[{"data":1,"prerenderedAt":204},["ShallowReactive",2],{"codex-\u002Fcodex\u002Fwhere-theres-a-right-answer":3,"codex-around-\u002Fcodex\u002Fwhere-theres-a-right-answer":192},{"id":4,"title":5,"body":6,"cover":158,"date":159,"description":160,"draft":161,"extension":162,"faq":163,"kind":176,"meta":177,"navigation":178,"noindex":161,"ogImage":158,"path":179,"readingMinutes":180,"seo":181,"seoDescription":182,"seoTitle":183,"standfirst":184,"stem":185,"tags":186,"updated":158,"__hash__":191},"codex\u002Fcodex\u002Fwhere-theres-a-right-answer.md","Where There's a Right Answer",{"type":7,"value":8,"toc":148},"minimark",[9,13,16,22,29,32,37,40,43,46,50,53,60,63,72,76,79,82,90,93,97,100,108,111,114,118,126,129,132,140,143],[10,11,12],"p",{},"HAUSVERBOT put a clip of Marcel Dettmann online in the same week Bandcamp banned\nAI-generated music from its platform. He is asked about the fear that AI will\nmake DJing irrelevant — that the robots will simply play the recommendations.",[10,14,15],{},"He is unbothered by that framing, and points somewhere else entirely:",[17,18,19],"blockquote",{},[10,20,21],{},"Optimizing stuff, it's the killer.",[23,24],"embedded-player",{"height":25,"provider":26,"src":27,"title":28},620,"Instagram","https:\u002F\u002Fwww.instagram.com\u002Fp\u002FDT0yzv0jeqc\u002Fembed","Marcel Dettmann on AI in music, HAUSVERBOT",[10,30,31],{},"His argument, paraphrasing: use it to be creative or to optimise whatever you\nalready did, and you get lost, because the possibilities are endless. You will\nnever be happy. You are looking for perfection. He is careful not to call it the\ndevil — it is a nice tool, and it is genuinely helpful here and there. Just not\nfor the creative part.",[33,34,36],"h2",{"id":35},"the-danger-is-not-the-machine","The danger is not the machine",[10,38,39],{},"The interesting move is that he declines the science-fiction version of the\nquestion. Nobody serious believes a model is about to take a Berghain booking.",[10,41,42],{},"What he describes is psychological and much more likely: a producer with\ninfinite versions available, optimising forever, converging on nothing. Before\nthese tools, the number of variations you could physically make was capped by\ntime. That cap was doing hidden work — it forced decisions. Remove it and\n\"finished\" stops being a state you arrive at and becomes a choice you have to\nmake against infinite unexplored alternatives.",[10,44,45],{},"Perfection was always unavailable. The difference is that it now looks reachable.",[33,47,49],{"id":48},"the-distinction-that-saves-the-tool","The distinction that saves the tool",[10,51,52],{},"Here is the rule I have settled on, and I think it resolves most of the\nargument.",[54,55,57],"codex-note",{"kicker":56},"The working rule",[10,58,59],{},"Use it where a correct answer exists. Never where one doesn't.",[10,61,62],{},"Some questions in music production have right answers, and they are checkable.\nIs there a resonant build-up at 240Hz that will make the mix muddy on a big\nsystem? Is the low end mono below 120? Is the master hitting its target loudness\nwithout pumping? Are two elements masking each other? These are measurements. A\ntool that finds them faster than your ears do is a better ruler, not a\nco-author, and refusing it is superstition.",[10,64,65,66,71],{},"Other questions have no right answer at all. Should this break at 3:20 or run\nanother ninety seconds? Should the hi-hat sit slightly wrong? Is this track\nabout something? Optimisation needs a target, and when there is no correct\nanswer, the only target available to a model is the average of everything it has\nalready been shown. Which is precisely how you get the frictionless, unskippable\nmusic I described in ",[67,68,70],"a",{"href":69},"\u002Fcodex\u002Fa-number-without-a-denominator","what the metrics measure","\n— pleasant, legible, and gone.",[33,73,75],{"id":74},"how-we-use-it-here","How we use it here",[10,77,78],{},"At Hybrid Flight the line is drawn exactly there, and it is not a moral\nposition, just a practical one.",[10,80,81],{},"AI-assisted tools are welcome for the technical layer. Spectrum analysis to see\nwhat my ears have stopped noticing at two in the morning. Mixdown checks. A\nsanity pass on the master. All of that is fair, and it makes the record better\nin ways that can be verified rather than argued about.",[10,83,84,85,89],{},"What those tools deliver is a ",[86,87,88],"strong",{},"technically correct foundation"," — and that is\nall they are permitted to deliver. A technically correct foundation is not a\ntrack. It is the floor the track stands on.",[10,91,92],{},"The one grey area worth naming honestly: using a generator or a random\nmodulation to produce a starting point is fine, and I have no problem with it.\nA generated pattern is raw material like a field recording or a happy accident\nwith a broken pedal. The rule is only that it never survives one to one.\nSomething generated is a suggestion, and taking a suggestion without changing it\nis how you end up releasing the average of a training set with your name on it.\nYour stamp goes on it or it does not go out.",[33,94,96],{"id":95},"the-other-half-of-the-argument","The other half of the argument",[10,98,99],{},"It would be easy to leave this as a defensive piece, so it is worth reading it\nagainst something less cautious.",[10,101,102,103,107],{},"Matt Bluemink's ",[104,105,106],"em",{},"The Future is Not Lost: On Music, Technology, and the Creation\nof New Worlds"," (Becoming Press, 2026) argues more or less the opposite mood.\nIt's a set of essays on music, technology, architecture and philosophy, drawing\non SOPHIE, Arca and Iglooghost, and its case is that the future has not\nevaporated — that it still gets imagined, and that music and technology are\nwhere that imagining happens.",[10,109,110],{},"That is not incompatible with Dettmann. SOPHIE's work was radically\ntechnological and sounded like nothing else, which is the whole point: the\ntechnology was in service of a decision nobody else would have made. The failure\nmode is not using new tools. It is letting the tool make the choice, because\ntools trained on the past can only ever propose the past, slightly rearranged.",[10,112,113],{},"Dettmann is warning about surrender. Bluemink is describing what happens when\nthere is no surrender. Both can be true, and if you only read the pessimistic\nhalf you will end up making conservative records for good reasons, which is\nstill making conservative records.",[33,115,117],{"id":116},"what-bandcamps-ban-actually-says","What Bandcamp's ban actually says",[10,119,120,121,125],{},"The context for the clip: in January 2026 Bandcamp became the first major music\nplatform to ban AI-generated music outright — nothing generated wholly or in\nsubstantial part by AI, no using it to imitate specific artists, and no scraping\nits catalogue to train models. Their reasoning was that listeners should be able\nto trust that what they find there was made by people. Subvert, the co-op I\nwrote about in ",[67,122,124],{"href":123},"\u002Fcodex\u002Fthe-equity-nobody-shared","the equity nobody shared",", took\nthe same position from the start.",[10,127,128],{},"A ban is a blunt instrument and enforcement will be genuinely difficult — the\nline between a mastering assistant and a generative model is not obvious from\nthe outside, and a determined uploader will not be stopped by a policy. But the\nsignal matters more than the enforcement. It is a marketplace stating that the\nthing it sells is human authorship, at a moment when the largest platforms are\nquietly discovering that they can sell something cheaper.",[10,130,131],{},"Which returns to the only rule I would hand anybody. Let the machine measure.\nDo not let it decide.",[133,134],"codex-cta",{"body":135,"href":136,"kicker":137,"label":138,"title":139},"Five tracks between Andalusia, Ibiza and Berlin — measured with everything available, decided by one person.","https:\u002F\u002Fhybridflight.bandcamp.com\u002Falbum\u002Fbetween-coasts-ep","The work","Buy on Bandcamp","Between Coasts EP",[141,142],"hr",{},[10,144,145],{},[104,146,147],{},"Source: Marcel Dettmann in conversation with HAUSVERBOT, clip posted 22 January\n2026; the full interview is on their YouTube channel. The quoted line follows\nthat clip. Bandcamp policy details are from its January 2026 guideline update as\nreported at the time.",{"title":149,"searchDepth":150,"depth":150,"links":151},"",3,[152,154,155,156,157],{"id":35,"depth":153,"text":36},2,{"id":48,"depth":153,"text":49},{"id":74,"depth":153,"text":75},{"id":95,"depth":153,"text":96},{"id":116,"depth":153,"text":117},null,"2026-08-31T23:00:00","Marcel Dettmann isn't worried about robots taking the booth. He's worried about optimisation. A working rule for where AI belongs in music and where it does not.",false,"md",[164,167,170,173],{"question":165,"answer":166},"Will AI make DJs irrelevant?","Marcel Dettmann's view, and the one taken here, is that this is the wrong thing to fear. The realistic risk is not a machine replacing the DJ but a producer disappearing into endless optimisation — chasing a perfection that does not exist because the number of possible versions is now effectively unlimited.",{"question":168,"answer":169},"Is it acceptable to use AI in music production?","For technical work with a measurable correct answer — spectrum analysis, gain staging, mixdown checks, mastering targets — it is a genuinely useful instrument. For decisions about what a track should feel like, where it should break, and what it is about, there is no correct answer to optimise toward, so a model can only pull the result toward the average of what it has already heard.",{"question":171,"answer":172},"Can you use AI-generated material as a starting point?","Yes, provided it stays a starting point. A generated pattern or a random modulation is raw material like any other. The rule that matters is never taking it one to one — the decisions about what to keep, cut, break and leave wrong are where the authorship lives.",{"question":174,"answer":175},"Why did Bandcamp ban AI-generated music?","In January 2026 Bandcamp updated its guidelines to prohibit music generated wholly or in substantial part by AI, along with using AI to imitate specific artists, and barred its catalogue from being scraped for model training. It was the first major music platform to do so, on the stated basis that listeners should be able to trust that what they find was made by humans.","Essay",{},true,"\u002Fcodex\u002Fwhere-theres-a-right-answer",5,{"title":5,"description":160},"Marcel Dettmann on optimisation as the real danger of AI in music. Why spectrum analysis, mixdown and mastering are fair game, why creative decisions are not, and what Bandcamp's ban actually signals.","AI in music production — where it helps and where it ruins the work","Use it where a correct answer exists. Never where one doesn't. Almost every argument about AI in music collapses into that one distinction.","codex\u002Fwhere-theres-a-right-answer",[187,188,189,190],"ai","music production","techno","marcel dettmann","YNtUhHwwTn8ApyWPwEZC-uY8QD_feKrS3-Af8cS4s0E",[193,199],{"title":194,"path":195,"stem":196,"description":197,"kind":198,"children":-1},"Two Brushstrokes","\u002Fcodex\u002Ftwo-brushstrokes","codex\u002Ftwo-brushstrokes","A DJ posting as @supafrisch describes a scene that presents itself through thirty-second transitions and new mixers, because almost nothing that actually happens in a club can be posted. On clips as advertising, and on what the advertising is supposed to be for.","Response",{"title":200,"path":201,"stem":202,"description":203,"kind":176,"children":-1},"Why a Room Dances Together","\u002Fcodex\u002Fwhy-a-room-dances-together","codex\u002Fwhy-a-room-dances-together","Several hundred people moving on the same beat, none of whom arranged it. What actually happens when a floor locks in — beat entrainment, raised pain thresholds, and why medium syncopation makes a body move.",1789031394040]