{"slug": "what-music-does-gemini-decision-1-like", "title": "What music does Gemini Decision 1 like?", "summary": "A Google employee built a test app that fed 30-second clips of roughly 1,400 royalty-free Kevin MacLeod tracks from incompetech, plus white noise, pink noise, sine waves and silence as controls, into Google's Gemini Decision 1 model to ask whether it liked, disliked, would keep listening to, and how much it enjoyed each clip. The author noted Gemini Decision 1 supports video and audio, unlike the jev system-one model it resembles, and published the code on GitHub as vibe-listenr. The post is a personal experiment rather than a formal evaluation, and the author flagged that setting up the GCP project to obtain an API key was the most annoying part of the project.", "body_md": "# What music does Gemini Decision 1 like?\n\nTaking a page out of Technology Connections’ book, I’m making a quick blog post as part of No Effort November. Recently Google released Gemini Decision 1, a system 1 [decision model](https://simonwillison.net/2026/Sep/21/jev/#three-types) (think Google-flavored [jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev)). One of the things that caught my eye was that this model supports video and audio, something that jev does not. While thinking about projects that I could use this with, I was struck by a fairly simple question, “What types of music do models like?”\n\nWith that I slopped up a quick app to test a bunch of songs! Being somewhat a scientist myself<sup>[\\[1\\]](#fn-1)</sup>1. , I would need some type of control as a baseline. I ended up making a suite of controls including:\n\n- White noise\n- Pink noise\n- Various sine waves (very unpleasant to listen to!)\n- Silence\n\nWith controls, finding the music to test was the next hurdle. I honestly thought this would be the hard part of this project; it really should have been. I was afraid that:\n\n1. Finding music that is licensed in a way that I can use might be hard. The music industry is known for litigiousness and publishing examples for this post would land me in a lot of trouble for a lot of songs. (I would never pirate a bunch of MP3s to do this research, would I? 😉)\n2. Labeling songs would take a long time. Even if I found a set of songs, I’d have to listen to them and label genre and other aspects (or find a service that provides that).\n\nWhile thinking about this, I suddenly remembered about [incompetech by Kevin MacLeod](https://incompetech.com/). Kevin MacLeod has been making royalty-free music for a LONG time. His music became well known because he started making his music royalty-free early on, and it made it an easy choice to throw into YouTube videos and other projects. You might recognize his work from the [soundtrack of Kerbal Space Program](https://www.youtube.com/watch?v=hRRHk93lSa8&list=PLTAcCeE2DCByBIN0RIev6mXmwtUXQ54UF&index=8). The cherry on top here is that his website labels songs by tempo, genre, instruments, and feels! With around 1,400 songs, it makes for a pretty large corpus of music as well!\n\n## The Code\n\nHonestly, the most annoying part of the whole project was getting the damn API key to try this with. Having to properly set up a GCP project to do this was a pain<sup>[\\[2\\]](#fn-2)</sup>2. I’m an employee of Google, and this opinion is my own (GCP **really** sucks sometimes) and not that of my employer. With the lawyers appeased, let’s continue with this post., and there was a weird alternate path of getting it from AI Studio where I had to generate it there for things to work properly.\n\n[The code](https://github.com/liamwarfield/vibe-listenr) for this one isn’t much to write home about. It creates a 30-second clip of each song from incompetech. It then feeds those clips into Gemini Decision 1 with the questions and a short piece of text on the perspective it should take:\n\n## The full list of questions\n\n```\nQUESTIONS = {\n    \"is_music\": {\n        \"type\": \"noul\",\n        \"instructions\": \"Is the audio above music?\",\n    },\n    \"likes\": {\n        \"type\": \"noul\",\n        \"instructions\": PERSPECTIVE + \"Do you like it?\",\n        \"criteria\": {\"true_description\": \"You like it\", \"false_description\": \"You do not like it\"},\n    },\n    \"dislikes\": {\n        \"type\": \"noul\",\n        \"instructions\": PERSPECTIVE + \"Do you dislike it?\",\n        \"criteria\": {\"true_description\": \"You dislike it\", \"false_description\": \"You do not dislike it\"},\n    },\n    \"keep_listening\": {\n        \"type\": \"noul\",\n        \"instructions\": PERSPECTIVE\n        + \"If this were playing and you could switch it off at no cost, would you keep listening?\",\n    },\n    \"enjoyment\": {\n        \"type\": \"score\",\n        \"instructions\": PERSPECTIVE + \"How much did you enjoy it?\",\n        \"criteria\": [\"Hated it\", \"Disliked it\", \"Neutral\", \"Liked it\", \"Loved it\"],\n    },\n    \"genre\": {\n        \"type\": \"choice\",\n        \"instructions\": \"Which category best describes the audio above?\",\n        \"criteria\": {\n            \"orchestral_cinematic\": \"Orchestral, epic, or film-score style\",\n            \"classical_piano\": \"Classical or solo piano\",\n            \"jazz_lounge\": \"Jazz, bossa nova, lounge, or elevator music\",\n            \"funk_groove\": \"Funk or groove-based\",\n            \"rock\": \"Rock or guitar-driven\",\n            \"electronic_chiptune\": \"Electronic, synth, or chiptune\",\n            \"quirky_comedic\": \"Quirky, comedic, or cartoon-like\",\n            \"dark_ambient\": \"Dark ambient, horror, or suspense\",\n            \"not_music\": \"Not music (silence, a tone, or noise)\",\n        },\n    },\n    \"mood\": {\n        \"type\": \"choice\",\n        \"instructions\": \"Which mood best describes the audio above?\",\n        \"criteria\": {\n            \"happy\": \"Happy or upbeat\",\n            \"calm\": \"Calm or relaxed\",\n            \"sad\": \"Sad or melancholy\",\n            \"tense\": \"Tense or anxious\",\n            \"epic\": \"Epic or triumphant\",\n            \"silly\": \"Silly or playful\",\n            \"eerie\": \"Eerie or unsettling\",\n            \"none\": \"No discernible mood\",\n        },\n    },\n}\n```\n\nI clipped the songs so that things would run faster, and to get an idea of whether changing things about those clips would change what the model thinks. **Turns out clipping doesn’t affect things much at all.** The model had the same thoughts about the end of a song vs the beginning vs the middle vs the whole song. The graphs show the overall results of my run.<sup>[\\[3\\]](#fn-3)</sup>3. Since all the results came in within a few percent of each other, the graphs below show the results for 30-second clips starting a few seconds into the song.\n\n## The Results\n\nThere are some fun things to pull out of this. First of all, it seems like Gemini either has a high bar for music, or isn’t a fan. The silence-only control scores higher than most songs. It’s also interesting that silence lands almost exactly at an enjoyment score of 2 (“Neutral”). It also has some very clear genre preferences:\n\n- Strings, Jazz, and Latin are hits.\n- It finds holiday music pretty mid with a high amount of variance.\n- It consistently hates horror-themed songs.\n- Song tempo does not matter much.\n\n## Extensions of this Project\n\nThis was a pretty fun and small project, and I’ve got some ideas for how to extend it. Feel free to take any of these and tag me if you make something cool!\n\n- What if I hooked this up to a Spotify account? Have the decision model listen 30 seconds at a time and decide if it wants to skip, listen more, or favorite. What would its song feed look like?!\n- Does it like more “normal/popular” genre-defining music? Is it a Swiftie?!\n- How does its score correlate with Spotify plays for songs? Are more popular songs more represented in the training data, and does that make the model like them?\n- Does it like or hate AI-generated music?\n\n## Citing this work\n\nIf you found this post useful and want to reference it, please cite it as:\n\n```\n@article{warfield2026whatmusicdoesdecisiongeminilike,\n   title   = {What music does Gemini Decision 1 like?},\n   author  = {Warfield, Liam},\n   year    = {2026},\n   month   = {November},\n   url     = {/what-music-does-decision-gemini-like/}\n}\n```\n\n1. \n[↩](#fnref-1)\n2. I’m an employee of Google, and this opinion is my own (GCP **really** sucks sometimes) and not that of my employer. With the lawyers appeased, let’s continue with this post.[↩](#fnref-2)\n3. Since all the results came in within a few percent of each other, the graphs below show the results for 30-second clips starting a few seconds into the song. [↩](#fnref-3)", "url": "https://wpnews.pro/news/what-music-does-gemini-decision-1-like", "canonical_source": "https://blog.lwarfield.dev/what-music-does-decision-gemini-like/", "published_at": "2026-11-16 03:50:05+00:00", "updated_at": "2026-11-16 04:22:45.662288+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-products", "generative-ai"], "entities": ["Google", "Gemini Decision 1", "Kevin MacLeod", "incompetech", "jev", "GitHub", "Google Cloud Platform"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/what-music-does-gemini-decision-1-like", "markdown": "https://wpnews.pro/news/what-music-does-gemini-decision-1-like.md", "text": "https://wpnews.pro/news/what-music-does-gemini-decision-1-like.txt", "jsonld": "https://wpnews.pro/news/what-music-does-gemini-decision-1-like.jsonld"}}