Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I've lost the link to the article, but there's still a certain something that Spotify lacks compared to a human radio playlist programmer.


Two things:

Ability to craft a mood or tell a story, relate to current events, culture.

The other is few humans get to do that directly now, and the experience is filtered through algorithms, and then cloned and shuffled.

Takeaway: "cool", relevant, "connected" come from minds, not rules, or limited AI in use today.

People are supreme at discerning the intent of other, through nearly any medium. Think prisoner tapping on the wall.

The more diluted, distant, muffled that intent is, the less compelling the expression of it is.

We quite simply get a lesser sense of the mind.

Think original painting vs a print.

Or live DJ making an evening, connecting with, relating to, and me sharing their local scene directly. By that I mean picking tracks they want, for human reasons, talking for their reasons, not delivering liners, even pausing for a few seconds to let something awesome sink in. Couple that with phones, social media, and it's a compelling experience, given that DJ understands and has mastery of their scene and can express it.

Now, that same DJ, voice tracked all over the nation, picking from a curated, "playola" list of "researched" tunes, and delivered via automation, play lists, etc...

The former can be pure magic. The latter ranging from pleasant to barely tolerable.

That is what is missing.

How many big companies have internal playlists? And if they do, what do people do, how do the tunes get on there and why do they matter?

Great questions worth really thinking about in human terms.

An example might be, say me making a music recommendation to a member with access to the list.

It starts with a quick exchange via SMS, email, whatever, "man, I'm waist deep in this driver code and just can't focus."

"Feel you brother / sister, give these guys a go, they are my goto, for low level assembly language programming. Give them 5 minutes, and you are there, in flow."

(ozric tentacles, would plug in here, for a music selection)

Later, "Man, they made a bazillion albums! This is great!"

On the company play list later on, "Lose flow? A buddy sent me these guys, thank him later, enjoy!"

Chatter happens.

Later still, someone tags selects a few tunes for flow, trance, whatever...

At each layer, the human spark of relevancy, meaning, context gets diluted, until it's just another data point tag like all the other tags.

Going back to the evening DJ.

Making a set, live, maybe lightly prepped for Friday the 4th isn't the same as a Friday goto set for summer, which isn't the same as researched, can't miss hits a program selects from on Fridays.

At each layer here the human to human contact is diluted, more distant, less expressive, less relevant, you get the idea.

Lastly, a human playlist programmer is one step removed from a human spinning tunes directly. But it's good, in that streaming eliminated the need for massive, broad generalization common to, say syndicated radio.

Not all expression forms can be done play list programmer style, but the basics can. Mood, time, feel, general matchup people can identify with cone across just fine, depending on how directly the play list is able to be specified.

People like and identify with those.

To connect, and get the real human mind to mind does require a person with considerable agency, perception, creativity, and talent.

Mass broadcast, as deregulation allowed it all to scale, diluted this. Secondly, the need for humans inhibited the ability to trade on the value they create without cutting them in and or yielding editorial control.

The product is a shell of what can be done, but it scales, can be reproduced, bought, sold, managed editorially.

Those are the basic dynamics as I see them.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: