Human kind
On how precedented unknowns expose our true selves, "AI slop" is everything, and how I've been using Suno
Way back years ago, I’d often say “Please ask to clarify” in online communications — this was borne from observing that there was a human bias towards (knee)jerk reactions that assumed false outcomes that were off-color, off-base, and otherwise demonstrably incorrect. None of us can change human nature at scale, but we can practice individual improvement.
Sadly, the slide downwards has accelerated. I write this to share and also record where I was at right now (Friendly greetings, future self!): I’ve touched before on the hostility towards neurodivergent writing styles, and I’ve had the displeasure of getting posts accused of being AI-written when they were handcrafted with no LLM-assistance, and also the inverse.
My keening observations have intuited this: just as alcohol destroys inhibitions to bring forth a warped funhouse/hellhouse mirror of that person you loved, AI is an even more intoxicating drug when (ab)used unwisely. It’s exponentially even more powerful and addictive than alcohol, because get this: someone doesn’t even need to be using AI to be directly affected by it, they react by discarding their rationality to become a raging frothing monster vs. other humans. Even those who engage in some sort of pseudo-intellectualism which seems refined but is actually “talking past” and not actively listening to what’s going on. The inability to know what you don’t know and the abject failure to engage new ideas hobbles intelligence and compassion.
While from a scientific perspective, booze does NOT necessarily uncover hidden truths but ruins impulse control, the ambience of AI has (like some zombie plague) infected different humans in wildly different ways. It turns some of us into hyper-augmented quasi-deities with more agentic arms than a hypocritical octopus colony, it transforms others into “I don’t like what’s happening but I must acknowledge it’s going on, and I’ll use it selectively” (nuance, YES!), and on the lowmind of the scale — raging, rabid sloganeering against all things AI.
I’m concerned that it’s those reactions (and not just “AI itself) that’s divided humanity further. They put up further walls to things like cultural differences and their accompanying views on AI trust — recursive spoiler: China is a lot more supportive — which in and of itself is a fascinating exercise. How often I’ve seen obstinate stubbornness to empirical recommendations like “try the tools to find out what they can/can’t do for you”.
“Overton window” can refer to gradually-shifting allowance of what’s OK to discuss in broad company. It’s generally characterized as slow rate of change, but what blows my mind is how rapid the concept has been whiplashing as vocabulary (and the attached “meaning”) rapidly becomes obsolete. “AI slop” has only been a mainstream thing since mid-2024, and while it was initially applied to blatantly stupid and obviously-generated crap, it’s now sloppily (heh) splooged as shorthand for:
Any and all AI-generated content, whether or not I can tell the difference beforehand, regardless of its quality. ‘Cuz I still disagree on ethical principles, environmental impact, etc.
So in effect, what someone using it as a pejorative is really trying to tell you is that it’s not “AI slop”, it’s “AI = Slop”. Doesn’t matter if it’s a beautiful photo of a Polynesian spa that’s indistinguishable from “the real deal”, it’s not the output at this stage but the process that matters more… and who can tell without being told?
It’s unnecessarily depressing to wade into the YouTube comments cesspool, where guitar virtuoso Lexi Rose is getting flak for using AI to add drums. The negativity is righteously overcome by her being bold to share, and also those like Shannon Forrest, who uses musical AI to enhance his traditional hardware chops on the drums, mixing desk, and more. These stories are worth zooming into as we simultaneously zoom out on the timeline, and keep surfing on these sine waves.
Let me tell ya: I’ve been using Suno as a songwriting partner. It listens to my riffs and, in true call-and-response fashion, hollers back with fresh ideas. Aside from the quaint initial comedy, I was boringly impressed by how readily it could formulate perfect pop songs. I’d wiggle in prompts to try to break it to the extremes, exploding bizarre disembodied laughter and chortles — this was more interesting to me than the blandized, actual AI slop-pop it could churn out. And yet, I felt something was missing.
Still, my wife encouraged me to trudge on and revisit, and with version updates, it eventually got to a v5 state with cover songs where… holy moly… I can insert a piano piece I played (handcrafted, gazooks!) and it comes back with a full orchestration. It follows my melodies, it mostly gets my chords (except for some of the weird jazzprogsteppy 13th stuff, for now), and even though it wants to continues past the ending (which means artificially extending it), it’s giving me spicy firepower like breakbeat toplines, Enya-esque breathy vocals, and heck —
I’d adequately describe Suno covers as sampling vinyl from a parallel universe. Like that show Fringe’s shipping lanes, only there are heaps of record crates, and I’m digging thru all the alt-dimension versions to find something I can lift off of. (Which in turn reminds me of sampling crimes.) I might be surprised with aw-geez how this universe’s tilted towards film noir and I get a smoky sax, and aw-shucks how that uni apparently cares a lot more about gospel gasping. (The multiverse sure is fucking’ vast!) The timing’s not always there, which also throws me back to the early Ableton Live days and the warp marker’ing (which Suno is also copping) that had to be done. The noisy artifacts upon stem separation also remind me of trying to do this to early 128 kbps MP3, so it’s come full circle.
Sure, I get conflicted thinking about “What was this trained on?” and also “Does it meaningfully matter if the output doesn’t resemble that at all… and it’s rather similar to ALT-ME?” which is why it’s infinitely more intriguing to train it on my material, because I consented and made a conscious choice. This is the closest $10/month (or $5 for one if you threaten to cancel that first time) I’ve come to peering into portals and scrying with my little eye-ears, what would other Torleys be doing? Long before Rick and Morty had theirs, I had a “Torley Council” in my mind’s eye, and being able to manifest it brings me such joy, which is a human emotion.
We’d do best to remember in these turbulent times that we’re “only” human, after all (that’s a shoehorned Daft Punk reference for ya), and what sets us aside from machines is being able to fuck things up for ourselves because of frailty, insecurity, and those flaws that also make for consistently remarkable art despite being a terribly inconsistent person.


