Hace unos días, en un post anterior, comentábamos como Andreas Weigend, anteriormente Chief Scientist de Amazon, había fichado como Chief Strategist de MusicStrands. Como muchos sabéis, Amazon es una empresa que me fascina, y las circunstancias de MusicStrands las conozco bastante bien por mi amistad con alguno de sus fundadores, que de hecho vinieron, en su momento, a comentar la idea con un grupo de profesores del Área de Sistemas y Tecnologías de Información del Instituto de Empresa.
Así que aproveché la ocasión para pedirle a Andreas que me contestase a algunas preguntas sobre cosas como los sistemas de recomendación, los efectos de la «long tail» sobre el mercado de música, la inteligencia artificial, las reacciones de la industria de la música contra sus usuarios y algunos temas afines. Creo que la entrevista ha quedado francamente interesante, así que la copio y la pego entera (además, es la primera vez que hago una entrevista en exclusiva para mi blog :-)
Interview by Enrique Dans, Professor at IE Business School
Corvallis, USA. Tuesday April 19th
Dr. Andreas Weigend, Chief Strategist, MusicStrands
- There seems to be an opportunity for music recommendation systems such as MusicStrands, as a sort of metamediary in between the users and the industry. Do you see people trusting music recommendations from computers more than from friends or from the MTV?
The power doesn’t lie in the computer, but lies in the connectivity of the social structure of the network. It’s not trusting the computer, but trusting the ability of the structures intelligence to extract recommendations. The recommendations it pulls from many more people then I can personally interact with.
Artificial Intelligence and machine learning is required to combine and enable filtering of this large number of people and their ratings. Plus other attributes such as combining the tags people put on the music to make them really useful for the end user.
- Following Chris Anderson’s article in Wired, “The long tail”… do you envision a future with less big hits? Music recommendation systems can go both ways, either recommending the same tune to everyone, or helping people dig into the huge variety of music available. What is likely to be the net effect?
I expect that we will see more of the fragmentation of musical taste. So, I expect that the slope of the curve will be more flat. We have already seen this fragmentation in news groups where people, irrespective of their location, age, origin; find each other based on common interest. And they often discover other people in the world who have similar interest; who, if they had only watched broadcast television would have never discovered their common interest. Think about minorities as an example, and the same will happen in music, and is happening in music already. Small groups who find each other and then empower each other, reinforce each other in their taste and ideas.
My prediction is that it will be less of a hit driven business and we will see more of the Indies, (independent artist) appear, which is one of the missions of MusicStrands; to help independent artists surface in those communities where they’re appreciated.
- The internet has created a channel for people to share and express themselves. One of the basics of MusicStrands is recommending music based on the taste of people, which means that people get to express themselves and share their taste in music. How do you see this tool MusicStrands offers in the scheme of people sharing?
We are very honored to have this individual with one of the most famous blogs in Spain. The same empowerment that allows him to reach hundreds of thousands of people in his blog, now empowers those people who have something to say and share with hundreds of thousands of people too.
If you take the shape of a pyramid in Egypt, the top being represented by the people who make the music and the citizens, the listeners are down at the bottom. The amazingly great thing about the Internet is that this pyramid has been turned upside down, down side up, that you now have, all the listeners expressing what they’re interested in. The music doesn’t get pushed anymore from the labels to the people, but to some degree pulled from the people. By them being able to express their interest and the ability to pair up together to form groups based on common interest.
So there is a reversal of the pyramid that music doesn’t get pushed, it gets pulled, that the meaning doesn’t get created by a few gate keepers, but by that of what the masses are interested in. You could have an aristocratic view and say that the masses don’t know what’s interesting and they don’t know what’s good for them, but I think there’s enough amazing people in the world, and that there is more intelligence and more creativity if you turn that pyramid upside down then the other way around.
- To what extent do you think it is possible to apply artificial intelligence based systems to tasks such as searching or translating? Are we getting closer to the AI revolution?
In some way the term Artificial Intelligence, (AI) seems to be something we think of as in the future. But if you think about the world we live in right now, that fact that I can not only search through 10 billion documents from my little mobile device anywhere in the world, using Yahoo or Google, but also through hundreds of thousands of books using Amazon.com. The fact that the results are ordered in a way based on the things I’m likely to be interested in, given where I am, given how I framed my query, are showing up as first on the list of my mobile phone. Well, if that is not artificial intelligence, then I don’t know what artificial intelligence would be.
- For the last five years or so, we’ve been witnessing a harsh fight between P2P file sharing systems and copyright holders… Will we see the end of P2P file sharing? Do you expect, at some point in time, all or most of the music will be digitally exchanged through legal online stores such as Apple’s iTunes, Napster and others? Will both systems coexist? What do you think of the tactics used by the music industry, such as poisoning P2P networks, legally prosecuting individual users, etc.? Where is the music industry standing right now? Better or worse situation than in year 2000, when Shawn Fanning came up with Napster?
As you might know, I spend a lot of time, several months a year in China teaching executives and that has given me somewhat a different perspective on copyright issues and also with business models that go with it. Everybody knows that for the essential cost of a blank CD or DVD, you can buy whatever you want to buy on that CD or DVD in Asia. So, what we have seen is a much smarter evolution of business models which took that as a given. For example, take games, specifically, (MMORPGs), Massive Multiplayer Online Role Playing Games. Companies in China and Korea knew that there was no hope trying to sell the software, so they gave the software away for free. As a matter of fact, when I was in Korea last two three weeks ago, they sent all the employees of one company to install the software for free. They installed their free software in all these PC bands in the net bars, in these PC centers where the youngsters go to.
These new business models emerged and those are that you pay for quantitivity, rather than for the software. And that is actually only a fraction of what the market has become. Namely, people trade those artifacts they create on these servers in those games.
So, that was Asia’s response, creating multibillion dollar market cap companies in the gaming space. While the software, the games themselves, propagating, basically without anybody paying for them. I would expect some smart people in the world coming up with similar models, I’m not smart enough for that, I don’t know what they are, but coming up with similar models in the music industry.