Personally I don’t see why the views of those that write software should really concern us, as long as the technical implementation is not biased. It’s open-source and people can take it and do with it what they please. No-one is forcing you to accept certain views or think about things critically (including assessing others viewpoints that may be different to yours). I feel like it’s a bit of a waste of time to worry about these things.
It is, in the instance chooser there is a “defederation score” that is “good” if you have defederated a sufficient amount of instances Rimu doesn’t like.
biases manifest in really strange and unexpected ways and you’ll fail even when you try to intentionally account for them; that’s why things like facial recognition success rate correlates to the darkness of your skin or why successful ai recognition of text/speech is related how different your language is to english or mandarin.
the only way to successfully gaurd against bias in software development is to have developer teams comprised of people can naturally keep each other’s biases in check.
I think facial recognition technology is very different to threadiverse software. The fact that those technologies are trained on predominantly-white data is no surprise, both of your examples are data-based (ML models) where the data itself contains the bias.
I am talking more of the open-source projects, it’s important; as you rightfully call out, that we have a varied group of opinions within the developer group 👍
it’s not just ai or training data for it; it’s developers themselves too, including fediverse ones.
the biggest non-ai/training examples that i can think of came from times before ai was ever a thing like:
usps had difficulty validating addresses because the software they obtained assumed euro-centric naming schema
airlines, health care providers, hotels, and state motor vehicle departments rejected registration/reservations because trans people have to option to select their sex
health care providers misdosed patients because the software they used didn’t account for highly-athletic/bodybuilding people or people with chronic conditions
there are SO MANY examples out there where the bias clearly comes from the developer instead of the training data and there’s no way that any piefed developer is immune or can even effectively mitigate their biases.
as long as the technical implementation is not biased.
Piefed has been baking in biases though, like adding a bunch of leftist/international sources in a huge blocklist labelled ‘qanon’, which if you then want to disable it’s got to be done manually.
Not to mention clear disregard for breaking activitypub interoperability.
Yeah, I’m not a huge fan of it myself and don’t see the need for it really. If that’s right about the baked-in blocklist, then this was the right choice. You should be able to craft your own blocklists, if you so choose to.
Hmm, you’re talking about the little red warning triangle? In theory this could be cool if it was something each instance could configure using custom url-lists (as you know we all don’t agree on which perspective is “right”) but if it’s coming packaged with PieFed, then I would call that inherent bias in the product. 🤨
there’s also a block component to that as well, his ‘qanon’ blocklist is something you toggle on and can manually edit it via ui; it pulls from a github he runs and contains leftist sites like grayzone as well as sites like wikileaks.org. here’s how the interface works, you either click every button on the list or edit the database/use a different one
quoting davel here:
domain blocklist, which is not mandatory on install—though most new instances choose it—and can be removed by an admin with either the patience or the database skill to remove it.
Personally I don’t see why the views of those that write software should really concern us, as long as the technical implementation is not biased. It’s open-source and people can take it and do with it what they please. No-one is forcing you to accept certain views or think about things critically (including assessing others viewpoints that may be different to yours). I feel like it’s a bit of a waste of time to worry about these things.
It is, in the instance chooser there is a “defederation score” that is “good” if you have defederated a sufficient amount of instances Rimu doesn’t like.
biases manifest in really strange and unexpected ways and you’ll fail even when you try to intentionally account for them; that’s why things like facial recognition success rate correlates to the darkness of your skin or why successful ai recognition of text/speech is related how different your language is to english or mandarin.
the only way to successfully gaurd against bias in software development is to have developer teams comprised of people can naturally keep each other’s biases in check.
I think facial recognition technology is very different to threadiverse software. The fact that those technologies are trained on predominantly-white data is no surprise, both of your examples are data-based (ML models) where the data itself contains the bias.
I am talking more of the open-source projects, it’s important; as you rightfully call out, that we have a varied group of opinions within the developer group 👍
it’s not just ai or training data for it; it’s developers themselves too, including fediverse ones.
the biggest non-ai/training examples that i can think of came from times before ai was ever a thing like:
usps had difficulty validating addresses because the software they obtained assumed euro-centric naming schema
airlines, health care providers, hotels, and state motor vehicle departments rejected registration/reservations because trans people have to option to select their sex
health care providers misdosed patients because the software they used didn’t account for highly-athletic/bodybuilding people or people with chronic conditions
there are SO MANY examples out there where the bias clearly comes from the developer instead of the training data and there’s no way that any piefed developer is immune or can even effectively mitigate their biases.
Piefed has been baking in biases though, like adding a bunch of leftist/international sources in a huge blocklist labelled ‘qanon’, which if you then want to disable it’s got to be done manually.
Not to mention clear disregard for breaking activitypub interoperability.
Yeah, I’m not a huge fan of it myself and don’t see the need for it really. If that’s right about the baked-in blocklist, then this was the right choice. You should be able to craft your own blocklists, if you so choose to.
relevant thread for what i’m referencing
Hmm, you’re talking about the little red warning triangle? In theory this could be cool if it was something each instance could configure using custom url-lists (as you know we all don’t agree on which perspective is “right”) but if it’s coming packaged with PieFed, then I would call that inherent bias in the product. 🤨
It’s essentially the Media Bias Fact Checker Bot as an integrated feature in the platform.
there’s also a block component to that as well, his ‘qanon’ blocklist is something you toggle on and can manually edit it via ui; it pulls from a github he runs and contains leftist sites like grayzone as well as sites like wikileaks.org.
here’s how the interface works, you either click every button on the list or edit the database/use a different one
quoting davel here: