Head of People

Exclusive: AI music generator Suno opens SF office. An AI startup that lets users make music from prompts has opened an office in downtown San Francisco, Axios has learned. The big picture: Suno's expansion is another sign of the AI boom fueling downtown's comeback, and brings to the city a company at the center of recent fights over using copyrighted music to train AI models. Driving the news: Suno's new office, opened last week, is located in the South of Market, near Mission and 2nd. * Opening doors in San Francisco "will be critical as we continue to scale," co-founder and CTO Georg Kucsko told Axios via email. * The Cambridge-headquartered company launched in 2023 and is "investing especially heavily in our machine learning team to develop our frontier music model," according to Kucsko. * The startup is planning to grow headcount by 70% across engineering, product design, data science and machine learning teams, and expects to double its workforce this year. Between the lines: The company has been embroiled in legal challenges over its product. * The Recording Industry Association of America sued Suno in 2024 on behalf of three major record labels, including Warner Music Group (WMG). The lawsuit accused Suno of "wholesale theft" by allegedly using copyrighted recordings to train its model without permission. * Suno maintains that its platform is focused on creating new music and constitutes "fair use" under U.S. copyright law. * WMG settled with Suno in November. The deal includes requiring Suno to create new models trained on licensed content and allowing WMG artists to opt in or out of the program. Suno says it's meant to help users make new music and denies requests that reference specific artists or non-AI-generated music. * Users are encouraged to report copyright infringement if they come across a song that belongs to another artist or incorporates elements of someone else's music. What's next: Suno plans to launch the models developed with WMG and other industry partners this year. * More than two-thirds of its team members are musicians, and they regularly host writing camps with producers and artists, per a Suno spokesperson.

Suno turns to customization with v5.5. Suno launches V5.5 update for its AI music model. Artificial Intelligence (AI) has taken center stage in the music industry, and Suno is not being left behind. The firm has just unveiled the latest update to its AI music model - version 5.5. This update is seen as a major milestone in the development of Suno's AI platform as it comes with significant upgrades and new features. Previous updates focused mainly on improving fidelity and creating more natural voices. However, version 5.5 seeks to hand more control to the users, enhancing their interaction with the AI music model. This upgrade introduces three new features - Voices, My Taste, and Custom Templates. Let's dig deeper into what each of these features brings to the table. New feature: voices. In the release notes, Suno indicates that Voices has been the most requested feature. This feature empowers users to train the voice model using their own voices. Users can upload clear acapellas, finished music pieces with accompanying music, or even sing directly into the microphone of their phone or laptop. The cleaner and better quality the recording, the less data the system needs to train the voice model. To prevent unauthorized use of another person's voice, Suno requires the user to also say a verification phrase. This is a security measure, although there might be possibilities of fooling the system with existing AI models of celebrity voices. Once the Voices feature is trained, users can then create an AI version of themselves singing along to uploaded music or AI-generated output. New feature: Custom Templates. The Custom Templates feature is another key highlight of the v5.5 update. This feature enables users to train Suno to their own music. To use this feature, users need to upload at least six titles from their catalog and name the personalized template. They can then use it to guide version 5.5's responses to the prompts, thereby personalizing the outputs. New feature: My Taste. The third feature, My Taste, is designed to learn and adapt to the user's tastes and preferences over time. It keeps track of the genres, moods, and artists that users frequently interact with for prompts and applies them when the magic wand is used to automatically generate styles. These features are set to revolutionize how users interact with Suno's AI music model. However, it's worth noting that while My Taste will be available to all users, custom voices and templates will be exclusively available to Pro and Premier subscribers. This significant upgrade underscores Suno's commitment to enhancing user experience and transforming the music industry using AI. With these new features, users can look forward to a more personalized and sophisticated interaction with the AI music model. For more details on the new features and how to use them, visit the official Suno website or check out the v5.5 release notes. Stay tuned for more updates and news on Suno's AI music model.

Suno introduces MILO-1080 step sequencer as it expands its AI music tools. Suno is moving beyond prompt-based music creation, with a new 16-track step sequencer. Suno's latest release suggests AI music tools are continuing to shift beyond simple prompt-based creation. Its new feature, MILO-1080, short for "Model-Integrated Loop Orchestrator", focuses more on structured music production than instant results. At its core, MILO-1080 is a step sequencer. This type of tool is typically used by producers to build rhythms and patterns in a more controlled way. That alone signals a move towards users with some level of production knowledge, rather than those simply experimenting with text-to-music outputs. This direction reflects a broader change in how Suno is positioning itself. Earlier this month, Suno announced it had hit over 2 million paid subscribers, highlighting just how quickly the platform is growing. The tools on offer are evidently evolving alongside that rapid growth, giving users more detailed creative control. MILO-1080 is currently available in Suno's 'Labs' section as a preview. It is described as "a quirky, fun-first, mixed compositional, generative, procedural, sample friendly, MIDI enabled, 16-track step sequencer and synth designer". In practice, it allows users to generate sounds via prompts, reuse clips from previous projects, or create sounds manually using a built-in synth engine. This mix of AI assistance and manual input is becoming more common. It also aligns with Suno's earlier push into production tools, including its own DAW launch last year. The focus appears to be on keeping users more involved in the creative process, rather than removing them entirely. According to reporting from Music Ally, this shift is deliberate. Developer Kieron Donoghue said, "MILO-1080 is a step sequencer. That's not a tool for casual users... that's a tool for people with some music production experience." He added, "Suno is clearly saying, we're not just for beginners anymore." Suno's current direction has been developing over time. It first gained attention for its prompt-based music creation, but has since expanded with more detailed editing features, the acquisition of a browser-based DAW, and the launch of its own studio environment. Altogether, this shows a gradual shift towards offering a broader set of tools for making music, rather than focusing on a single feature. As Donoghue put it, "Each step brings them closer to being a full creative platform for musicians at every level." Beyond creation, there are also questions around discovery and data. Music Ally also reports that Donoghue has built a prototype tool called SunoCharts, which demonstrates how analytics could work if Suno introduces an API. It suggests a future where trends, genres, and even prompts could be tracked more clearly. However, there are potential barriers. Ongoing legal challenges around copyright may affect how far Suno can open up its data and tools. Donoghue pointed out the risks, saying, "Third parties might use Suno's data or capabilities in ways that create legal or brand risk... And once you open an API, you can't easily close it without it leaving a really bad taste in the ecosystem." Overall, MILO-1080 highlights how AI music platforms are continuing to develop more complex tools. While this may offer more control for some creators, it also raises ongoing questions about how these technologies fit into the wider music industry.