It's been a couple of months since my last book and I'm well under way with Stellarforged now. I'm pretty sure it'll be my next book now, but not sure exactly when.
I'll keep the news section fairly brief, but there's some rambling nonsense below if you like to read my thoughts on some stuff.
The Heretic Spellblade Epilogue audio comes out on June 30th. Podium bumped the release up a week and the new date is on the Audible page now.
Stellarforged is still underway. I'm aiming to get Book 1 done by the end of August at the latest, so that it can hopefully get an earlier audiobook release (as it takes a lot longer to schedule recording than in the past). Chapters should continue coming out, but it'll be in fits and bursts. We're around 1/3rd of the way into the book, by my rough estimate.
New Stellarforged chapters hopefully next week. I'm working out some ideas for the arrival on Rylamis with my beta reader, as I think my original plan may have easily slipped into an "idiot ball" plot and violated my rule of not including stuff I personally hate seeing in other novels.
Early July, I'll be putting the entire Spellblade series on a 99c special as usual with the end of series anniversary (US and UK only, due to Amazon). I've waffled on whether to include the Epilogue, but it's already in the omnibus so is technically discounted, so it'll be part of it.
I don't have a lot of other direct news. I'm still avoiding talking too much about the future and what I want to do, simply because it tends to be used against me rather than to help people know where I'm going or what I'm thinking.
Anyway, there are a couple of rambly topics below.
AI Stuff
One of the most foundational myths about AI is that it's undetectable. Well, recent research shows that AI detectors have improved drastically, to the point one (Pangram) has an extremely low false positive/negative rate.
These come from a couple of papers I read about the AI spam on Amazon. One's an NBER paper on the economic impacts of AI books on Amazon (https://www.nber.org/papers/w34777) and the other is just a evaluation of detection methods (Artificial Writing and Automated Detection by Brian Jabarian and Alex Imas).
The first paper is an interesting read, although it's quite dry and the exact sort of thing I'd expect from a place like the NBER (I've dealt with similar Australian institutions and attended some conferences of the ilk in my past life). It's not a pro- or anti-AI paper, and while you could argue with some of the statistical methods around how they define some of the stuff they're researching, it falls into the category of "if you disagree with the paper, you might have a problem with its facts."
In short, over half of the books published on Amazon today are AI-generated (according to Pangram's detection). The researchers measure quality using reviews/ratings, which is probably the only way to do it, and are finding that the "quality" of AI books is increasing, which could also just mean that there's a growing audience for them. Although they do mention that most AI books really suck.
One finding they make without much elaboration (because it wasn't an ancillary part of their research) is that human authors active prior to ChatGPT and LLMs are unaffected in aggregate by the AI spam. I don't think they've made a more specific finding or argument about how or why, or whether authors who started post-LLM are able to rise above the tide of slop.
But overall, they're interesting papers. The fact AI can currently be detected is definitely an unpopular fact, as it's taken as truth that it can't ever be reliably detected. While I'm sure changes in AI models will mean Pangram will make more mistakes, it is an interesting change that I'm sure will be ignored in every argument about AI witchhunts. Especially as I'm pretty sure Pangram is commercial.
I'm not sure I'll comment much on AI for a while after this. AI remains in a weird, almost political place. There's noticeable pushback whenever anyone suggests not reading AI stuff, or being critical of it in anything more than a general manner.
The comparison I've made before is still valid to me. AI slop is like spam calls. People are going to naturally become sceptical, due to being repeatedly burned, and authors need to work around it. Telling people to just accept the AI slop and deal with it is like telling them to answer every call from an unknown number in case it's important. Nobody does that, and it's stupid to suggest it.
The Audible Buzz
If you've spent any time in harem communities this year, you've probably seen authors talk about how audiobooks are now riskier, make less money, are harder to produce etc. I've mostly seen it talked about fairly obliquely, but I'm certain authors have been blunter in places.
In short, Audible has forced all authors on to a new model that drastically reduces royalties earned by most audiobooks. Authors reliant on omnibuses are (likely) hit a lot harder than others due to the specific nature of the changes, which affects harem a great deal as many harem authors rely on omnibuses.
The changes will slash author audiobook earnings by 25%-50% depending on how badly they're hit. I think I lost around 30% now it's shaken out on Podium's end, and it's made me thankful I'm not reliant on audiobooks.
The funny thing about this change is that it's not actually new or a surprise. This new royalty model rolled out ages ago, but was in a "beta test" restricted to mid-large publishers and self-pub authors who joined it voluntarily. Me and every other author published with Podium got enrolled in it last January, saw how absolutely horrendous it was, and got ignored by many of the authors who are now panicking.
Authors involved with bigger publishers saw the change a little later. If you're aware of the discussions about Audible changes from August last year, mostly spearheaded by Daniel Greene and Robin Sullivan, then you'll know this really has been going on for a while. The only change is that Audible is now forcing it on everyone, but that was a foregone conclusion if you had a brain.
It's a real "you reap what you sow" situation, as Audible has been boiling the water for years now while a lot of authors actively defended how hot it got, and I'm honestly too tired to really give a shit.
There'll probably be some big movement later this year about Audible alternatives and attempts to get readers to start buying outside Audible. It's difficult for me not to be very cynical about this. There might have been a real chance to build a proper Audible alternative that attracted major fans if it had been attempted sincerely years ago, but the many attempts to make one over the years (and other authors feathering their own nest by dragging them down) have fragmented the author community. Some authors who have tried (and failed) to break away from Audible are already in the "acceptance" phase of the changes. Good luck to those who try to break away this time, because they'll need it.
If you do care about the actual details of the royalty model changes, I've included a reasonably detailed explanation.
On the Kindle Store, royalties are very simple. If you buy a book, 70% (usually) of what you pay goes to the author. If you borrow it via Kindle Unlimited, the author is paid per page you read, and the payout per page is calculated as a split of a monthly fund determined by the total subscriptions of Kindle Unlimited. Amazon does control the size of the fund, and therefore the payout, but it's all fairly transparent.
Audible is a complete mess by comparison.
First, cash purchases work like on Kindle. If you buy an audiobook via Whispersync or a cash sale, the author gets whatever royalty percentage they should receive. This used to be 40%/25% (depending on if the book is exclusive or not). Now it's 50%/30%.
But credits and subscription purchases are weird. Rather than pay the author a percentage of the amount paid by the user for the credit (that would be too sensible), Audible instead determines its own value per credit sale. The value is largely based off the retail price of the book (you know, that absurdly high price for each audiobook that nobody ever pays?). This retail price is used as a proxy of an audiobook's length, ostensibly meaning longer books get paid more per credit sale. But it means that setting a higher price means you receive more than other authors, and it means omnibuses receive a much higher cut (particularly as they can have retail prices over $100). This has been a known "exploit" for ages (if you can call it that), and is a reason why omnibuses are so damn popular with authors.
The new royalty model changes two things:
First (and most importantly), credit sales are now weighted based on "activity." So even after you buy an audiobook, the author's payout depend on how much of it you listen to (plus other nebulous BS Audible can use to justify paying whatever they like). The length or price of an audiobook mean jackshit. This crushes the payout of omnibuses, although it's somewhat alleviated by the fact they should have higher listening activity due to being longer.
Honestly, I'm going to say this is a good change. Part of me wants Audible to just pay a fixed percentage of the credit, but their model is trying to entice authors to make longer audiobooks. If authors receive more for longer books, then they'll release longer books, rewarding readers (who pay the same no matter what). And yeah, I'm biased because I release longer books and don't rely on omnibuses. Sue me.
The second change is the real source of a lot of angst, and probably why many authors overlooked the change last year (to their own detriment). Audible is paying for listening "activity" not sales. This means free listens on Audible Plus get paid, and that money comes from credit sales of other books. Every time you buy an audiobook on Audible with a credit, you are now subsiding someone else's Audible Plus subscription.
Audible is pushing authors to enter Plus hard. Unfortunately, Plus is an absolute wasteland right now. The vast, vast majority of books in it are AI narrated slop. I think every real publisher bailed on Plus once it became clear it was a disaster. Podium was one of the last in it, but has been pulling all its books this year as the 12-month enrolments ended (Demon's Throne and Neural Wraith left Plus earlier this year).
In the end, it's a disaster for authors. It's also one that has been predicted for a long time. A lot of people are making the surprised pikachu face when the company that has repeatedly done everything in their power to fuck over authors has, in fact, fucked over authors once again.
Anyway, that's why I give my kids Audible Plus. It's not the best platform, but it's the one we have and it's allowed me to... [generic infomercial blurb fades into the distance]
- - - - -
If you're reading this, you (maybe?) read the rambling above. To be honest, I nearly didn't include them. The extent to which I care about the broader stuff going on is far less than it used to be, because it doesn't really affect me. 2026 is a year where things just continue on. The AI slop tidal wave hasn't impacted my sales. The Audible shit shook out last year.
Mostly, I'm keeping on and writing. Stellarforged is a new challenge, given the huge shift in style and content. We'll find out within 2-3 months how much the broader community likes that shift.
Until next time. Hope you enjoy the Stellarforged chapters. They should resume next week.