I just wish they fixed the problem with the bands disambiguation. I just hate everytime I get a recommendation in the release radar for a band that happens to have the same name as the one I like but stylewise they have nothing to do with the music I usually listen.
I'm still stuck wondering if one artist I found recently had a very different style originally, or if it lumped two unrelated acts together.
They do lump unrelated acts together. It's somewhat obvious when one of them is a current young musician and other had their career in 60s.
That's the exact reason I switched to YouTube Music. I remember even about 10 years ago, last.fm's algorithm on bringing songs I would enjoy basing on previously listened/liked songs was far more accurate than Spotify's today.
I think Youtube Music does a decent job on recommendations but its android app is annoying. How does it still not have a horizontal screen mode?
The iOS and web app are _bad_. And knowing Google, it won't get any better anytime soon.
I have seen this article before. The author are so close to understanding why Spotify does NOT need a dislike button >To dislike a song is easy – to like one, you have to actually invest time in it. Likewise, to detect if a user dislikes a song is easy since the user will not invest time listening to it: the skip button is the dislike button. In contrast, detecting if a user likes a song is more difficult: there need…
Sometimes your just not in the mood for a specific song or genre. A skip doesn’t imply I’d never want to hear a song again.
So basically they’re developing a product marketed to streaming platforms to identify what’s “more attractive” to consumers. The algorithms are still optimized for increasing sales and engagement, not improving recommendations (for the user). That’s exactly why I’m unhappy with Spotify lately. I feel I’m being sold a bunch of corporate portfolios on what’s currently in vogue, and not just artistically but politically…
Quoted post unavailable.
With respect, your comment illustrates the issue - a burden to reeducate people by shaping their culture. Music can make a statement, sure, but a world where people are force-fed curated playlists designed to mold them into some Standard Issue set of beliefs, this is a dystopian vision.
With respect, your comment illustrates the issue - a burden to reeducate people by shaping their culture. Music can make a statement, sure, but a world where people are force-fed curated playlists designed to mold them into some Standard Issue set of beliefs, this is a dystopian vision.
Typically there's another p-word used to describe this.
An option to hide all podcasts would significantly improve spotify's recommendations. I don't listen to podcasts (though I have accidentally clicked on one, when a podcast interviewed a musician I like... blech). Why is my screen filled with podcasts multiple times per day? I like the music recommendations, "so and so artist radio," etc., but they make me hunt for them. Also, while I like a broad variety of genres, I…
My guess is that they don't have to pay rights for the podcasts, so for them it's a win/win if you spend your attention on their app listening to free content.
So basically they’re developing a product marketed to streaming platforms to identify what’s “more attractive” to consumers. The algorithms are still optimized for increasing sales and engagement, not improving recommendations (for the user). That’s exactly why I’m unhappy with Spotify lately. I feel I’m being sold a bunch of corporate portfolios on what’s currently in vogue, and not just artistically but politically…
That's the exact reason I switched to YouTube Music. I remember even about 10 years ago, last.fm's algorithm on bringing songs I would enjoy basing on previously listened/liked songs was far more accurate than Spotify's today.
YouTube makes recommendations around engagement metrics only (1 political video gets you Fox News recommendations for years), spotify still gives me plenty of small bands - I think if you’re listening to very popular music you’re screwed either ways
I have seen this article before. The author are so close to understanding why Spotify does NOT need a dislike button >To dislike a song is easy – to like one, you have to actually invest time in it. Likewise, to detect if a user dislikes a song is easy since the user will not invest time listening to it: the skip button is the dislike button. In contrast, detecting if a user likes a song is more difficult: there need…
But there is a difference between "I don't feel like this now" and "I hate this" that skipping a track doesn't capture.
Surely they have some context as to what "now" means?
I read that YouTube Music recommends different music depending on time of day, whether I'm at work / home / in the car, and even takes the weather into account. I'd be surprised if Spotify didn't have some context awareness as well.
I work in this industry and actually we’ve found from online experimentation customers generally don’t want to tell you they dislike something. The expectation is customers get what they like (relevant) and sometimes they want to express stronger affinity for specific things. The converse hasn’t shown to be true. Generally customers don’t have a clear picture of how their explicit signals impact recommendations. Thou…
> Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
So, basically if I dislike a song but want to give it a chance and then decide that you want to never hear it again, your industry thinks that I liked it. That explains a lot.
You are exhibiting google's mistake, caused by over educated people with zero real experience thinking they know better. Or, I should correct that - some code they wrote knows better than the complex brain of the actual person. This results in "AI" (big if/then loop) trying to figure it out, and completely blocking any input from the customer. Here's the issue with "inferred from other activity." You know how you can…
> What this results in, is you have an inferior product, your customers leave, then you leave and go screw up another company's product without ever learning your lesson. I'm not sure what you're suggesting and I don't need to qualify my experience in the software industry or within the domain of recommenders and personalization.
I did not suggest anything. I very clearly stated what I stated, in plain English that is very easy to understand. Your qualifications are irrelevant. Your point of view on the issue we are discussing is wrong for long-term product success, and destroys the hard work others have put into the idea, in order to feel personally superior to your users. I don't know how more plainly I can put it for you if you had trouble understanding that. I do believe however that not understanding something, is not something that has ever prevented you from action or strong and wrong opinions on a subject matter.