Tyga makes history with a 0.0 pitchfork rating

Today is a historic day for AI music, cause generative music tools just helped Tyga's latest album receive the lowest Pitchfork score that any album has gotten in 19 years. The reviewer, hilariously, whose name is Drew Millard, suggests if you want to know what the album sounds like, you should just go into an AI music generator and you can generate something that might sound like it would have belonged to that album anyway. The reviewer also talks about how the first time Tyga was asked if the album was entirely human made, he said no, but when asked a second time if AI was used on the record, he said yes, I did use it. It's very much like how Auto Tune was being received a couple decades ago and tried to make the argument that if people like the music, then it shouldn't really matter how it was made. Well clearly people don't like the music, so maybe it does matter. I think you should go read the review if you want a good chuckle. But I do think the thing in this review that I really like that is mentioned is that we're not gonna see this as the death of AI music. We really won't. This author is talking about a human to AI ratio in terms of using AI as a tool. That clearly is not perfected on this album, but does not mean that people will not continue to try. Working with humans can be hard and challenging. They require money and they require time to make something that's good. Whereas AI is always going to cost less, at least while it's still being propped up by so much VC money. And will be in all of the non-human ways that it is a lot more efficient than asking a person to do something. In any commercial industry that's optimizing for cost, if you can get something cheaper and faster, you will likely opt for the cheaper and faster solution. However, if it makes the output that you used the cheap and fast solution for less commercial, then people might start to swing in the other direction. But as soon as we have an AI generated hit, I can prove to be as commercially successful as any other song made by a person. Maybe that's the point at which we would continue to see AI generated music. Right now, though, it seems like we're marked safe from any major Tyga radio singles because according to Pitchfork, this is the worst album of the last almost 20 years. There's also a lot of conversation in the reviews from Pitchfork about the history of sampling in music and specifically hip hop music and pulling on inspiration from local communities and local genres to make new sounds. Can AI access those references a lot faster and a lot easier? Probably. But doesn't an artist also lose control over how that output is generated, no matter how many turns they take with prompt engineering the tool to give them what they're asking for. It reminds me of why people got mad at Hank Green because he was doing research using AI for his videos and people said we don't believe that. Your sources are the ones you would have found if you went and looked for this on your own. The output of any model is always going to be the amalgamation of the training data, which might differ slightly from one tool to the next to the next, and differ based on the weights that are put on those systems and how those models were trained by the companies who trained them. I will say until I'm blue in the face that AI is not a neutral technology. And that would be just as true for researching something to make a YouTube video about as it would be for generating an 80s synthpop sounding sample. Maybe we don't always know the references that our favorite artists are pulling into the music that they make. But the fact that they can point to a song where a sample came from, or point to a genre that inspired them to make the music that they're making is going to be something that an artist that uses AI to generate those references will never be able to do because all of these models are so entirely black box.