Yesterday, Apple released developer beta 1 of iOS 26.4, which among other things adds a feature to the Music app that uses Apple Intelligence to generate a playlist from a short description of what the user wants to hear. That immediately reminded Federico and me of The Sentence, a Beats Music feature that sadly didn’t survive the app’s acquisition by Apple.

The Sentence allowed subscribers to describe the music they wanted to hear based on a Mad Libs-style sentence construction. Every sentence was structured as “I’m [location] & feel like [mood] with [person/group] to [music genre].” The feature was a fantastic innovation that made playlist creation fun and easy. As Federico described it in 2014:

It’s The Sentence, though, that steals the spotlight in how it combines regular, Pandora-like song shuffling with a context/mood-based menu to tell Beats what you want to listen to. The Sentence, as the name implies, lets you construct a sentence using variable tokens for location, mood, user, and music genre. You can request things like “I’m at my computer and feel like dancing with myself to pop”, “I’m in the car and feel like driving with my friends to indie”, or more absurd contexts such as “I’m underpaid and I feel like shoveling snow with my lover to metal”. As reported by Re/code [Ed. note: This is a dead link], Beats explained that “the content, and the filters, are selected and tuned by humans, and an algorithm generates the playlist from your choices”.

As you can tell from Federico’s examples, The Sentence could range from the ordinary to more “out there” concepts. Here are several additional examples from reviews that I found today:

• I’m at a party and feel like BBQing with my BFF to Dance-Pop.
• I’m at my computer and feel like going back in time with strangers to jazz vocals.
• I’m at work being productive with my co-workers to the 2000s.
• I’m in the future and feel like no kicking back with my boss to Hip-Hop.
• I’m bored as hell and feel like hiding in a closet with the paparazzi to Dance.

Of course, I had to try some of these classic sentences with Music’s new Apple Intelligence integration, which, it’s worth noting, requires an Internet connection to work and can take anywhere from under 10 seconds to a minute or more. Here are the results:

I’m at a party and feel like BBQing with my BFF to Dance-Pop

I’m at my computer and feel like going back in time with strangers to jazz vocals

I’m at work being productive with my co-workers to the 2000s

I’m in the future and feel like kicking back with my boss to Hip-Hop

I’m bored as hell and feel like hiding in a closet with the paparazzi to Dance

Alas, that last prompt from Federico (of course) was a little too much for Music. It failed two attempts, although to be fair it worked once earlier in my testing. Otherwise, I think Music’s new feature did pretty well. Judging from the titles, Music seems to have mostly ignored the more nonsensical aspects of The Sentence’s prompts, focusing on mood and genre to generate its 25-song playlists.

If you’re wondering whether there are guardrails on the feature, there are. Prompts like “I’m in bed, and feel like waking up, with your mom, to musica Mexicana.” that I found in Ellis Hamburger’s review for The Verge failed every time. There’s mood music and then there’s mood music, and the latter is off limits.

Of course, it’s also worth keeping in mind that this is developer beta 1, so the feature is likely to evolve before it is publicly released, but if you’re on the beta train and feeling nostalgic for The Sentence, Music’s new Apple Intelligence playlists are a fun way to semi-relive it.