For this Feedback Loop feature, I spoke to Tom Hull of Hull Productions about what it means to be an up-and-coming artist today. Working across production and mixing, Tom often sees artists at the early stage, when songs are still being shaped, identities are still forming, and the pressure to release, promote and stand out is already there.

A lot of the conversation came back to the same problem. As Tom put it, “it’s probably easier than ever to put music out, but harder than ever to actually get people to care.” Artists can record at home, release independently and promote themselves through TikTok, Instagram and Spotify without waiting for a label. But because everyone can do that, even strong songs can get buried quickly.

Tom also pointed out that newer artists are not just expected to write and release music anymore. “They’re writing the songs, booking rehearsals, paying for recording, sorting artwork, posting content, trying to get gigs, trying to get playlisted. It’s a lot.” That pressure can be difficult for artists who are strong writers or performers, but do not naturally think like content creators. This is where Feedback Loop’s purpose becomes clearer. Algorithms can recommend music, but they do not always explain why an artist is worth listening to. As Tom said, “algorithms don’t really explain anything. They just throw songs at you.” A playlist can place one track next to another, but it rarely gives the listener a deeper reason to care.

That is why Feedback Loop uses bigger reference points as a route into smaller artists. If someone likes Fontaines D.C., Wolf Alice or Oasis, that can become a route into a newer act. Tom saw the value in this, explaining that “if you just give someone a random band name, they might not bother. But if you say, ‘If you like Fontaines D.C. or Wolf Alice, you might like this,’ then suddenly there’s a reason to click.” From the production side, Tom said the most exciting artists are usually the ones with a clear identity, even if the recording is still rough. “Some artists come in and they’re still chasing a sound they think they should have,” he said. “Others come in with a rough demo and, even if it’s messy, there’s already a clear feeling there.”

For Tom, polish is not always the answer. Sometimes a strained vocal, a rough guitar part or an imperfect demo carries the character of the track. “The job isn’t always to clean everything up,” he explained. “Sometimes it’s knowing what to leave alone.”

That point also connects to AI. Tom wasn’t completely against AI, especially when it is used for admin, ideas or helping artists get unstuck. “I’m not fully against AI,” he said. “I think there are ways it could be useful, especially for admin or idea generation.” The concern is when AI starts replacing the creative choices that make music feel personal. In April 2026, Deezer reported that AI-generated tracks made up 44% of new music uploaded to its platform, with around 75,000 AI-generated tracks delivered each day. For up-and-coming artists, that adds another layer to an already overcrowded space. Tom added that for smaller artists, it can feel like they are now competing with music “that can be generated in seconds.”

The issue is not whether AI music sounds good or bad, it’s about what it does to the wider ecosystem. If streaming platforms are flooded with generated songs, human artists have to fight even harder for attention. As Tom said, “people still care about human artists, but they need to know what they are listening to.”

For Tom, authenticity comes down to choice. Music has always used technology, from guitar pedals to Auto-Tune, but the difference is whether those tools are helping an artist shape an idea or starting to replace the idea itself. “For me, authenticity is about the artist making choices,” he said. “Even if they use technology, there still needs to be a person behind it deciding what feels right.”

He also raised concerns about copyright and training data. Artists spend years building a sound, writing songs and developing a style. If AI systems are trained on creative work without permission, payment or credit, then the technology benefits from the labour of artists without properly recognising them. As Tom put it, “artists spend years building a style, writing songs and developing a sound, and then technology comes along that can learn from all of that without necessarily paying or crediting them. That doesn’t sit right with me.”

Tom’s advice for newer artists was simple: make something that actually sounds like you. “Don’t just chase what’s working on TikTok or what another band is doing,” he said. “References are useful, but there needs to be something personal in it.” At the same time, artists still need to think carefully about the story, artwork, live presence and audience around a release. That balance is difficult. New artists need to be strategic, but not so strategic that the music starts to feel soulless. Tom summed this up well: “You need to care about how the song reaches people, but you also need to protect the thing that made it worth releasing.”

For Feedback Loop, this conversation reinforced why human recommendation still matters. Not every good artist will go viral, and not every strong song fits neatly into an algorithm. As Tom said, “Sometimes it just needs a person saying, ‘This is worth your time.’”

Technology will keep changing how music is made and discovered, but it cannot replace actual taste. In a crowded scene shaped by algorithms, content pressure and AI-generated music, human context matters more than ever.