It's a great idea with a lot of utility, but my question is: How does dispute resolution work? I.e. How do you prevent bad actors from beating the original artist to the punch, and registering someone else's artwork as theirs on the blockchain?
The short answer is reputation. Consumers of the data should be able to filter by cryptographic signature to prioritize users that have more reputation over ones that don't. A timestamp is just one piece of metadata.
For example, on a platform like Soundcloud there are many Rhiannas but it is obvious which account the "real" Rihanna is because she has the most followers, likes, etc. Popularity can be another data point.
Projects like Blockstack already offer solutions that take a similar approach, allowing you to link your social media accounts so your identity is more trustable.
We also envision developers and trusted organizations like Creative Commons or DPLA creating indexes on top of Mediachain to help users filter the data.
Reputation is not as easy to measure as simple popularity. There are very many "content creators" like LadBible who are orders of magnitude more popular than the actual content creators they steal from.
Okay dispute happens Alice has Reputation X and Bob has Reputation Y. So first who or what does Alice dispute to and second so what: if X > Y, Alice becomes the owner?
Hey Denis,
What differentiates your company from the hundreds of other "blockchain-based" identity companies that do almost the same thing?
I've seen a lot of companies throwing around these exact buzzwords and fancy landing pages for 5 years now and most just seem to be raking in VC funding while fading into obscurity a year later.
We're creating a completely open, decentralized media library and our goal is to help creators reach their audience directly through their content.
We believe that value should follow the content, instead of flowing from a platform.
Today, Mediachain helps preserve attribution for creative works wherever they go, so you can know the creator of that anonymous viral image you see in your feed.
In the future the same infrastructure will enable financial exchange directly between a creator and their audience through their works.
Right now, filecoin is only a whitepaper. There's no implementation yet. Possibly never.
Now, Filecoin seems to, from the whitepaper, gotten it right. We all know of the tragedy of the commons. Their solution was an elegant way around that, by granting filecoin to whom share their bandwidth and local storage. And then, others will take your coin for duplicating files you want copied.
But right now, it doesnt exist. And IPFS still works flawlessly.
Filecoin is one of several proposed approaches for encouraging seeding ("pinning") in the IPFS network. In our network (Mediachain), we will probably start out by using tit-for-tat pinning between peers for this purpose.
I have used IPFS without any filecoin that I've been aware of. It was actually really simple like early versions of rails or symfony simple or simpler.
And if there was a filecoin scam in there I didn't notice it neither on my pc nor on my bank or cash.
What happens if someone crops and scrubs the metadata of the original image and then shares it? Can mediachain identify the original source of the new image?
Any metadata that is stored in EXIF or part of the file can be easily stripped out. This is why Mediachain uses perceptual recognition technology similar to Shazam or Google Image search to identify media based on how it looks or sounds, automatically resolving to metadata stored in Mediachain. Near duplicate image detection is quite far along and works very well even for images that are cropped, distorted, etc.
> Imagine being able to connect with the artist of a viral GIF you see in your feed, learn the history or origin of any image, or automatically reward a musician whenever you press play.
OK, I've gone and imagined it. I'm not entirely convinced that any appreciable number of people actually want to do these things though.
Is this a clever application of a trendy technology creating a solution that's searching for a problem? What are the real life business model use cases for such a tool?
Our goal is to make the monetization of media follow the content, instead of flowing from the platform like it does today.
Today, Mediachain can automatically link an image to who made it. That same channel can be used to transfer value directly through content in the future. We believe that this will create a huge opportunity for developers to build new media applications that reward creators directly through their content, no matter where it is.
The perceptual recognition technology is similar to Shazam or Google Image search (or Tineye), identifying media based on how it looks or sounds.
This allows Mediachain to always point back to information about an image, even if its metadata is stripped out as it goes viral, for example. Near duplicate image detection is quite far along and works well for images that are cropped, distorted, etc.
And what's the utility to end-users, who just want to share a cute picture of a cat, that they should care? And thus that a media host should care to implement this?
Perhaps the media host would receive less DRM take downs and a more accurate 'relevant content' prediction model since the true origin of content is confirmed.
Well, I made a GIF that went viral. It was posted on all of the major websites and attributed to someone's twitter who found it on someone else's twitter. I could see how it would be nice to have made a few dollars/cents from that image. Although, is it my content or CSPAN's (the original makers of the video?)
How does your perceptual recognition technology distinguish piracy from Fair Use? Given the impossibility of knowing with certainty which jurisdiction the end user is in, which Copyright scheme are you assuming? Can works be registered as Public Domain or as subject to a permissive license such as Creative Commons?
This sounds like you're making an Internet-scale DRM scheme. What precautions are taking to protect people from creators trying to take more than copyright provides?
The goal of Mediachain isn't to enforce scarcity of media online, but to help creators benefit from the scale of sharing. Today, an image can easily go viral without viewers knowing who the author was. Mediachain makes sure that the creator can be present wherever the content goes, automatically.
Our view is that bits are easily reproduced and there is no way to stop that, and until now there's been no easy way for creators to benefit from that. Mediachain allows creators to connect directly to their audiences through the content itself. This way, gratitude can be exchanged through attribution, and in the future, even a payment directly through the content.
It's a partial screenshot of the article in my browser. I intend to use it in a lecture as an example of the complications behind fair use. (irl I really do teach such a lecture). Who is it attributable to? The artist of the drawing, me, techcrunch or mediachain?
In short: I am suspicious of any service, and there have been a few, that claims to be able to identify the owner or creator of a work. It's a layering problem, with nearly every digital image being at some level the property of a great many people.
(fyi: if this was a law exam, the best answer is "me" in that I created the composition and am well-covered by fair use in my use of it here.)
Declaring ownership and provenance with complete certainty is very hard, which is why we don't profess to do it. Our approach is to nondestructively aggregate potentially conflicting claims and allow for more nuanced resolution at read time, as appropriate for a given situation. In your particular example, we would have a chain of derivative works.
Obviously, in certain cases (payment routing) a definitive answer is required, but this is already something that PROs and similar organizations deal with on a daily basis, and we can support their arbitration.
If you plan on linking money to content, you better have a really definitive answer. That movement of money is essentially the purchase of a license. If a business or individual relies on your answer as to who that license should be purchased from, they do not want to be hit with a copyright complaint should the real owner appear. And if people end up abusing your system to monetize content that they should not, expect all sorts of legal docs in the mail.
33 comments
[ 4.5 ms ] story [ 81.7 ms ] threadFor example, on a platform like Soundcloud there are many Rhiannas but it is obvious which account the "real" Rihanna is because she has the most followers, likes, etc. Popularity can be another data point.
Projects like Blockstack already offer solutions that take a similar approach, allowing you to link your social media accounts so your identity is more trustable.
We also envision developers and trusted organizations like Creative Commons or DPLA creating indexes on top of Mediachain to help users filter the data.
Okay dispute happens Alice has Reputation X and Bob has Reputation Y. So first who or what does Alice dispute to and second so what: if X > Y, Alice becomes the owner?
This Co.Design article does a great job of explaining how Mediachain helps creators: http://www.fastcodesign.com/3060426/could-blockchains-solve-...
Here is our official fundraising announcement with more details: https://blog.mediachain.io/mediachain-labs-funded-by-union-s...
I've seen a lot of companies throwing around these exact buzzwords and fancy landing pages for 5 years now and most just seem to be raking in VC funding while fading into obscurity a year later.
There's a list if you scroll down on this link: http://www.blockchaintechnologies.com/blockchain-companies
We believe that value should follow the content, instead of flowing from a platform.
Today, Mediachain helps preserve attribution for creative works wherever they go, so you can know the creator of that anonymous viral image you see in your feed.
In the future the same infrastructure will enable financial exchange directly between a creator and their audience through their works.
We've written some more about the implications of Mediachain here: https://blog.mediachain.io/the-gif-that-fell-to-earth-eae706... https://blog.mediachain.io/what-a-blockchain-for-music-reall...
It's not clear to me that Filecoin is any more or less of a scam than Bitcoin, nor that involving "blockchains" adds any value at all in this case.
Now, Filecoin seems to, from the whitepaper, gotten it right. We all know of the tragedy of the commons. Their solution was an elegant way around that, by granting filecoin to whom share their bandwidth and local storage. And then, others will take your coin for duplicating files you want copied.
But right now, it doesnt exist. And IPFS still works flawlessly.
This is what every blockchain does.. (lottery mining)
>But right now, it doesnt exist. And IPFS still works flawlessly
Reading the IPFS website, I was under the impression IPFS nodes were utilizing filecoin. My bad.
And if there was a filecoin scam in there I didn't notice it neither on my pc nor on my bank or cash.
http://www.mediachainlabs.com/jobs/
More on our approach in this post: https://blog.mediachain.io/perceptual-resolution-9c00ad5ca55...
Details on the implementation in this RFC: https://github.com/mediachain/mediachain/blob/master/rfc/med...
OK, I've gone and imagined it. I'm not entirely convinced that any appreciable number of people actually want to do these things though.
Is this a clever application of a trendy technology creating a solution that's searching for a problem? What are the real life business model use cases for such a tool?
Today, Mediachain can automatically link an image to who made it. That same channel can be used to transfer value directly through content in the future. We believe that this will create a huge opportunity for developers to build new media applications that reward creators directly through their content, no matter where it is.
Does your system require metadata along with the image? If so, what about once an image escapes the metadata?
This allows Mediachain to always point back to information about an image, even if its metadata is stripped out as it goes viral, for example. Near duplicate image detection is quite far along and works well for images that are cropped, distorted, etc.
More on our approach in this post: https://blog.mediachain.io/perceptual-resolution-9c00ad5ca55...
Details on the implementation in this RFC: https://github.com/mediachain/mediachain/blob/master/rfc/med...
This sounds like you're making an Internet-scale DRM scheme. What precautions are taking to protect people from creators trying to take more than copyright provides?
Our view is that bits are easily reproduced and there is no way to stop that, and until now there's been no easy way for creators to benefit from that. Mediachain allows creators to connect directly to their audiences through the content itself. This way, gratitude can be exchanged through attribution, and in the future, even a payment directly through the content.
It's a partial screenshot of the article in my browser. I intend to use it in a lecture as an example of the complications behind fair use. (irl I really do teach such a lecture). Who is it attributable to? The artist of the drawing, me, techcrunch or mediachain?
In short: I am suspicious of any service, and there have been a few, that claims to be able to identify the owner or creator of a work. It's a layering problem, with nearly every digital image being at some level the property of a great many people.
(fyi: if this was a law exam, the best answer is "me" in that I created the composition and am well-covered by fair use in my use of it here.)
Obviously, in certain cases (payment routing) a definitive answer is required, but this is already something that PROs and similar organizations deal with on a daily basis, and we can support their arbitration.