Here’s my latest music video. The song was created using Suno and remixed with Reason. The great thing about Suno is the ability to mix genres: you start with an electro/trance thing and then introduce harmonica, as I did here for example. It might take a couple of dozen attempts before you find something that works but it’s so much fun along the way.
The video was built with clips generated by Hailuo, Genmo and Luma and you can hopefully see how improved the details are as I zoom in on the eyes, for instance. Syncing the music to the video is probably the hardest part, but it’s the best way to learn. Content is free now folks, so get to it!
I’ve been having so much fun in the last few months I thought I’d better share it with you. I must’ve dabbled, downloaded or signed up to over fifty different AI programs in order to get some sort of a handle on the revolution that is currently engulfing is. So, this is a kind of beginner’s guide to the creative AI tools which really can build stunningly beautiful, utterly believable, mind blowing content out of thin air…for (almost) zero cost.
I’ll include all the links so that you can play with them yourselves, because, frankly, if you have the tiniest creative bone in your body you will be dying to have a go. Genuinely, I’ve been waking up like an excited nine year old desperate to get back to the AI world.
Portraiture
Right now the top two contenders for creating photo realistic images of people that never existed are Leonardo and Midjourney. There are a thousand others but these two seem to be battling it out for the most lifelike detail, and they’re both improving week on week. The more detailed the prompts the better, although, like most artists, they’ll both completely ignore you at times. Specify the lighting (dappled or light leak for example), the blurriness of the background (or not) the texture and the depth of colour and you cannot be disappointed. You can even upload a reference image and specify the percentage of artistic license you want it to have. And now, with the new ‘consistency’ feature you can create a character and store it in its memory so that you can come back to it later. This means that you’re rugged cowboy on a ranch, looking pensively into the sunset can also be pictured dressed as a giant tortoise pushing a trolley around a supermarket… if that’s what floats your boat. Of course, both these ingenious platforms will also produce any illustration style you can imagine, so the possibilities are endless.
Illustration
For instant creative illustration with terrific fantasy interpretation skills Dalle or Bing (which uses Dalle) are hard to beat. Looking for a 1940s style pen and ink drawing of King Kong smoking a cigarette atop the Empire State Building? This is your platform. Other contenders, equally capable, are Ideogram, Tengr and Openart.
Photoshopping
Funny how things pan out. Adobe’s Creative Suite (which includes Photoshop) has recently become prohibitively expensive for anyone other than a full time professional. If you’re a dabbler or just like to enhance your personal photos occasionally, like Kate Middleton for example, thirty quid a month is a bit rich. On top of that, Adobe has sparked some controversy recently with new terms and conditions that give it access to your private images. Seriously worrying. But fret no longer because as Adobe slowly commits suicide an army of AI photo-fiddling substitutes is coming over the hill to the rescue. There’s Vivid, Bria RMBG (simple background removal) and fun programs like Akool for face swapping, or PhotoDirector and YouCam Perfect, to name a handful I’ve tried. It really depends on what you use Photoshop for, but surely the monthly subscription to Adobe’s Creative Suite must’ve taken an almighty hit.
Music
Oh boy, I’ve had so much fun making music I can’t tell you. I’ve ‘written’ songs in Suno for a couple of my talks (ok, you can roll your eyes) including two Ibiza style dance trance tracks for a conference in, yes, Ibiza. In a slightly less embarrassing fashion I’ve created an album of AA Milne poems set to music of various genres. You can find them here. There’s a dancy, chill-wave one, an Andrew Lloyd Webbery style thing, a Gen Zeddish moody ballad and a couple of heavy delta blues numbers. I think they’re all amazing, of course. But that’s because, whether AI did it or not, none of these songs would exist had I not conceived of them. That’s art as far as I’m concerned. Suno’s big competitor is currently Udio which I’ve tried but found a tad limiting, but there are many more to choose from including Soundraw, Beatoven and Soundful, each of which will be great for specific genres as well as having different price structures for various levels of access. Suno cost me just short of seventy quid for a year’s subscription btw, so it really isn’t an expensive hobby. And the latest feature allows you to upload a sound or music loop so that you really can control the output. Serious fun.
Clones and Avatars
I’m currently working on creating a clone of myself. Sounds horrific, right? Well, it’s only a bit of fun. With a platform called Delphi you feed it everything that exists of you online: every article, blog or video, so that eventually it will know precisely how you think and how you tend to phrase things. Combine this with a lip-syncing avatar and you have a mini-me (or you) that can answer questions vocally, on your website, for example. I’m halfway there but I don’t think Delphi is quite as ready to roll as it pretends it is. We’ll surely get there though. Of that I have no doubt. In the meantime both Hallo and Hedra can create convincing, speaking avatars from any image you upload.
Text to Video
This is the Holy Grail for AI enthusiasts at the moment. Type in what you want to see and AI will deliver a video almost instantly. You can see the Sora showcase here and it is truly mind blowing. It hints at a future where we will create our own ‘movies’ with friends each contributing to the storyline, adding characters and dialogue as we explore the universe that opens up before us. This will happen, but for now Sora, Pika and Kling are battling it out behind the scenes with only a privileged few being allowed in on the fun. In the meantime, Dream Machine by Lumalabsis available, free and pretty damn impressive.
Truth is, things are moving so fast that in a couple of weeks I’ll need to update all this info, but the point is that right now, today, is the perfect time to start playing around with these incredible tools as they develop and mature. After all, if AI will ultimately rule the world and turn us all into slaves we may as well have some creative fun with it in the meantime.
Howard Saunders is a writer, speaker, Retail Futurist(and AI enthusiast)
