AIToday
Audio & SpeechThe Verge AIPublished: Jul 20, 2026, 04:00 JST3 min read

Suno AI music finally doesn't sound terrible—here's why

Suno AI music finally doesn't sound terrible—here's why

3 Key Points

  1. What happened

    Artist 1010Benja released a song called "Semiramis' Dream" on his latest EP using Suno generative AI, but in a more deliberate way than typical: he wrote lyrics, recorded his own vocals, collaborated with a producer on the beat, then iteratively fed elements into Suno, edited the output in Ableton, and fed it back into Suno hundreds of times.

  2. Why it matters

    The result is genuinely engaging—it avoids the "dead giveaways" that make most AI-generated music sound obviously synthetic. Unlike the typical Suno output, Benja masked the AI-generated parts (like the backup choir) by leaning into hyperpop's already-artificial vocal processing, showing that generative AI can work when paired with significant human artistry and intent rather than used as a shortcut.

  3. What to watch

    Benja is transparent about his process (Deezer marks the track as AI-generated) and candidly acknowledges the ethical baggage—but argues he uses AI tools out of financial necessity, not because they are essential to the sound. The real takeaway is whether his iterative, hands-on approach becomes a model for how generative AI could meaningfully serve artists, or remains an outlier.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Context & Analysis

The article hinges on a genuine surprise: music made with Suno AI that actually works. The writer, typically skeptical of generative AI in music, finds themselves enjoying "Semiramis' Dream" despite their stated bias against AI music—a contradiction that makes the piece worth attention. The reason for the success, however, is instructive and somewhat deflating to the AI-music narrative: Benja didn't let the tool drive the process. He started with intentional human work (lyrics, vocals, beat collaboration), then used Suno as one step in a multi-step production pipeline, iterating hundreds of times and manually editing in Ableton between cycles. The AI wasn't the primary creative engine; it was a tool deployed within a labor-intensive, deeply human creative process.

Benja's candor about the ethical weight of AI—he acknowledges the environmental damage and offers a wry, nihilistic shrug ("The good guy doesn't win this one")—adds texture. He frames his use of generative AI not as a principled choice but as economic necessity: he cannot afford session musicians. Yet the article also notes a crucial detail: sonically, little in "Semiramis' Dream" could not have been achieved without AI. The backup choir is convincing only because Benja deliberately masked it within hyperpop's own heavily processed aesthetic, a choice that neutralizes rather than exploits the AI tool. The takeaway is subtle but important: when generative AI works in the hands of a skilled artist with clear intent, it works not because the tool is powerful, but because the artist is doing the real work—writing, performing, and editing—and using the AI as one component among many.

FAQ
How did Benja actually make the song with Suno?
Benja wrote the lyrics and recorded his own vocals. He worked with collaborator Patrick Holder on the beat. Then he fed those elements into Suno, took what it generated, edited it in Ableton, added more layers, and fed it back into Suno—repeating this cycle hundreds of times to arrive at the final version.
Is Benja hiding that the song uses AI?
No. He told Hearing Things in an interview: "When you look in Deezer, it marks the song as AI-generated. We're not hiding it. We don't feel any shame about it."
Why did Benja use generative AI in the first place?
He is a struggling artist who has experienced homelessness and says: "The only reason I'm using gen at all is because my visions are currently far beyond my means." He couldn't afford to hire a professional choir, even though the AI-generated choir in the track isn't particularly convincing.

Get the latest Audio & Speech news every morning

For example, today's edition would include:

  • Google releases Lyria 3.5 music model in Gemini appTHE DECODER · 1d ago
  • Meta releases real-time audio model Muse Voice TranscribeTHE DECODER · 1d ago
  • AWS launches WhatsApp ordering assistant with AgentCoreAmazon AI Blog · 2d ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleConstellation invests in Blue Energy's prefab nuclear plants for AI data centers