From: redpointai
AI has fundamentally transformed the music industry, with Suno emerging as a prominent example due to its viral success and user engagement. With over 10 million users who have generated songs and a recent fundraise of $125 million, Suno’s CEO, Mikey Shulman, highlights the crucial role of community and user experience in its development and future vision [00:00:12].
User Engagement Models
Suno identifies two primary categories of users:
- Casual Users (Soundtracking Life): These users leverage Suno to narrate their daily lives musically, creating songs about happy, sad, funny, or memorable moments, such as a barista getting their name wrong or unexpected package deliveries [00:06:04]. Music serves as a way to tell stories, a practice humans have engaged in for a long time [00:06:25].
- Power Users (Creative Outlet): This group uses Suno as a significant creative outlet, enjoying both the process of making music and the final product [00:06:45]. They spend hours crafting songs to tell a specific story or actualize a sound in their head [00:07:05].
A key learning from power users is that many people possess great musical taste and ideas but lack the traditional means or complex tools (like Ableton or Protools) to execute them [00:07:26]. Suno aims to reimagine music creation processes through technological breakthroughs, making it accessible and enjoyable [00:07:52].
Overcoming the “Blank Canvas” Problem
Many AI products face a “blank canvas” problem, where users are unsure how to begin [00:08:14]. While Suno hasn’t fully solved this, they acknowledge it’s an area for future improvement [00:08:40]. Current approaches include:
- Suggestions in the text box: Providing a starting point for prompts [00:08:48].
- Themed experiences: Like a Valentine’s Day experience, where the reason to make a song is clear [00:09:03]. This helps users realize that reasons to make a song already exist, similar to sending a text or taking a picture [00:09:21].
- Future intuitive prompting: Moving beyond text to allow users to express themselves through mood, visuals, or sounds (e.g., humming a melody, tapping a beat, or turning everyday sounds into music) [00:10:01].
Early-stage AI tools are often text-driven, indicating significant room for growth in developing more intuitive user interaction methods [00:10:35].
The Social and Collaborative Future of Music
Suno emphasizes the social aspect of music, aiming to foster co-creation and shared experiences. This aligns with the idea that making music with others can be one of life’s most enjoyable moments [00:02:49].
- Sharing: Suno already enables users to share songs with small groups of friends [00:13:52].
- Multiplayer Creation: A major future focus is on “everything multiplayer,” allowing users to make music together synchronously (like a jam session) or asynchronously (sending half a song for completion) [00:14:31]. This envisions music as a fluid conversation, allowing users with varying skills to express ideas, react, and riff off each other [00:15:20].
- Interactive Concerts: Observing a Twitch streamer using Suno to perform an interactive digital concert to a large audience demonstrated the potential for group experiences [00:16:55]. The ability for viewers to micropay and interact with the streamer transforms a sterile digital concert into an engaging, communal event [00:17:15].
Community’s Role in Evaluation and Feedback
Unlike text models with objective reasoning benchmarks, music quality is highly subjective [00:20:08]. Suno’s approach to model evaluations incorporates:
- Automatic Metrics: For objective elements like audio quality, though these are often flawed [00:20:29].
- Human Evaluation: Relying on people who deeply love and enjoy music to make judgment calls, as “aesthetics matter” [00:20:44]. The ultimate test is how much users love the music produced and the level of control they have over it [00:21:11].
- User Feedback Loop: Suno benefits from a large and engaged user base who provide implicit feedback (model usage, choice between models) and explicit feedback through their Discord community [00:22:18]. This community helps identify issues, such as songs ending unreliably [00:23:14].
"The more you actually spend time looking at your data the better you're going to understand your model" [00:23:39].
Learnings from Discord Community Engagement
Initially, Suno anticipated a prolonged reliance on Discord, similar to other AI consumer apps like Midjourney [00:48:45]. However, upon launching a thin web app, 90% of usage shifted to the web within five days [00:49:10].
- Experience vs. Platform: Discord, while an amazing messaging platform, was not the right place for an “all-encompassing music experience” [00:49:26]. The entire user experience needs to be pleasant [00:49:28].
- Community Value: Despite the platform shift, the Discord community remains an invaluable resource for product development, providing critical feedback and insights, effectively “crowdsourcing prompt engineering” [00:49:56].
- User Pride: Users often edit song titles to include their names when their creations hit the trending page, demonstrating a strong sense of pride in what they’ve made [00:47:58]. This pride is something Suno aims to lean into, making it easy for users to share and feel good about their creations [00:48:06].
Broader Market Perspective
The market for AI music is seen as incredibly vast and Green Field [00:39:40]. Mikey Shulman believes that multiple companies will thrive, catering to different niches:
- Professional Artists: Tools for professional artists will emerge and grow, increasing the quality and quantity of music [00:40:21].
- Background Music: A huge business, as seen in the prevalence of background music in YouTube videos [00:40:37].
- General Consumers: Suno’s primary focus is on building experiences for the average person, expanding music’s importance and joy in their lives [00:40:51].
Regarding IP partnerships and artist imitation, Suno has intentionally avoided direct artist partnerships that allow users to create new songs “in the style of” specific artists without explicit consent [00:42:10]. Such viral moments are considered a “flash in the pan” that ultimately fade [00:43:38]. The focus remains on enabling users to create music relevant to them and enjoy the process of creation itself [00:43:27].
Hot Takes
- Open Source AI is Overhyped: While open-source software is amazing, open source AI is challenging due to the significant compute barriers and lack of explicit financial incentives required to produce state-of-the-art models [00:46:21].
- Music is Underhyped: Beyond AI, music as a fundamental part of people’s lives and its potential for greater impact is generally underhyped [00:47:07].