Christmas, Cryptic, Coke, Charli XCX And All The Rest

I didn't post yesterday, even though I could have. It felt like skipping school, by which I don't mean it felt fun. I was never one of those kids who enjoyed an unjustified day off. I could never quite forget I ought to be somewhere else. This was a little bit like that. Especially since I knew I had a busy few days coming up, what with work and other things, so I legitimately wouldn't be posting then. 

Also, there's the post count to think about. I hit the big three hundred in 2024 and it's on the cards for 2025. I need another 48 posts in thirty-three days. With the Advent Calendar to come it's doable. But only if I don't keep bunking off...

Speaking of which, let's start there.

The Inventory Full Advent Calendar 2025 

2022 was the first year I did one. I went with stock images and a single tune, all of which I sourced myself. It was hella fun so I did it again in 2023.

That was the year I tried to get the LLMs to find me some obscure Christmas songs, a job they signally sucked at, choosing instead to just make stuff up. AIs were still cute and funny then, like puppies. 

Not that everyone agreed. I had more success using AI image generators for the pictures but it lost me at least two readers, who made a point of letting me know they were going in the comments. I imagine quite a few more just stumped off in silence.

It didn't stop me calling on AI for help the following year but I did at least make some allowance for reader sensitivity. In 2024 I had the brilliant idea of doubling up with two tunes for each day, one Naughty, one Nice. I picked all the songs myself but for the images I used my own photos for the nice numbers and had AI do the nasty ones. That way, people could opt out of AI and still get to open a door every day.

It was a lot of work, not just the double-posting but because I also decided not to look at any pre-existing lists of Christmas songs. Too much work, in fact, which is why this year it's back to one window a day, although some days, when you open the flaps, you may find more than one song. 

This year, too, in keeping with my mild disillusionment with the way AI is going, there will be no AI involvement at all. The pictures will all be Public Domain images sourced from art galleries, museums and any other archives I can find. 

That has turned out to be a lot more work than you'd think. I can see why people use AI. It saves so much time. I'm already thinking that next year I might start with an AI image and then distress it, the way I do the non-AI pictures I use for music posts. That feels like it would be both faster and more fun and also legitimately involve at least some degree of human creativity.

As for the songs, I learned from last year's experience. Trying to find a couple of dozen unusual or obscure Christmas songs without looking at any lists or collections or playlists of unusual or obscure Christmas songs anyone else has done is a) really hard work and b) exhausting. 

This time I haven't even attempted to be original. I looked through a lot of other people's Christmas lists and cherry-picked. That led to me finding other songs that weren't on the lists,   so at least some of it is my own work. It sure went a lot faster this time, anyway, and I felt a lot less burned out by the end.

It's still quite annoying not to talk about the songs as I post them but I think that would kind of defeat the point of an Advent Calendar. The last couple of years, Redbeard has been kind enough to leave a comment on each day's entry, which gave me the chance to monologue about them in reply. If no-one comments this year, which is more than likely, I'll just have to do a separate post at the end. Don't think I won't!

Next up...

Oh, this is a Grab-Bag post, by the way. Did I not mention that?

Good News/Bad News

 

Which is it? Hell if I know. When I read that Embracer was selling Arc Games and Cryptic and that it was a management buyout, I had a couple of thoughts. The first was "I guess that's good news for Cryptic" and the second "Who is Arc Games?"  

Generally speaking, I'd say that if you're playing an MMORPG and the company that owns it sells it to someone else, you'd probably prefer it went to the people already making the content for it. Always assuming you're enjoying the game, that is. It seems like the safest option, aesthetically and emotionally, although possibly not financially.

Also, if your game was currently owned by Embracer, I guess you'd be forgiven for thinking anyone else would be an improvement. Then I read the article about the buyout on MassivelyOP, which clarified some things and confused others. 

Arc Games is what Perfect World Entertainment (PWE) turned into. It looks like it's some of Arc's management team that's mostly involved in the buyout, funded by Chinese company XD. As MOP points out, though, Embracer had already gutted Cryptic and handed the remnants over to one of Embracer's own divisions, DECA, while apparently Arc Games remained responsible for publishing Neverwinter, Champions and Star Trek Online. So that's as clear as concrete.

I haven't played any of Cryptic's games in years but I have at some point played them all. Champions I just didn't get on with. STO, I thought, was a decent enough game but I have very little affection for the Star Trek IP and I think you'd need quite a bit to want to keep playing for long.

Neverwinter, though, I did like. I played it a fair bit, on and off, although in common with most of the literally hundreds of MMOs I've dabbled with over the years, I never got all that far. There are a couple of dozen posts about it here, though, and of the three, it's the only one I occasionally think of going back to visit.

Whether this latest development makes that more or less likely is hard to predict. If it's just a case of changing the names on the board-room door then no, I guess it doesn't. If it actually leads to a material difference in the way the game is developed, then yes, maybe.

The real question, as always, is who is playing any of these games now anyway? I guess someone must be because the buyer, fancifully named Project Golden Arc, paid  Embracer $30m, which seems at the same time a lot and not much. Thirty million for three established and operating MMOs sounds pretty cheap when you consider EG7 paid ten times that for Daybreak just five years ago. On the other hand, I repeat, who plays Cryptic MMOs in 2025? Looked at that way, it sounds like a lot of money after all.

I'm guessing this will be an ongoing story. I look forward to further developments.

Oh God, Not More AI Nonsense...

I could probably do two or three posts a week just on news items about AI that turn up in my feeds. Most of them I read and forget about but a few really demand some kind of comment or examination. Especially the ones about music. Here's an example.

Right after I published yesterday's post, I got the news that Warner had done a deal with Suno. This follows a deal Udio signed with UMG last month. Udio and Suno are the two leading generative AIs involved in creating music, Suno being by far the superior of the pair, in my opinion.

Until very recently, both were facing lawsuits from the three giants of the record industry, two of which are Warner and UMG. As far as I can tell, Warner is still suing Udio and UMG is still suing Suno. The third, Sony, is suing both of them. I guess the music stopped when it came to Sony and the parcel was already all unwrapped. Pity there wasn't a third AI in there, then they could all have gone home with prizes.

This is exactly what we should have expected, anyway. The way the whole AI scene is going now feels like a mirror of what happened to all the previous technical innovations that threatened the huge media conglomerates. First they try to close them down and then they buy them out. After that it's either close them down themselves or assimilate the technology and use it for their own purposes. 

It's great for the sell-outs, who get a big payday and don't have to go to jail, but it's rarely - probably never - good news for either the current users or the artists trying to protect themselves from their predations. Arguably, it may be good news for all the hundreds of millions of people who didn't even know the tech existed until the majors repackaged it and sold it to them. That's how we got Spotify, after all. Then again, look how well that turned out...

Immediately I read the news, I went straight to Suno to see how they were spinning it. I was not impressed.  They claim "You’ll still be able to create original songs the way you love today." but I very much doubt that will last long. They've already announced changes, coming "soon", meaning "a paid Suno account will be required to download songs from the product, with each paid tier enabling a specific number of downloads each month."

That seems weird to me. I guess it depends how many downloads you get. If it exactly matches the maximum number of songs you can make under a given payment plan, then it's a non-issue. If it's fewer, though, then it makes very little sense. Why let anyone download a percentage of the songs you allow them to make? Unless, I suppose, you intend to charge them extra for the rest. Double-dipping, I think they call it...

We'll see the details soon enough but it's not going to affect me directly. I have already downloaded all my finished songs as MP3s and burned most of them onto CD so they can't do much about that. I have a few days left before my current Pro subscription runs out and I'm currently downloading everything again, this time in .wav format, along with a bunch of the also-rans, too. 

I'm sure as hell not going to download everything I've made. There are several thousand versions of songs of mine on Suno now because you literally cannot delete anything you ever make. Most of them I never want to hear again, so I won't be wasting hard drive space on them. I still own the copyrights under Suno's own agreement anyway, so they can just sit there forever as far as I'm concerned.

I will not be renewing my subscription when it ends, either. Unless, of course, they offer me another huge discount. Then I might. I also might come back for another go when all those Warner artists, who must be just so excited so be signing away their rights so we can all start remixing their work, which looks like being the game plan. That sounds like a fun toy. I have a whole post on that but it's still brewing.

You Don't Have To Be An AI To Deepfake

Is deepfake a verb? It is now. 

This made me smile. We hear so much about AI copies of famous people, both their images and their voices, it was oddly refreshing to read about such an old-school, analog way of doing much the same thing. 

Johnny Cash's estate is suing Coca Cola over a commercial Coke has been running for a few months, in which - allegedly - a Johnny Cash tribute singer was hired to sing in the lamented country star's signature style. For about five seconds, while some guy (If we're supposed to know who he is, I don't..) wanders through a crowd of sports fans. 

It hardly seems worth complaining about, let alone paying lawyers, but I guess if you have a brand you have to protect it somehow. Shame to think of Johnny Cash as a brand but that's the 21st Century for you.

Some TV Talking 

I had plans today to review the second season of Man on the Inside, along with a quick sidebar on the first episode of Stranger Things, Season 5 but a full review is a post of its own so I'll just say of MotI I liked it but not as much as the first season. Not sure the format can stand a third. It already seems to have turned into more of a family dramedy than a comedy crime show. A full review may or may not follow.

ST 5.1 was a trip, though. I was genuinely taken aback by just how big a deal it felt, especially given I came late to the party. If this was the old days of broadcast television it would have been one of those water-cooler moments, I guess.

Netflix has taken the Wednesday option and pumped it up, dropping the season in three parts, with the final episode also being screened in 350 movie theaters across the USA. It is the streamer's most successful series ever, so I guess they want to make the most of this last hurrah.

The first tranche is four episodes and I did consider watching all of them back-to-back. They're all somewhere around 70-80 minutes long so it's an afternoon, not a whole day. 

In the end I opted to stick to my established practice and just watch one every evening. I watched the first last night. It started slow and ended fast. I was gripped from the intro to the credits and I am not going to say anything about it other than that. I'm doing my absolute best to avoid all spoilers, including what anyone else thinks about it, which believe me has not been easy, so I won't contribute to anyone else's problems if they haven't already seen it.

Always Leave 'Em Singing

Yes, the obligatory musical ending. I have so many songs bookmarked now. Any opportunity to shed one gratefully taken.

House - Charli XCX (Feat. John Cale)

From the soundtrack of Emerald Fennel's adaptation of Wuthering Heights, already controversial before it's even been released. My first and so far only experience of Emerald Fennel's work was her novel Monsters, which at the time I thought was the most amoral Young Adult book I'd ever read. 

It still is and I still do. Only now she's famous so we have it faced out with a Recommends card. Which I did not write. But I would have, if anyone had asked me.

John Cale, of course, among his many, many accolades, is the man I saw perform at the Cambridge Corn Exchange, only a handful of years after the crimes in question, wearing a copy of the signature mask used by the infamous Cambridge Rapist. Even in a hall full of worshipers and admirers, it caused something of a stir. 

Among this company, Charli looks like the nice, polite, well-behaved one...

It's Real If I Say It's Real

Yeebo dropped by the comment thread on the last post I wrote about Blue Protocol: Star Resonance, to mention how he'd started playing Honkai Star Rail a while back and was somewhat enjoying it, until he spotted a sale on Neir: Automata, which he'd heard described as "one of the best games ever created by humans". And that was it for his time with HSR.

That got me thinking about a lot of things, some of which came up in my reply, like how I really never get on with the combat in Honkai Star Rail and how any combat that isn't really easy puts me right off any game these days. Then I got to thinking about my gaming habits in general and how they've changed, both over the whole of my life and more specifically in the last few years.

I was going to write something quite specifically related to that today but then before I got down to it, I read this article at NME about a large-scale survey commissioned by French music streaming platform Deezer to find out if listeners really can tell the difference between music made by AI and music made by humans. And this one at GamesIndustry about the absence of player pushback over the use of AI in mobile games. And this one from the same source about Nexon suggesting everyone should assume every game company is using AI...

All of which made me think even more. Which is why this post is all over the place. I'm still thinking. But I have to start somewhere...

Let's begin with the shift from MMORPGs to Open World RPGs and Open World Survival Games, which for simplicity's sake I am going to lump together. I could also have linked a bunch of articles on that but I'm going to stick to my personal experience and some general observations because why do proper research when you can just wing it? That's how all the best columnists do it, anyway.

It seems hard to argue that these kinds of games, gacha or otherwise, haven't largely eaten the MMORPG sector's lunch. Still  as Sony/NCSoft's announcement of the in-development Horizon MMORPG, Steel Frontiers, proves, there is still a degree of interest in and commitment to the genre outside of its established, specialist niche market. MassivelyOP, who always nitpick over genre tags, were very keen to point out the acronym "MMORPG" appears right there in the title screen. 

And it is something worth mentioning. A lot of developers and publishers in recent years have gone out of their way to call their games anything but MMORPGs, believing to do so would harm their chances in the wider market.

Is that true? No idea. How would I know? Certainly, every new AAA game that claims to be an MMORPG seems to attract a million players on launch day. But then 90% of them are gone in a month and the rest a month after that. 

Was it because the games weren't MMORPG enough? Too MMORPG? Just not very good? Or were most of those people only there because it was the Big New Thing and a Thing can only be Big and New for so long?

Search me. No idea. And neither do the devs, apparently, because it keeps on happening.

What I can say, though, is that the naming of things is important. We should all know that as fantasy fans. To know a thing's name is to control it. 

Why, though? Why does a name, a true name, hold so much power?

It's all about authenticity, isn't it? That elusive, nebulous, indefinable quality that we know when we feel it. The Massively editorial team knows when a game is an MMORPG, regardless of what the press release they just received tells them. Just like we all know an NPC we're listening to was voiced by a human, not by AI. 

Except, do we? I can't help but think of the old Coke vs Pepsi test. That wasn't a notional thought experiment. It wasn't even something set up in some side-room in a University somewhere. It was an actual, physical test you or I could try for ourselves, when we went into town to buy an album or some new trainers.

If you search "Pepsi Challenge" you'll get the idea it only happened in America but I remember seeing the van parked in the shopping precinct near the bus and railway station in a city where I lived. I just can't remember which city... 

I didn't try it myself but I was always absolutely certain I could tell the difference. Pepsi is a lot sweeter than Coke, to my taste at least. Now, if it had been Coke vs Canada Dry, my all-time favorite cola, I'm not so sure. 

Which is kind of the point. Maybe I can tell the difference between Coke and Pepsi. Certainly, to me, they don't really taste all that similar. Between Coke and Canada Dry cola, though? Those two are so close I wouldn't like to put money on it.

Whether I can tell or not, though, I think I know. Do I want to put it to the test? Not really. Why would I want to find out I was wrong? How would that make my life any better?

I have listened to a lot of music in my life and until very recently all of it was made by humans. There wasn't really any other option. Despite that intense and continued exposure, I don't doubt that, like 97% of the 9,000 people Ipsos tested for Deezer, I would not reliably be able to differentiate between music produced by AI and music produced by humans.

Mrs Bhagpuss says she can. She really likes the music I've made with AI but she says it all sounds "pink". She can hear the pinkness in it the way some people see auras. I cannot hear the pinkness. I did think about playing her something extremely similar that was made by humans to see if she still thought it sounded pink but that's an experiment that's going to stay firmly in my thoughts only.

She also says the vocals sound "too perfect".  I hadn't noticed that but after she pointed it out... I still couldn't hear it. What I think I can hear is the AI being imperfect on purpose sometimes, which isn't the same thing at all. And I certainly couldn't pick an AI singer out of an audio line-up based on perceived flaws.

But then, the AI vocals I'm listening to, the ones in the songs I've caused to be made, like a Renaissance artist overseeing a workshop filled with talented but anonymous craftsmen, are my vocals. The imperfections are frequently my imperfections, replicated as though they were intended.

I remember Tipa, who also uses Suno, mentioning a while ago that, while she liked the music she made with the software, she hadn't found much she wanted to listen to by anyone else there. I'd go further than that. I haven't heard anything there that I haven't found intensely irritating. The app defaults to playing the next tune on some playlist or other if you don't stop it, so I've been forced to hear snippets of lots of AI tunes and there hasn't been a single one I haven't almost broken the keyboard trying to silence.

But is that because they don't sound like they were made by humans or because they're just terrible songs? I could do much the same on YouTube and many of the tunes would be "authentic". They'd just be awful. A lot of people who can't play an instrument or sing also have terrible taste in music. Suno lets them share their lack of talent with the world, too. AI is really egalitarian that way.

What are we really lookng for when we listen to music or look at images or play video games, anyway? Authenticity or entertainment? Is a real, bad thing better than an artificial, good thing? And anyway, what even is authenticity?

At the moment I prefer Open World RPGs to MMORPGs and I prefer Eastern games to Western games. I'm not saying this is a permanent change or even a lasting one. It's a snapshot, like all preferences. It may stick or it may slide.

Looking at that preference as objectively as a person can observe a subjective preference of their own, I'd question some of the assumed positions on authenticity that come up repeatedly when games and especially RPGs are being discussed. There's long been a trend in the discourse over automation. It predates any queasiness over the use of generative AI, although that does seem to have intensified the and polarized the debate considerably.

MMO players in the West have tended to react very negatively to many of the things that are currently drawing me towards both open world RPGs and mobile ports and which a few years ago led me to appreciate a number of imported games that were calling themselves MMORPGs.  

One day I'll write a proper post about why I like these sorts of games but for now, here are a few of the more obvious reasons. 

 

I like the brighter colors and the flatter surfaces of the graphics, for a start. I like the cleaner textures. A lot of older or more traditional MMORPGs look gritty, somehow. Dirty, even. I can deal with that look but I'd rather not have to.

I like the stories, which seem a lot more modern and relatable than those in Western games. The characters are younger and more enthusiastic. The themes are stronger; the emotions clearer. 

There's a tendency to call them "anime" games but they could as easily be called "YA" games instead. I read a lot of YA novels (The acronym stands for Young Adult, marketing-speak for what publishers used to call "teen fiction".) and the characters in many of the games I'm now darting between remind me very much of the ones I meet in those books.

Ironically, these games, clearly aimed at a younger demographic than the traditional Western MMORPG, also tend to have a lot more time and respect for older characters. In most of the MMORPGs I've played, the characters are much of a muchness when it comes to age. 

The Elves all live forever so they're ageless. The dwarves are all old even if they're young. The humans are inevitably somewhere in their 30s or 40s. The anthropomorphic races (And the gnomes.) tend to be child-like. Mostly, though, unless a character has to be a specific age for a plot point, age barely even rates a mention.

The open world rpgs and anime games give me stories across the full age range, from small children to the elderly. And those stories frequently reflect the kinds of concerns real people in those age ranges would have. It's not all gods and mosnters. Sometimes it's homework or rivalries at work or the way your hip doesn't want to let you climb the stairs like you used to.

That feels more authentic to me but I'm betting it's a black mark against authenticity for anyone looking for the traditional, high fantasy MMORPG experience. Still, a lack of authenticity in the story is nothing compared to what happens in the gameplay. 

As I said at the start of the post, I bounced off Honkai Star Rail partly because I found the combat too much like hard work. I dropped Genshin Impact because I literally couldn't beat a boss to carry on with the story. Not all of these games have Combat For Babies enabled. Just the ones I like.

After a quarter of a century and more, I think I can say I'm officially over finding combat in MMORPGs fun for its own sake. I never liked it that much but it did used to have its moments. Now, it's almost always a means to an end. The easier it is, the better I like it. I like one-shotting mobs. All of them, if possible.

The received wisdom is that making combat too easy turns players off. They get bored and go somewhere else if the challenge isn't there. Not me. I get bored and go somewhere else if the challenge is there. One of the things I like best about BP:SR is the auto-combat. I use it in every fight. It's even better than one-shotting mobs because I don't even have to press a key.

Except I do press some keys, sometimes, because that's fun, too. I dodge a bit now and then. Jump about. Change position. Not sure if it makes any difference but it makes me feel like I'm involved. Because you want to feel like you're doing something, don't you? You just don't always want it to be true. 

Authenticity is in your head. There may be an objective reality out there but you do not have access to it. You think Coke tastes better than Pepsi because your eyes tell you so when you see the label on the bottle. Your taste buds have no say in it.  

That NPC you hear, the one that sounds so flat and uninflected? It might be AI. Or it might be a voice actor who isn't doing such a great job. That song you like? The one that came up on that auto-generated playlist that's by someone you never heard of before. Are you sure it's a real person singing? 

Yes, you know. Of course you know. But how are you ever going to know?

And what about the fun you had, playing that game?  No, wait...   

I won't say the fun you had. You may not have had that fun. I'll say what about the fun I had, playing that game where the AI (Different kind of AI, of course. The old, good kind.) did all the fighting for me. It even did the running, there and back. All I had to do was take the quest and hand it in. Did I really have fun or was I just imagining it?

Maybe I was having more fun all those times I spent an hour trying to beat some stupid boss in a Guild Wars 2 Living World instance. One of those times I lost so often all my armor fell off and I had to give up and leave. That time I had a headache for an hour afterwards. When I felt like uninstalling the damn game, I hated it so much. 

That was some authentic gameplay there, right?

Yeah, I don't miss any of that. What is authentic isn't the experience but  how I feel about it. If the game feels like it was fun, it was fun. If the song sounds good, it's good. If the voice acting is convincing, it's convincing. 

And that would appear to be why I prefer the games I prefer just now. They're authentically entertaining. Whether they're any good...

Well, that's an entirely different question. 

Nobody Wants This More

I'm back to watching an hour or two of TV every night, right before I go to sleep. Partly that's because we're coming into the peak season for returning shows, when the days get dark before it's even nighttime and all the stations and platforms start to make a grab for attention by bringing back all those shows they think people liked last year. 

And it works! I can't say I've been losing sleep waiting for the next season of Hazbin Hotel but I was unironically delighted when I logged into Prime Video last night and saw the familiar logo splashed across the top of the screen, somewhat predictably just in time for Halloween.

It'll be a while until you see my review of that one here. Amazon has chosen to employ some faux-retro scheduling, dropping a pair of episodes every Wednesday for a month. Expect a review in December.

The returning show I'll be reviewing today, whose second season I've just finished watching, is Nobody Wants This, the Kristen Bell/Adam Brody romcom. It's on Netflix, who decided to let us have the whole thing at once in the equally old-school style of a DVD box set. I didn't binge but I did end up watching more than my regularly-scheduled one episode per evening, which may offer a clue as to how much I enjoyed it.

The other reason I'm back to watching TV at all, let alone having the time for two episodes of the same show some evenings, is that I've stopped making songs with Suno. For the last six months, I've spent two or more hours, pretty much every night, working on original material. It's been verging on an addiction.

The reason I stopped has nothing to do with losing interest or even feeling I ought to take a break. I just ran out of source material. In half a year I've completed "definitive" versions of well over a hundred original songs, for which I had to generate thousands of works-in-progress (All of which still exist because you literally cannot permanently delete anything you make on Suno.)

The first batch were all actual songs from decades ago that I pulled off of decaying cassette tapes but the great majority (And by far the better results.) are scores of new songs I created by cutting up prose I wrote in the nineties. About a week ago, I finally hit the point where I couldn't find any more sections or paragraphs that felt like they could be converted into lyrics. So I stopped.

I didn't plan on stopping. I was going to start going through the older, less-successful songs and work on better versions. I've found I have by far the best success with Suno v4.5Pro, which didn't come into existence until May and which wasn't as solid when it was introduced as it is now. (Don't talk to me about the downgrade that is  v5.0...)  

That hasn't happened. Not yet, anyway. I did spend some time playing around in Suno, getting Gemini to give me prompts for certain artists or bands, then having the AI do covers of my songs in those styles, but that was more of a party game than any kind of serious project. For the moment, I think I'm done with making new songs with AI. The next thing is to figure out what to do with the half-dozen albums-worth of material I already have. 

I was going to say that I've unsubscribed from the app but when I went to do it a few minutes ago, I got one of those "Please don't leave - we'll  give you a really good deal" offers and it was so good I took it. Suno is already ridiculously cheap at $10 a month but at half-price it seems daft not to keep the sub going for another month. Also, I'm curious to see how many times they'll try to bribe me to stay.

That's a long pre-amble to the substantive purpose of the post, which is for me to tell you what I thought of Nobody Wants This, Season 2. It's possible I might be playing for time. I'm not entirely sure what I did think about it...

I liked it. I enjoyed it. I'd watch a third season if there is one. It hasn't been confirmed but apparently the stars are up for it and "the writers' room is working" on one so the omens are good. I'm just not sure if I buy the premise that the central characters are made for each other...

Starting at the end, since that seems to be where I've gone, the season doesn't really finish with one of those annoying cliff-hangers, nor even with an obvious bunch of loose ends left blowing in the breeze, all begging to be neatly tied up next time around. There are plenty of unfinished storylines, for sure, but they all have more of a slice-of-life feel to them, the inevitable unraveling of lived lives we just happen to be seeing play themselves out, rather than anything that demands urgent resolution.

Time for the spoiler warning, I guess...

That final episode! Hah! I loved it! Also, it made me want to watch When Harry Met Sally again, which is never a bad thing.

Everyone breaks up. Well, the three central couples. Then one pair gets immediately back together, one goes "on a break" (We all know what that means...) and the third is fucking cooked - thank God! All so very satisfying. If you're going to invoke the spirit of Nora Ephron in your title you better have a goddam plan and they did.

So the wrap-up left me feeling like I'd gotten closure. Always what you want, coming out of a season. But how was the build up?

Hmm. Confusing, I'd say. I mean, come on, I love Kristen Bell. We all love Kristen Bell. But do we love Joanne? I'm not sure we do.

She's kind of annoying, isn't she? And a lot more so in Season 2 than last time. In the first season it seemed like sister Morgan was the self-centered, solipsistic one but this time around she came across as considerably more vulnerable and capable of self-analysis. Also sadder. 

Okay, the pair of them are terrifyingly judgmental and prone to acts of petty vindictiveness not usually seen outside of high school but at least Morgan seems to be learning something, occasionally. Joanne is mostly just doubling down.

And then there's Esther, the sister-in-law, who acted like such a terrifying hard-assed bitch in Season 1.  She's much warmer, more nuanced and yes, again, vulnerable. A lot of character growth there, too. Possibly more at times than seems entirely likely but then much of what was there before was obviously an act. Obvious now, that is.

Even Rebecca, the fiancee who gets dumped at the start of the first season and who just gets a couple of cameos here, even she seems to have grown. Joanne, though? She might have shrunk.

All of which is in the writing and the playing. It's a great performance as always by Kristen Bell but it isn't very endearing. There were multiple episodes where I didn't like Joanne all that much. Of course, she was frequently at her funniest when being at her least likeable and this is a sitcom so it's a valid trade-off. Plus "endearing" isn't really Kristen Bell's thing, is it, now I come to think about it. 

In the end, though, for a romantic comedy to work, the audience has to want the prospective couple to get it together and there were plenty of times I was rooting for them to realize they were dead wrong for each other and call it quits. 

Which, I think, might even be the idea. You're supposed to feel that until the very end, when they turn it around. And they do. It is, nonetheless, a lot of Shawshank before the redemption, as the saying goes. Also, it might feel a lot different on a re-watch. And no-one just watches these things once, do they?

So that's Joanne. How about Noah?  Geez, but he's infuriatingly un-ininfuriating! I feel like I ought to be able to say I wanted to punch him in the face but it would be like punching one of those inflatable bosses they supposedly have in the basements of office buildings in Japan. He'd just roll around a bit and come back up for you to punch him again and it wouldn't get either of you anywhere.

He's so fricken' unbelievably nice without really being very nice at all. He has so many barely-concealed anger issues with which he is not dealing at all but he has a seemingly-bottomless well of niceness he can smear all over any emotion he feels to tamp it down so he can pretend it isn't there. And it works! For him it works, anyway. Not so much for anyone around him or indeed anyone who has to watch him doing it.

The other thing about Noah is that he's hardly ever funny. He's mostly a straight man for others but he does occasionally do a kind of pained fish-out-of water thing, like when he's running about between the Strongly Agree/ Strongly Disagree markers with the teens at his new job and getting it wrong every time, something that passes for humor only until you think about it. 

Also, is he really that good-looking? I have issues with the scraggly facial hair but I guess I'm not the target demographic.

He does a job, anyway. It's fine. And he is good when it matters, in the non-funny bits. He convinces me every time he cracks and tells Joanne how he really feels about her. If Adam Brody couldn't sell that, the show would be in real trouble. But he can.

The really funny characters, though, are... all the others. Another generic problem with romcoms, of course. The couple at the center can't always be making it with the one-liners and the funny faces and the pratfalls. They have to gaze meaningfully into each others' eyes once in a while.

The show is full of hilarious couples that aren't Joanne and Noah:  Esther and Sasha, Sasha and Morgan, Morgan and Dr. Andy, even Joanne and Morgan's parents ffs. Every time any of those pairs gets going it's like someone switched the "Sitcom Filming" light on. 

I would personally love a spin-off starring Justin Lupe and Timothy Simons as Morgan and Sasha. They have a ridiculous chemistry that just burns off the screen. Not specifically a romantic one, either, although if the script took things that way it would be entirely convincing. It's the way they communicate so comfortably, intuitively and naturally that makes me want more of them together.

The whole dynamic between Sasha, Morgan and Esther in Season 2 is so unexpectedly lovely it almost overwhelmed the main plot for me. It's rare, even now, to see this kind of male/female friendship between characters, who could viably have a romantic relationship but have something different and deeper instead; to see it alongside recognition and acknowledgment of its value and purpose from a partner, who could very plausibly be expected to be threatened and concerned by it, is astonishing.

It is a clever show, though, which shouldn't come as any kind of surprise. The one thing that ties all Kirsten Bell's projects together is their intellectual rigor. They all make you think. Hard.

At which point I ought, I suppose, to say something about the central conceit, which is that he's Jewish and she's not. Only I'm not going to, other than to say I found it confusing and distracting in the second season in a way it wasn't (So much.) in the first. I feel as though it's adding something to the show that makes it harder to analyze and understand than a romcom normally would be and harder still to talk about. 

Did I learn anything from it? Hmm. Maybe. Maybe not. Except the bit about how you can apparently just decide to be Jewish. That was new. And some people make a big deal about gender!

What I would say is that regardless of whether Noah is a rabbi or a vicar or a policeman or the Wichita lineman, anyone who openly states that marrying them would be, first and foremost, marrying a job does not make for a very convincing romantic lead, even without all the angst over changing religion. There's a lot of talk about "red flags" in Dr. Andy's behavior that could easily apply here, too.

"Nobody Wants This" is a very interesting title. It's the name of the sisters' podcast but it also refers to the central relationship, in which it's never clear who actually does want it. The text is that even if the pair at the center do (And at any time one or both of them might not.) any number of outside forces certainly don't. The subtext is that all those outside forces might be right. Maybe the universe doesn't want it either.

I do, though. The show, that is. There were plenty of moments where what one or other character was doing annoyed me more than amused me but that's the grit that builds the pearl. I'm still not remotely convinced that Joanne and Noah ought to be together. As things are now, I do kind of hope that the series finale, when it eventually comes, sees them go their own separate ways. I think the prospects of them having a long and happy marriage are vanishingly small. 

Until then, though, I'm more than happy to watch them struggling with the same doubts for another ten episodes, this time next year. And for a few years after that, if Netflix so allows.

Don't You Know Who I Am?

Naithin posted something the other day about his experiences with the AI music generation app Suno, which I've been using for most of this year. As I said in the comment thread to his post, it's largely replaced playing video games for me, in the way playing games once replaced watching television. I very much doubt it'll hold my attention for more than two decades, the way MMORPGs did, but for now it's still feeling quite compulsive.

There's probably a whole series of posts I could write about various aspects of making music this way - practical, philosophical, aesthetic, ethical, legal - but not right now. Today I just thought I'd share an idea I had for a parlor game, to be played using your LLM of choice, an idea that only occurred to me last night as I was asking mine, which happens to be Gemini, a question.

Mostly, when I'm working with Suno, I make up my whole prompts from scratch. I do re-use a lot of them, often with certain key phrases changed, trying to finesse the exact result I'm after, but I'm not usually trying to re-create the signature sound of a specific band or singer, more the sounds I hear in my head.

Occasionally, though, I look at something I've written and wonder what it might sound like if it had been covered by someone I like. Or, for that matter, don't like. Even more occasionally I find myself wanting to make something that sounds entirely unlike anything I'd come up with on my own, both for inspiration and variety.

To that end, now and again, I go to Gemini and ask it to give me a prompt for Suno that describes how a particular genre or band sounds. Feed those into Suno unedited and the results aren't often all that close. You can always recognize the influences but it's not like you can copy and paste from one AI to another and pump out convincing ersatz albums by your favorite acts like some kind of amateur Top of the Pops knock-off operation. Well, not yet, anyway.

It would be an amusing party game to generate a bunch of those, play them and have people try to guess who they were supposed to be. I can imagine that being a lot of fun with enough drink taken. If I thought anyone would actually take the time to click through and listen, I might have knocked out a few myself and included them here.

But that would be a lot of work and I'd be surprised if more than a handful of readers would be interested enough to listen to even a few bars. That sort of thing really requires a bunch of people together in a room to be fun. 

What would be much easier and also a better game for a blog would be if I just copied and pasted a few of Gemini's prompts to see if anyone can recognize the acts from the descriptions.

Gemini is pretty good at describing what singers and bands sound like. Disturbingly good, actually. The shorthand explanation of how "LLMs work by analysing vast banks of data to ascertain the next most likely word" seems to collapse pretty quickly when you look at the results. Self-evidently, there's more to it than that, or the output wouldn't be so convincing. When I saw the descriptions I was getting, my first thought was that Gemini had to be precising a whole bunch of reviews and picking out the most commonly-expressed opinions but even if that's all it is doing, it's weirdly accurate.

What it mostly makes me think is that well-known bands and singers have such a definitive, known, recognizable style it can quite easily be encoded in a couple of paragraphs. A test of how accurate that encoding is might be to ask someone if they could tell the band from Gemini's description of what they sound like.

And if that's not a parlor game, I don't know what is.

Oh, one other thing. When I started doing this, Suno had a two-hundred character limit on prompts. That forced Gemini to be very succinct. I think those make for a better game, although they're far less interesting in themselves than the ones it puts up for the new thousand-character limit. Those are more like entries in a musical dictionary.

Shall we try a couple? Of course, there is the slight problem that some of the bands I may have been asking about aren't as well-known to the readership of this blog as they might be but we'll see how we go. Let's start with a few short ones, all of them pretty famous.

1. "Epic rock ballad, theatrical vocals, powerful piano, soaring melodies, dramatic drums. Over the top emotion."

 2. "Clanging percussion, gravelly vocals, bluesy growl, broken piano, dark carnival vibe. Storytelling song."

3.  "Dissonant blues, fractured rhythms, raw vocals, free jazz elements, odd time signatures. Experimental rock." (Gemini apologized before offering this one, saying "It's challenging to precisely replicate the uniquely abstract and complex sound of ------ within the constraints of a short Suno prompt."

 4. "Synth-pop, melancholic lyrics, electronic beat, dramatic vocals, catchy melody. 80s/90s dance-pop."

5. "Dark, narrative song, brooding vocals, piano/organ, gothic rock feel, intense atmosphere."

Answers, in classic magazine style, at the bottom of the page. Or better yet, click through the number to see the act in question in full flight on YouTube.

I think most of those are fairly easy to guess but it's always hard to tell, if you already know the answers. 

Much more intriguing and educational are the longer prompts. Some of these really made me think about the artists in question, sometimes in a different way to how I would usually have considered their work. Unfortunately, these are really too long to include more than a handful in a blog post and, more problematic still, most of the ones I've done have been for bands I wouldn't really expect anyone to get.

This, for example:

A blend of chamber pop and indie rock with dual vocals

Create a song with a delicate, chamber-pop feel. The melody should be intricate and pastoral, driven by acoustic guitar, cello, and oboe. The vocals should be a male-female duet, singing with a gentle, melodic, and slightly ethereal quality. The rhythm should be subtle and understated, possibly featuring soft, brushed drums. The overall mood should be whimsical and melancholic, like a folk-tale set to music, with a touch of art-pop sophistication

Anyone? Anyone? Bueller?

Okay, that's Shellyan Orphan. And I think it's pretty close, except there's not always two of them singing. But speaking of duos...

This might be easier.

Duet: Psychedelic Cowboy Lounge

Create a psychedelic, country-lounge duet with a distinct male and female vocalist. The male voice should be a deep, world-weary baritone, delivered with a cool, nonchalant swagger. The female voice should be a soft, breathy, and slightly detached soprano. The music should have a lazy, sun-drenched feel, with a loping, cinematic rhythm. Instrumentation should feature twangy, reverb-drenched electric guitars, a prominent bassline that walks slowly and deliberately, and a gentle, brushed drum beat. The atmosphere should be both sultry and melancholic, like a slow-motion scene from a western film, with a sense of fatalistic romance and hazy desert heat."

Now that is eerily accurate, I'd say. It's Lee Hazelwood and Nancy Sinatra and I specifically asked for them "as they sounded on their version of Summer Wine." 

I was curious enough about the provenance to try googling individual phrases with Lee or Nancy's names to see if I could track down a specific review or article from which they'd been taken but there was nothing. However Gemini is doing it, lifting descriptions wholesale from published texts doesn't seem to explain it.

Now three more I won't immediately reveal. At least one is blindingly obvious, I'd say, and the other two are guessable if you're familiar with the fairly-famous acts in question, although on of the two was only really well-known in the UK.

Psychedelic Chamber Pop

Create a richly orchestrated, orchestral pop song with a sense of wistful innocence and profound introspection. The soundscape should be built on a foundation of lush vocal harmonies, layered meticulously like a choir. The instrumentation should be a kaleidoscopic mix of traditional rock instruments and unconventional sounds: think harpsichord, French horn, cello, and subtle, otherworldly electronic sounds like the theremin.

The production should feel like a "pocket symphony," with a dense, "Wall of Sound" quality where every instrument and vocal part is given its own space. The arrangement should be dynamic, with unexpected melodic and rhythmic shifts, moving between serene, introspective sections and soaring, anthemic choruses. The mood should be both joyful and deeply melancholic, capturing a feeling of bittersweet wonder. The song structure should be more like a suite, with distinct musical sections seamlessly spliced together to create a unified, emotional journey.

Dramatic Orchestral Pop

Create a dramatic, operatic pop song with a grand, majestic feel. The vocals should be a deep, resonant baritone, delivered with a formal, theatrical, and deeply melancholic tone. The orchestration should be lush and sweeping, featuring soaring strings, blaring horns, and a full choir. The melody should be stately and complex, with unexpected shifts and a sense of cinematic tension. The lyrics should be poetic, dark, and enigmatic, telling a story of romantic tragedy or existential dread. The overall sound should be one of epic grandeur, blending chanson and Broadway with a noir sensibility, creating a sophisticated and deeply moving soundscape."

Ornate Art Rock

Create a theatrical and eccentric art-rock song with a baroque pop feel. The vocals should be a hyper-dramatic, trembling vibrato, delivered with a campy, highly stylized theatricality. The melody should be intricate and somewhat ornate, with a prominent, crying violin and a lush, romantic piano. The rhythm section should be understated but propulsive, providing a dynamic backdrop for the more flamboyant elements. The sound should feel both raw and refined, combining the glam-rock energy with the delicate precision of a chamber orchestra. The atmosphere should be one of decadent, melancholic glamour, full of quirky flourishes and a sense of sophisticated playfulness."

That's enough for now. Also, it's most of the ones I've done. I said I didn't do it often.

I might do it more now I've realized what a good game it makes, though. And obviously it doesn't have to be music. You could get your pet LLM to give you a short description of an artist 

("Large-scale canvases featuring soft-edged, rectangular blocks of color that seem to float and radiate, aiming to evoke profound human emotions from the viewer" - Mark Rothko.) 

or a writer 

("gritty realism, complex world-building, and moral ambiguity. He uses a multi-protagonist, limited third-person perspective to tell an intricate, sprawling narrative where beloved characters can be killed off unexpectedly, reflecting the harsh, unpredictable nature of his fantasy world". - George R.R. Martin) 

or a game's art style

("a highly recognizable and influential stylized high-fantasy look. It is characterized by exaggerated proportions, vibrant colors, and hand-painted textures that prioritize readability and a timeless, "cartoonish" aesthetic over photorealism". - World of Warcraft)

Back in the eighties, when my pals and I used to sit around into the early hours making up exactly these sorts of games, this would have been iterated on over and over. The potential for making your own entertainment with AI is huge and if you ignore the drain on the environment it seems like one of the more harmless uses of the technology I can think of. 

I might be using it for my Advent Calendar this year. I have a few ideas... 

 

 

Answers down here for those with an aversion to hypertext.

 

 

 

 

A bit further

 

 

 

 

Here we go

 

 

 

 

 

1. Meatloaf

2. Tom Waits

3. Captain Beefheart

4. Pet Shop Boys

5. Nick Cave

 

Psychedelic Chamber Pop - The Beach Boys (Specifically God Only Knows and Good Vibrations.) 

Dramatic Orchestral Pop - Scott Walker

 Ornate Art Rock - Cockney Rebel.

It's All just A Game To You, Isn't It?

Not that I want to act as unpaid hype man for Neverness To Everness but damn! Have you seen the latest trailer? 

No? 

Well, here it is. 

Enjoy!

Disclaimer: Not actually a trailer. 

It's better than that. It's the opening sequence of the game itself, apparently. More annoyed than ever about the lack of beta invite now. Grrr.

The only problem is, I want to watch that anime quite a lot more than I want to play that game. It feels like the game is the consolation prize because there is no anime. Maybe that'll swap round when I do get to play it. I hope so but I kinda doubt it. 

Raises a question, though, does it not? All these very-to-extremely successful anime-inspired games... wouldn't it make sense for them all to spin off into actual animes? Movies or shows, either way. Is there really not a Genshin Impact movie in the works? Surely there must be...

Wait...

I'll check...

Well that was more confusing than helpful. Maybe all that guff about the death of the web does have some validity after all because it certainly is full of nonsense these days. And I don't mean AI halucinations or deep fakes or any of that trendy buzzword clamor, either.

Has anyone actually tried to use IMDB for anything meaningful in the last, oh, five years, for example? Nothing there has made much sense since at least the pandemic. I used to be able to go there and get solid, straightforward details on movies, just like looking them up in a book. Now look at it! 

Case in point: do a simple google search for "Genshin Impact movie" and the top result is IMDB - Genshin Impact (TV Series 2019...) Hmm, interesting, you think... how did I miss that? And did Genshin Impact really come out that long ago? 

No, it did not. It launched at the end of September 2020. So, what then? The game is an adaptation of an existing IP? How did I miss that?

No, it isn't. There was no "TV series", not in 2019 or any other time. The entry lists 250 episodes but it looks as if what IMDB has chosen to do is list every cut scene and trailer for the game as if it was an episode in TV show. 

That's not nearly confusing enough, though. We can do much better than that. It seems HoYoVerse has been hyping a TV show based on the game for years. Or a movie. Or both. There are plenty of links to times they've said one in the works, complete with a few details on which studio might be producing it. No actual show has yet appeared or even been scheduled.

Well, except for that one that one that ran from September 2020 to August 2031, of course. I mean, we all know about that one, don't we? It was "one of Netflix's most streamed Asian programmes" and it won "Anime of the Year Award in 2022 at the International Anime Awards". 

Don't say you didn't hear about it? It's right there on the web in black and white (Actually white on black because in some parts of the internet it will be forever 1998.)

Someone put a lot of work into that Fandom Wiki entry. That is not AI. I mean, I'm not saying whoever did it didn't get an AI assist but I'm pretty sure that's all one person's fantasy there. And if they'd not chosen to project it all the way through to 2031 as though we were past then already, it would have been very convincing indeed.

So, no, as it happens I don't think AI is going to render the world-wide web unusable for any practical purposes when it comes to research. Humans got there first and are already making a pretty good job of that.

None of which was really what I was going to write about but now I've started, I might as well throw in something else I've been meaning to say about AI. Has anyone noticed how some of the AI applications that started out as services or resources are slowly turning into games?

It took me a while but once you have your eye attuned to it, you start seeing it everywhere. NightCafe, the AI image generator portal I've been using by preference for a long time, has slowly morphed from a sort of curation point for multiple models to a social media platform for enthusiasts and now it's made the full jump to becoming a game by adding... quests.

They've been using pop-ups to suggest ideas I might want to make pictures from for a long time, as though I'm going there to kill a few minutes, not to make specific images for a defined purpose, but now they're amping the casual gaming approach up to offer two-week-long challenges they've chosen to label "quests". 

Suno hasn't gone that far yet but it's well into its social media stage, with its own star creators and featured performers. It has rankings of a kind, too. I'd guess it's maybe one more iteration away from going full game.

This is the beginning of how it's going to be. AI is still bloody awful at the things the big tech companies are demanding everyone uses it for but it's tailor-made for casual time-filling. Forget the supposed danger that AI could replace the creative work of actual artists and musicians and writers in true video games. It's nowhere near ready to do that and even if it was the pushback from a very vocal segment of the audience is going to delay the (Still inevitable.) drafting of the technology into mainstream entertainment production for quite a while longer. As something for idle hands to do when there's nothing better, though? For that, it's perfect already.

The threshold for game-like entertainment at that end of the market is orders of magnitude lower than  in games aimed at console or PC gamers. It's not as though the standards of human-made games there are stellar and the expectations of the audience are infamously relaxed. If it kills thirty seconds at a bus-stop it's likely good enough.

Or  maybe that's just the view of someone who doesn't understand the market or the players. Easy trap to fall into. Still, I have heard enough about the way the Suno mobile app's being used already to understand that for many of its users, typing in a few words and having it make a song you can laugh about is considered a decent enough way to pass a lunch hour. Especially now it will happily use rude words.

Luckily, for now anyway, these apps do still retain the backend that lets them be applied to the purpose for which they were originally created, namely as creative tools. How long that will last is another matter, which is one reason I'm getting as much done now as I can before I log in one day to find the whole thing has been turned into a game. I suspect the window of opportunity for doing much other than the AI equivalent of quizzes, word-searches or sudoko may not stay open that much longer.

And once again, that won't be because of anything AI has done. It'll be because some people actively enjoy making things worse. 

Let's hope that's overly pessimistic. I'm certain there will still be professional tools using AI tech. You'll just have to pay professional rates to access them. The free stuff will be aiming at an entirely different demographic.

One thing I will commit to: it'll be a long, long time before we see anything from an AI to match the Nevernesss To Everness cut scene at the top of the post. Humans may use AI as the tool to create something like that but the AI will have about as much say in the finished product as your pen has in the words you write.

I'm not even saying I'm happy about it, either. Who wouldn't want to be able to type "Urban post-apoc magitech anime TV show with animorphs - sassy dialog, lots of explosions" into an AI app and get ten twenty-five minute episodes as good as that trailer in thirty seconds? Geez! The phrase "I'd never leave the house" would really come into play then, wouldn't it?

It's the 2020s equivalent of the 1960s jetpacks and flying cars though. Always going to happen, never does happen. 

Speaking of which, where is my jetpack? I've had it on order since at least the nineteen-seventies...

Turns Out More Is More After All

What I want to talk about today, even though I'm pretty sure no-one but me wants to hear it, is Suno. I wasn't expecting to write about AI music again but that's because, as Vic Reeves used to say, I hadn't thought it through.

Here's the thing: I spent two months generating literally hundreds of versions of about thirty songs until I had what I thought was a definitive version of all of them. I moved on to making the videos and I didn't need Suno for those, so I thought I was probably done with it, for a while at least.

I was just about to unsubscribe then, just before the sub was due to renew, Version 4.5 arrived. Which was a complete surprise to me. Even though the "Cover" feature I'd been using was clearly flagged as "Beta" and there was a drop-down menu of previous models in the app itself, it never occured to me that meant development was still going on. 

I mean, I knew, obviously. I just didn't know.

Several things changed with the update, all of them potentially significant for the project. For a start, the cover feature was much improved in 4.5. I'd been very satisfied with the 4.0 beta version but things can always get better so naturally the first thing I did with my remaining credits was generate a few covers for comparison. The difference was hard to miss. 

There was a noticeable improvement in quality and the annoying glitches almost completely vanished. Most importantly, the new covers I was generating stuck much more reliably to the uploads. Under 4.0 that had not always been the case.

One of the reason I generated so many covers of the same songs was to get one where the AI didn't decide it had a better melody for the bridge than the one I'd given it or that it could phrase my lyrics more evocatively than I could. Which might have been okay except that nine times out of ten the AI was wrong. The original was better. (The other reason I keep making more and more covers, of course, is that it's incredibly satisfying and tons of fun. I'm not saying it'll never get old but it sure hasn't yet.)

With 4.5 the AI rarely deviates from the template it's given. Hardly at all. When it does, it's usually because I'm trying to make it accomodate my lyrics and melody to an entirely different and wholly unsuitable genre, something I do mostly for my own amusement. I can hardly blame the AI for losing patience with me there.

As well as producing much more accurate versions of my originals, the new model can also make them longer. Output from 4.0 was limited to four minutes. That's doubled in 4.5. 

Not that I've got anything that's eight minutes long. This isn't prog rock. But I was having problems with a handful of songs that naturally run just over four minutes. They kept getting cut off in the coda or the final verse. 

There are ways to get around that with extensions in the old model but it's fiddly and usually doesn't sound quite right. The new model just keeps going until the song's over, which is infinitely better. And it gets the timing right so the songs have an actual ending rather than just stopping like the machine's been switched off.

All of that meant I've had to think again about the whole project. I might need to go back over all the "finished" versions to see if they really are as good as they could be. I suspect some of them are not. More to the point, I'm now aware there may be further improvements to the AI in future releases, so I'm just going to have to draw a line somewhere for sanity's sake. 

Either that or keep playing around with it until I finally get bored, something I'm showing absolutely no sign of doing so far. I have to say that at £8 a month I'm getting at least as much use and entertainment out of the subscription than I ever got from an MMORPG sub. It's cheap entertainment by the Dollar per Hour scale and you have something to show for your time and money at the end.

Perhaps the most intriguing addition to Suno's capabilities in this latest model is the new Prompt Enhancement Helper.  The AI as a whole has had its interpretative and creative powers increased and extended, something that's apparent from the results. That's probably the most important change but the most fascinating is the way it now lifts the curtain to show how it's doing it.

I had been under the impression that to get the best results I needed to keep my prompts as concise and tight as possible. I didn't think long, descriptive prompts full of adjectives and metaphors did all that much. I was convinced that even listing multiple genres or naming more than one or two specific instruments led to the AI cherry-picking a couple of them and ignoring the rest.

Based on the new feature, I had it all backwards. What happens when you use the Prompt Enhancement Helper is that your brief summary gets translated into a long paragraph, full of clauses and imagery and opinion. The exact opposite of what I expected, in other words. 

For example, if I prompt 

"1960s garage rock, female vocals" 

the Prompt Enhancer turns it into 

"Raw, punchy 1960s garage rock led by gritty female vocals. Begins with fuzzed-out guitar riff and tambourine accents; bass and drums drive a stripped-back rhythm. The chorus bursts open with call-and-response backing vocals and harmonized guitar lines. Short, energetic instrumental break before a brisk finale."

That's some "enhancement". 

Because I can't help myself, I tested the results on a few songs using the basic and enhanced prompts and the enhanced ones always come closer to what I was imagining. It's still true that the end result appears to have ignored a lot of the specifics but the overall effect is much closer to the intent. That's great in itself but the best part is that, now I know what the AI is looking for, I can write more detailed prompts myself.

Of course, once I'd started playing with the prompt enhancer, I started getting ideas. The first thing I thought of was to see just what the AI would make of that mysterious, seemingly meaningless prompt I wrote a whole post about a while back, Post-Post-Vibe Cassette. 

With enhancement

"Post-Post-Vibe Cassette, female vocals" 

becomes

"Dreamy lo-fi "post-post-vibe" cassette track: begins with washed-out, tape-warped synth pads and muffled drum machine groove. Minimalist bass pulses subtly beneath airy, melodic female vocals, layered and filtered. Interspersed analog samples weave throughout, adding texture."

Seriously? Where is it getting all that detail from? As far as I can tell, the original prompt is void of all context other than "female vocals". You can't be "post-post" a "vibe" unless you know what the vibe is and "cassette" isn't a genre. Cassettes don't even sound significantly different to other forms of recording (Except worse.) And the enhanced prompt admits that by putting the phrase in quotes. It doesn't recognize it as a genre. It just takes my word for it that it might be one.

I was curious to see if the prompt enhancer would always interpret post-post-vibe cassette the same way. It doesn't. It's not a fixed response. 

Here's another version it gave me:

"Atmospheric synths lay a hazy, lo-fi bed, layered with warbly cassette textures and saturated drums. Minimalist bass pulsates beneath crisp, delayed guitar plucks. Female vocals glide with dreamy reverb, while tape flutter and subtle noise accent shifting, moody sections."

That explains why the results vary when I just use the basic prompt. Even so, the long prompts are very similar. It's clearly not just making stuff up at random. I have no clue how it's landing where it does but it clearly thinks post-post-vibe cassette means something fairly specific, which explains why the variations always sit within a relatively limited range of styles and sounds. They're always a bit woozy and untethered, which is how the descriptions feel.

I had fun playing with that for a while and then I started to wonder... if the AI can interpret an abstract phrase as effectively as that, what could it do with something more concrete? Quite a lot, it turns out.

All the time I'd been using Suno I'd been seeding my prompts with musical genres and styles, specific instrumentation, production and arrangement and some mood descriptors like "sad" or "wistful". It never occurred to me to leave all that out and just tell the AI the context instead.

So I tried this:

"music heard coming from the local radio station of an orbiting space habitat over a vast conurbation in an anime set in the far future, female vocals" 

Suno didn't blink. It gave me two very nice interpretations, neither of which felt particularly "spacey" but either of which I could easily imagine hearing in the scenario described.

I have to say that one of the best things about the generative AIs is the way you can ask them to do all kinds of things that would be too embarassing to ask a person. They really don't care. Bearing that in mind, I asked Suno to enhance that description:

"Dreamy synth layers open the song, with glitchy electronic percussion and pulsing sub-bass creating a weightless atmosphere. Bright, airy pads ebb and flow, supporting spacious female vocals processed with subtle delay and reverb. Futuristic textures blend with downtempo beats, weaving ethereal melodies and brief robotized vocal fragments between verses and choruses. The arrangement gradually intensifies, introducing swirling arpeggios, metallic accents, and digital noise, cresting with a lush, enveloping bridge before gently dissolving into haunting echoes and cosmic ambience."

 I see the 200 character prompt limit is a thing of the past, then...

Obviously, I wanted to hear the results and I wasn't disappointed. They were excellent, although once again I can't quite see what's futuristic about them. But then, it's an impossible task, predicting the future, even in musical style.

Clearly I am nowhere near done with Suno yet. It's a fantastically entertaining toy. I have a bunch more ideas of things to try and projects to work on. I certainly feel more excited about playing around with it than I do about playing games these days.

In fact, I'm starting to wonder if generative AI might not end up becoming a new entertainment medium in its own right. It's clearly not great at facts but it's pretty good at making stuff up. Maybe that's what it'll end up being used for. I can imagine a whole raft of purpose-specific AIs being marketed and sold just as games are now. 

And on that happy thought I'll leave you. I have some songs to generate.