above priorities. As we are also documenting on the jointly authored newsletter Paolo & Praveen on Tech, that work is also producing some initial prototypes following the Storya app experiment. More on that in upcoming issues on both newsletters.
I believe it is also important to acknowledge that the mental health and creative difficulties post-shutdown of Storya, which I have written about extensively, have gotten only partially easier in recent months. Finding my creative and professional stride continues to be an exercise in patience and resilience.
Prompt Engineering (profession or fad?)
One way I have tried to leverage all the learnings from that experience has been through the consulting work I mentioned. I have been working on and pitching prompt engineering projects for clients in the publishing space. It has given me some interesting opportunities to test and refine my AI skills in the “real world”.
That being said, my overall feeling (at this admittedly early stage of prompt engineering as an actual profession) is that most companies approaching AI still struggle. There is still reticence to take a deep dive into how the applications of generative AI can help businesses. There is a combination of legal and employee concerns about AI replacement and compliance that is preventing a lot of experimentation going on in official capacities.
I say “official” because, as many surveys and studies are finding, employees and managers are already secretly using chatbots and AI tools across their tasks for professional purposes, often without notifying their bosses in light of the concerns above.
That being said, I thought I’d share some key prompt engineering learnings, something we have touched on previously without going into sufficient details. So, what are these learnings?
Use the best models only: see above for recommendations, and stick to the key providers without getting scammed by the many services providing what are essentially prettier interfaces for less capable models.
ALWAYS break down the tasks by asking the model to PLAN its response by “THINKING STEP BY STEP”.
Anytime it is possible, provide SAMPLES: whether it is your own work or a work you admire, help the AI focus on the task at hand by giving it two or three examples of good work you want to inspire yours.
Roleplay: whenever appropriate to the task, ask the AI to “Act as ABC” role. This is another effective way to get the AI to focus on the task at hand. Even better when this is combined with the samples just mentioned.
These techniques are talked about with fancy names (“few-shot”, “chain-of-thought”, “personas”) but really the summary above gives enough context to start getting better results. To improve upon these simple frameworks, the magic of prompt engineering is really just iteration, which allows one to build longer “master” prompts for specific case studies. If you are keen to see some examples, of course, do let me know in the comments and we can cover those in future newsletters!
Looking ahead, I think that, while prompt engineering remains an interesting new opportunity for consulting work, it may struggle to take off under this exact guise. Perhaps new service providers will emerge packaging the idea of prompt engineering as something different and more acceptable for companies? Time will tell.
Fiction Writing (and lots and lots of editing)
The second major focus of my work has been the editing of my second novel, Path of the Nemesis. As some of you know, this is the more mature sequel to my debut novel, Path of the Guardian, part of the planned epic fantasy trilogy called Portal Wars.
Now, I have been a professional editor for financial publications covering complex capital markets topics across long-form and news articles for a decade. Even so, the challenge of effectively writing and editing fiction remains the highest (virtual) mountain I find myself climbing whenever I am sitting at a keyboard.
First, some context and numbers. It took me ten years to go from starting to draft my first novel in Italian (at the tender age of 16) to seeing it published in 2011 by a small Italian publisher. I started not understanding what I was doing as either a writer, a fictioneer, or a self-editor. Writing that novel was my creative writing teacher, and it was a very harsh one! I am documenting some of the nerdy journey that led me to the decision to write a novel at a young age on an author X account I recently set up, so feel free to check that out. But the gist of it is that there was very little guidance I had access to in my hometown, and the Internet was not nearly as powerful in the early 2000s as it is today. I mean, Skype and MySpace were the pinnacle of the Internet experience, enough said!
After moving to Asia, also in 2011, and beginning my career as an English journalist and editor in Hong Kong, I shifted my authorial efforts to the English publishing markets, which continue to be far more promising than those of Italy (sadly, and especially for speculative fiction). That transition is, of course, a whole complex topic in and of itself, which I am happy to touch on in future newsletters (let me know if you’d like to hear that part of the story!).
Getting from that published though immature first Italian novel to its English edition turned out to be a complete rewrite that took me nearly another decade. I finished and self-published that book, with the title Path of the Guardian in 2020). It was a slow, painful surgery of an edit, cutting deep into the very bones of that first book project to lay more solid foundations for the sequels.
So, what about these sequels?
Well, I have since written and self-published the second novel in that trilogy in Italian (also in 2020, “thanks” to the pandemic). If that was not enough on my plate, I drafted the entire third and final novel of the trilogy in Italian, also in 2020.
I remember it like a fever dream. For two, long pandemic months ( April and May) I woke at 3 am every day, while the kids slept, and wrote till my fingers hurt and my eyes got blurry. The result still leaves me stunned to this day: a 160,000 word monster of a draft and, at least in my very partial and biased eyes, the best creative writing I have ever produced. I have not touched it since and plans for how I will edit such a beast remain partially formed, but more on that in a moment!
After the pandemic, and after relocating to Singapore (also in 2020!), began the slow work of rewriting my second novel in English. It has been a slow but more methodical process than I previously experienced for many reasons. Technology, which this newsletter is at least 50% about, has played a massive role in helping me there. So let’s touch on that.
Translating a Novel
Since 2021, before I embarked on my Storya startup journey, I was already scouring academic research and available translation tools (starting from good ol’ Google Translate). Fast forward to the present, and AI models like Anthropic’s Opus and Google’s Gemini Pro have become incredible at handling fiction translation, even for longer texts. I can really see the nuances of my own writing in the translations they produce, and I work with detailed prompts to preserve my writing style as much as possible on a line-by-line basis.
Of course, this would be very hard (or prohibitively expensive) to achieve with a human literary translator. I typically start by getting the two top AI models mentioned to provide each their own translations of an individual chapter, then make them assess each other until I find a winner. That text becomes the rough marble to carve through the rewriting and editing process, still keeping the original Italian side-by-side.
So far, I have found that this process has increased my motivation and improved the quality of the output, if not necessarily the raw speed of the work.
Rewriting and Editing
Here, things are somewhat messier in terms of whether I see some substantial improvements to my work.
Some tools have been around for years, such as Grammarly, ProWritingAid, AutoCrit. These are all fantastic assistants for any fiction and non-fiction writer, and they are themselves adding new AI-based capabilities to go beyond line editing and offer more “developmental style editing” to the writer. I typically find their AIs underwhelming. It is clear their in-house models cannot compete with state-of-the-art models, at least for now. In some cases, I suspect all they are doing in the backend is use an API to GPT3.5 with some custom prompt to spit out a response. Unfortunately, to put it plainly, GPT3.5 is just terrible at most complex tasks, including editing or reviewing fiction writing.
That means editing remains a challenging, slow slog with many steps, even with technology involved. For me, it starts with a manual review, followed by a deep dive with models like Claude Opus for a more structural/developmental review using custom prompts. Opus at the moment ranks highest for its ability to work with fiction writing (GPT4 was the choice model until Opus came out). This is the part I enjoy the most, as it allows me to go off incredible, sometimes absurd tangents for character development, world-building and more. These chats often spawn ideas that I end up jotting down for use in other parts of the novel.
After that, as a last step, I move on to using the above mentioned line editing softwares and call it a day.
Do these tools really help?
Despite the challenges, my experience over the past year (basically since the AI tools have gotten any good), would lead me to say YES confidently.
My efficiency as a self-editor has increased considerably as I feel I always have a helper at hand to bounce idea and offer suggestions and alternatives whenever I am struggling. This is important when dealing with the long and drawn-out process of editing a 120,000 word novel!
Are the AIs going to replace human editors?
Not even close, as I find constantly needing to be harshly critical of every word that comes out of the AI when offering editing suggestions. So far, they tend to be too way generous and afraid to hurt the human’s feelings to be comparable to how an editor approaches a poor writer (I speak from experience having been on both sides of the fence as both an editor and a reporter!). But it is a helpful support and I am glad to have it.
The Master Publishing Plan
So, where am I? As of today, I am working on a final self-edit of chapter 25 out of 35 chapters in my second novel, which will be titled Path of the Nemesis. With external editing, I am hoping to have it ready to either self-publish or traditionally publish before the end of the year.
So, as that writing project ends, I will soon need to answer the question of how on Earth am I going to get through that monster, the unfinished draft of my third novel I introduced above.
So far, my thinking is that I will not be even attempting to edit it in the original Italian, instead using it as a base to help me write the English version directly. This will probably shave a couple of years off the timeline? That is my hope, at least.
I have invested very little energy in marketing my books in Italian (and only slightly more in English, at least in the last few years). But as I get close to publishing my second novel, marketing and sales will be something I need to take more seriously and, given all other duties and responsibilities in my life, the compromise will be that the Italian market will have to wait while I figure out my primary book marketing efforts. Perhaps that is also a topic worth touching on in future issues?
And that’s it for today. Thanks for reading this far!
I’d love to hear your thoughts on the processes I talked about, the tools I discussed, author careers, and anything in between, so please reach out in the comments or at paolo.danese@me.com! My inbox is always open.
Till next time,
Paolo
P.S. I plan another issue of this newsletter soon with a very simple goal: sharing as many useful links to tools and resources from the past three months, so look out for that! It will be a quicker read but hopefully a useful one.