5 ms·

This looks like really cool technology, but what is the use case for "normal" users? The Subconscious site [1] which leverages this library lists the key value

This looks like really cool technology, but what is the use case for "normal" users? The Subconscious site [1] which leverages this library lists the key value propositions as "Own Your Data," "Decentralized," and "Open Source," which are three things average users demonstrably don't care about. I'm also concerned about the immutability of information stored in this way. [2] There are a number of reasons why immutability (in this case a fully public version history) is an anti-feature, with significant real-world risks. I would push the protocol authors to fully consider the ramifications of doxxing, harassment, accidental self-disclosure, etc—especially in authoritarian nations. If they don't have a story for how they'll handle this, it's not an appropriate foundation for the internet of ideas. [1]: https://subconscious.network https://subconscious.network [2]: https://github.com/subconsciousnetwork/noosphere/blob/main/design/explainer.md#immutable-data https://github.com/subconsciousnetwork/noosphere/blob/main/d...

CrypticShift 4y ago
>but what is the use case for "normal" users? >three things average users demonstrably don't care about I always ask this question myself (and have an interest in all things no-code). But this is about the underlying Internet layers. For decades, from the ARPANET to the IETF, a small clique were unchallenged and unhurried in presenting their solutions methodically in RFC after RFC. They did a great job. Now The upper OSI model is breaking up. Everybody has a vision and a voice. Every day brings new projects like this one here in HN. There is not even large consensus around that leading Web3 banner [0]. So, IMO this natural selection process will take us at least a decade (or two?) to coagulate into standards that support mainstream "normal users" use cases [0] https://news.ycombinator.com/item?id=33469978 https://news.ycombinator.com/item?id=33469978
mattdesl 4y ago
A lot of users do care about this when push comes to shove. An example is happening on Twitter: a lot of users want to escape Twitter onto another platform, but they are 'locked' to the walled garden because they don't own their data in a way that allows for an exit. IPFS isn't perfect but does give you the ability to migrate to other (competing) hosting platforms without having to change your URLs. I'd argue that a similar thing happens with end-to-end encryption. A lot of users don't care or even think about whether their Twitter or Mastodon direct messages are encrypted, until a malicious actor reads these messages or a data breach has them released online.
deafpolygon 4y ago
How does "decentralized" and "own your own data" fit together? Presumably, decentralized means your data is on everyone else's server, rather than on one provider (the 'cloud') or your own ('self-hosted')?
- cdata 4y ago
  When we say "decentralized," we don't mean that your data is deliberately scattered around the network; we mean that the underlying infrastructure providers are fungible. In other words, you own your data because you can change infrastructure providers as it suits you, and other users who interact with you on the network generally won't notice or otherwise care that you did that.
cdata 4y ago
Hi there, I work with Gordon (OP) on Subconscious and Noosphere. This is a topic that we take seriously and think about a lot together. We've had some good discussions in our Discord, and some of our motivating goals have manifested as issues on Github: https://github.com/subconsciousnetwork/noosphere/issues?q=is%3Aissue+is%3Aopen+label%3A%22Topic%3A+Privacy%22+ https://github.com/subconsciousnetwork/noosphere/issues?q=is... Our work is open source, and we are building it in the open hand-in-hand with our community. I personally respond to as many issues and stray comments on Discord as I can. If this is a topic that you care about, your feedback, comments and ideas would be very welcome. > I would push the protocol authors to fully consider the ramifications of doxxing, harassment, accidental self-disclosure, etc—especially in authoritarian nations. If they don't have a story for how they'll handle this, it's not an appropriate foundation for the internet of ideas. My personal opinion is that we need more than a story. Our users need tools that empower them to be in control of their data, and that includes being able to keep private data private, fix mistakes and execute countermeasures when they are threatened with abuse. And it's true: immutable data is a double-edged sword. The primary benefit we seek from it is content addressing, which enables us to look up some data in a way that makes data providers fungible. This has an advantage when it comes to censorship: if censored network A can't give you your data, you may be able to find it on uncensored network B as long as your know the content address. It also presents a challenge: sometimes a user may rightly wish for data to be removed from the network; content addressing creates the possibility that an abuser may have access to data in a way that further dis-empowers users who are already vulnerable today. Here is a strong opinion (weakly held) that I have about mutability: the "mutability" of the hypertext web has not spared us from the kinds of danger described above. We have leaned on mutability as a privacy crutch, and in turn we have excused ourselves from providing capabilities and interfaces that enable real privacy and agency for users under threat. For example, a user may think a Snap message is extra private - from their perspective, the message is ephemeral - but others will still find ways to make copies and propagate the data across the network, and Snap the company certainly has access to the data (perhaps in perpetuity). What recourse does that user have once their data escapes the mutable sandbox? It's possible that immutable data will enable us to build better tools for taking pro-active action against abuse. For example: today on the hypertext web, the recourse for most users to head off the propagation of harmful data is to reach out to service providers individually, file abuse claims and hope that the content is eventually removed. Victims are burdened with both identifying that the data is stored with a provider, and also the clerical task of reporting the abuse over and over again. In a world with content addressing, anti-abuse filters full of content addresses may propagate fluidly from provider to provider, which suggests a far less burdensome process for the abuse victim - they don't need to know that the content has surfaced on some provider in the distance, because the provider is already aware of the abuse claim and can pro-actively refuse to propagate the data. In the end, I think the best solutions to the problems of this kind of abuse will be social/legal ones, supported by a technical foundation that can enable pro-active countermeasures. All that said, I'm not an expert on this topic, I just like to spend a lot of time thinking about it. If you or someone else reading this is an expert, and you have some time to spare, please come hang out in our Discord or on Github and teach us what you know!

CrypticShift 4y ago

>but what is the use case for "normal" users? >three things average users demonstrably don't care about I always ask this question myself (and have an interest in all things no-code). But this is about the underlying Internet layers. For decades, from the ARPANET to the IETF, a small clique were unchallenged and unhurried in presenting their solutions methodically in RFC after RFC. They did a great job. Now The upper OSI model is breaking up. Everybody has a vision and a voice. Every day brings new projects like this one here in HN. There is not even large consensus around that leading Web3 banner [0]. So, IMO this natural selection process will take us at least a decade (or two?) to coagulate into standards that support mainstream "normal users" use cases [0] https://news.ycombinator.com/item?id=33469978 https://news.ycombinator.com/item?id=33469978

mattdesl 4y ago

A lot of users do care about this when push comes to shove. An example is happening on Twitter: a lot of users want to escape Twitter onto another platform, but they are 'locked' to the walled garden because they don't own their data in a way that allows for an exit. IPFS isn't perfect but does give you the ability to migrate to other (competing) hosting platforms without having to change your URLs. I'd argue that a similar thing happens with end-to-end encryption. A lot of users don't care or even think about whether their Twitter or Mastodon direct messages are encrypted, until a malicious actor reads these messages or a data breach has them released online.

deafpolygon 4y ago

How does "decentralized" and "own your own data" fit together? Presumably, decentralized means your data is on everyone else's server, rather than on one provider (the 'cloud') or your own ('self-hosted')?

cdata 4y ago
When we say "decentralized," we don't mean that your data is deliberately scattered around the network; we mean that the underlying infrastructure providers are fungible. In other words, you own your data because you can change infrastructure providers as it suits you, and other users who interact with you on the network generally won't notice or otherwise care that you did that.

cdata 4y ago

Hi there, I work with Gordon (OP) on Subconscious and Noosphere. This is a topic that we take seriously and think about a lot together. We've had some good discussions in our Discord, and some of our motivating goals have manifested as issues on Github: https://github.com/subconsciousnetwork/noosphere/issues?q=is%3Aissue+is%3Aopen+label%3A%22Topic%3A+Privacy%22+ https://github.com/subconsciousnetwork/noosphere/issues?q=is... Our work is open source, and we are building it in the open hand-in-hand with our community. I personally respond to as many issues and stray comments on Discord as I can. If this is a topic that you care about, your feedback, comments and ideas would be very welcome. > I would push the protocol authors to fully consider the ramifications of doxxing, harassment, accidental self-disclosure, etc—especially in authoritarian nations. If they don't have a story for how they'll handle this, it's not an appropriate foundation for the internet of ideas. My personal opinion is that we need more than a story. Our users need tools that empower them to be in control of their data, and that includes being able to keep private data private, fix mistakes and execute countermeasures when they are threatened with abuse. And it's true: immutable data is a double-edged sword. The primary benefit we seek from it is content addressing, which enables us to look up some data in a way that makes data providers fungible. This has an advantage when it comes to censorship: if censored network A can't give you your data, you may be able to find it on uncensored network B as long as your know the content address. It also presents a challenge: sometimes a user may rightly wish for data to be removed from the network; content addressing creates the possibility that an abuser may have access to data in a way that further dis-empowers users who are already vulnerable today. Here is a strong opinion (weakly held) that I have about mutability: the "mutability" of the hypertext web has not spared us from the kinds of danger described above. We have leaned on mutability as a privacy crutch, and in turn we have excused ourselves from providing capabilities and interfaces that enable real privacy and agency for users under threat. For example, a user may think a Snap message is extra private - from their perspective, the message is ephemeral - but others will still find ways to make copies and propagate the data across the network, and Snap the company certainly has access to the data (perhaps in perpetuity). What recourse does that user have once their data escapes the mutable sandbox? It's possible that immutable data will enable us to build better tools for taking pro-active action against abuse. For example: today on the hypertext web, the recourse for most users to head off the propagation of harmful data is to reach out to service providers individually, file abuse claims and hope that the content is eventually removed. Victims are burdened with both identifying that the data is stored with a provider, and also the clerical task of reporting the abuse over and over again. In a world with content addressing, anti-abuse filters full of content addresses may propagate fluidly from provider to provider, which suggests a far less burdensome process for the abuse victim - they don't need to know that the content has surfaced on some provider in the distance, because the provider is already aware of the abuse claim and can pro-actively refuse to propagate the data. In the end, I think the best solutions to the problems of this kind of abuse will be social/legal ones, supported by a technical foundation that can enable pro-active countermeasures. All that said, I'm not an expert on this topic, I just like to spend a lot of time thinking about it. If you or someone else reading this is an expert, and you have some time to spare, please come hang out in our Discord or on Github and teach us what you know!