Nov 13, 2024 · 47m · big-technology
Spotify Co-President Gustav Söderström on their future with Generative AI
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Alex Kantrowitz interviews Spotify Co-President Gustav Söderström about how generative artificial intelligence, multi-format media integration, and evolving recommendation architectures are reshaping audio streaming while preserving human artist value and listener agency.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 31.6% of the talking time here. How this is scored →
Alex as informed peer 5.1 Guest teaching 5.5 Guest disagreement 2.1 Alex pushing back 4.1
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 25:27 Pushing back against AI DJ failure narrative
Gustav directly rejects Alex's claim that listeners are abandoning the AI DJ, asserting that hard platform numbers prove it is one of their most successful and heavily used features.
Hardest push from Alex ▶ 7:59 Demanding clarity on AI slop and Shrimp JesusAlex refuses to let Gustav dodge the question of platform contamination, invoking Meta's viral Shrimp Jesus to demand whether Spotify welcomes pure prompted AI music.
Biggest teaching moment ▶ 36:50 Deep dive into generative recommendation architecturesGustav explains how recommendation engines hit a performance ceiling with traditional deep learning, and how Spotify is re-architecting them as generative sequence token models that scale like LLMs.
Alex holds their own ▶ 13:28 Challenging the human-artist moatAlex articulates a well-reasoned counter-thesis to Gustav's romantic view of human musicians, arguing that rapid AI improvements are increasingly disproving the necessity of human connection in music.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome and Overview with Gustav Söderström at Spotify Headquarters | 3 | 5 | 1 | 3 | Alex opens with an engaging premise about AI music tools like Suno, asking if they represent a threat or opportunity. Gustav reframes the conversation with a comprehensive historical timeline of musical technology from Bach to DAWs and Avicii. | |
| Platform Ethics, Creator Compensation, and Copyright Frameworks | 5 | 4 | 2 | 6 | Alex presses Gustav on whether fully AI-generated tracks are welcomed on Spotify, citing Meta's proliferation of 'Shrimp Jesus' engagement bait. Gustav draws a clear line between hosting legal creator tools and Spotify generating cheap in-house content. | |
| Ambient Music Utility Versus Human Artist Connection | 6 | 5 | 3 | 6 | Alex challenges Gustav's claim that human connection will always safeguard human musicians like Taylor Swift, arguing that AI quality leaps may shatter that assumption. Gustav responds by arguing economic scarcity will make authentic human artists even more valuable. | |
| NotebookLM, Conversational AI Audio, and Long-Tail Content Economics | 6 | 5 | 1 | 2 | Alex details his firsthand workflow using NotebookLM for interview prep and cites industry frameworks around long-tail content. Gustav enriches the discussion by applying economic principles of zero marginal production costs and complementary goods. | |
| Algorithmic Recommendations and the Evolution of Spotify's AI DJ | 4 | 6 | 4 | 5 | Alex bluntly challenges the success of Spotify's AI DJ, citing anecdotal user drop-off among his peers. Gustav immediately counters with hard platform metrics, explaining why it is among their highest retention listening sets. | |
| Balancing Algorithmic Automation with Active User Agency | 6 | 6 | 2 | 5 | Alex cites Kyle Chayka's New Yorker critique regarding algorithmic force-feeding and the loss of user choice. Gustav acknowledges the tension, explaining that moving to zero-friction algorithmic delivery destroys valuable user feedback signals. | |
| Overcoming Filter Bubbles and Decoding Social Listening Habits | 6 | 7 | 3 | 3 | Alex questions algorithmic flattening and individualization using external commentary from Chayka and Fontana. Gustav educates Alex on modern generative recommendation architectures that scale like LLMs, and presents internal data showing social listening remains massive. | |
| Multi-Format Audio Strategy: Podcasts, Audiobooks, and Discovery | 5 | 6 | 1 | 3 | Alex asks about the business logic of bundling audiobooks and podcasts into one application and touches on podcast discoverability problems. Gustav explains CAC dynamics in mobile ecosystems and why multi-format super-app architectures beat standalone apps. |
Statements from this episode (22)
Söderström: Many major artists are already using AI in song production
“But the truth is that much of music being made per day made today is a combination. I think many of the big artists are using AI for parts of their songs or parts of the track or the drums, et cetera.”
Söderström: It Will Become Very Difficult to Define an AI Song
“So I think there's actually a scale between zero AI and a hundred percent AI. And I think we're on this progression where it's actually going to be very difficult to say what is an AI song.”
Söderström: Generative AI music is currently in an early P2P-like phase
“Before Spotify, the technology sort of preceded the business model. It was great for consumers. They could now get all of this music for free, but it didn't work for, Creators. And I think we're in the same period of time now where the technology has preceded …”
Söderström: Spotify commits to not generating synthetic AI music in-house
“Should we generate all the music ourselves? And that's where we're saying, no, we're not going to generate that music and other platforms maybe will because it's cheap content, right? So that's the key difference of we decided what we want to be in this world.…”
Spotify uses detection systems to take down infringing derivative works
“We have detection systems for if you are if it's a derivative of work of something that already exists. So we have systems to take these down.”
Söderström: AI-generated music will proliferate in functional and ambient use cases
“For certain things, maybe you could create better white noise, maybe you could create better you know, always varying ambient music for your studying, maybe for gaming, maybe that music should automatically adjust what's happening on the screen. So I think we'…”
Söderström: AI will not replace human artists like Taylor Swift
“I do think the human need for having someone to believe in an actual artist that you care about, I don't think Taylor Swift will be Replace by an AI, not because the music couldn't sound similar, but because the whole point is Taylor Swift and belonging to som…”
Söderström: Human connection will become more valuable in an LLM-dominated world
“What tends to happen in these worlds is that the thing that is scarce gets even more valuable. So one bet would be that true human connection gets more valuable than ever. When a lot of what you talk to in the future may be LLMs.”
Söderström: NotebookLM's breakthrough was generating dialogues instead of monologues
“What I think was the great innovation of Notebook LM was that people generated monologues and what humans really respond to are dialogues.”
Söderström: AI will cause an explosion in niche podcast catalogs
“We're going to have enormous amounts of content around niches where it didn't make sense to produce a podcast. So one way to think about it is just like the cost went to zero. So I do think that the catalog is going to explode.”
Söderström: Superstars will grow bigger despite an explosion of niche AI content
“I also think you're going to see the same thing as we see in music. The superstars will actually also get bigger.”
Söderström: LLMs will enable literal relationships between consumers and brands
“And I think what is going to happen with these LLMs is at least for some brands, you will start having literal relationships with them.”
Söderström: Spotify's AI DJ usage exceeds Discover Weekly among active users
“Well, for the people that use it, it's actually their biggest set. It's bigger than their Discover Weekly usage.”
Söderström: LLM-generated music storytelling significantly boosts Spotify app retention
“So what we've done since then is we've invested quite a lot in this is quite recent that is rolling out in LLMs that actually tell interesting stories about the music. And we see very strong effects on this, on the retention of the application.”
Söderström: Spotify is re-emphasizing manual user playlisting
“So we're actually reemphasizing playlisting quite a lot.”
Söderström: Eliminating user effort harms apps by destroying vital feedback signals
“Going to zero user investment seems good in the short term, but I don't think it's good in the long term because you actually lose signal from that user. And at the end, I think they feel less participatory in the experience.”
Söderström: Purely data-driven product design produces weird applications
“If you only treat it as statistics, the application is going to be very weird at the end of the day. So you have to combine some sort of vision and conviction, but you have to be still very data driven.”
Söderström: Generative recommendation systems exhibit scaling laws unlike older deep learning
“These deep learning based systems, they had flattened out in terms of if you added more user data or more parameters, they did not get better like the LLMs. There were no scaling laws. It's just like, it is what it is, and you could move at .2%. There's someth…”
Söderström: Double-digit percentage of Spotify listening occurs in group settings
“When we survey users and we ask them what percentage of your listening is with others, it's a huge percentage. Double digit percentage. So music is actually a very social activity still.”
Söderström: Apple held roughly 98% of podcast market without investing
“Apple hadn't invested in it and they had like nine to eight percent
Of the market.”