Blaine
@blaine.bsky.social
Sociotechnologist, activist. First employee @ Twitter, built that and pioneered social decentralization, created OAuth. Shy cloud-observer, mostly. Building for Local at @wearenewpublic.
Thanks, Joe! 💜 Can't wait to share more!
December 2, 2025 at 3:12 AM
Thanks, Joe! 💜 Can't wait to share more!
We also hope that not having engagement-bait algorithmic "optimization" plus robust moderation and community management tools will make our feeds better, and we have features that are designed to be "evergreen" spaces for members of the community to meet, grow, and learn about their community.
December 2, 2025 at 3:12 AM
We also hope that not having engagement-bait algorithmic "optimization" plus robust moderation and community management tools will make our feeds better, and we have features that are designed to be "evergreen" spaces for members of the community to meet, grow, and learn about their community.
Re: the feed format, we do have feed aspects, but there's a lot that's perhaps not obvious from the screenshots. For example, our "Highlights" community homepage is a steward-curated section that actively tries to break the feed paradigm.
December 2, 2025 at 3:12 AM
Re: the feed format, we do have feed aspects, but there's a lot that's perhaps not obvious from the screenshots. For example, our "Highlights" community homepage is a steward-curated section that actively tries to break the feed paradigm.
That said, we've poured our hearts and souls into the design. We agree with everything you've said, and we'll be sharing more in detail about the design. Our current launch is focused on learning what does and doesn't work with our design, and iterating to support healthy communities.
December 2, 2025 at 3:12 AM
That said, we've poured our hearts and souls into the design. We agree with everything you've said, and we'll be sharing more in detail about the design. Our current launch is focused on learning what does and doesn't work with our design, and iterating to support healthy communities.
I'm not the best person to answer all of them, but in short we've spent _a lot_ of time making sure that resources, training, and support (peer and organizational) are available to our stewards. You're absolutely right that this is more important than the technology, by far, so it's our #1 priority.
December 2, 2025 at 3:12 AM
I'm not the best person to answer all of them, but in short we've spent _a lot_ of time making sure that resources, training, and support (peer and organizational) are available to our stewards. You're absolutely right that this is more important than the technology, by far, so it's our #1 priority.
Excellent questions, among many others that we've been very carefully considering in the design and implementation of the product. 💜
December 2, 2025 at 3:12 AM
Excellent questions, among many others that we've been very carefully considering in the design and implementation of the product. 💜
One of the things our research has found is that the toxicity often comes from the platform! Enshittification in the form of "people can't look away from a good fight, and eyeballs = revenue" is a hell of a drug when you're Facebook.
December 2, 2025 at 1:05 AM
One of the things our research has found is that the toxicity often comes from the platform! Enshittification in the form of "people can't look away from a good fight, and eyeballs = revenue" is a hell of a drug when you're Facebook.
This isn’t another toxic, chaotic space — Roundabout is designed to be useful, fun, and helpful and it isn’t motivated by stealing your data and selling you stuff.
There's so much more to say on this, but in the meantime so excited to get this out to communities and learn and improve our approach!
There's so much more to say on this, but in the meantime so excited to get this out to communities and learn and improve our approach!
December 1, 2025 at 10:53 PM
This isn’t another toxic, chaotic space — Roundabout is designed to be useful, fun, and helpful and it isn’t motivated by stealing your data and selling you stuff.
There's so much more to say on this, but in the meantime so excited to get this out to communities and learn and improve our approach!
There's so much more to say on this, but in the meantime so excited to get this out to communities and learn and improve our approach!
In practice, the cryptography is 2% of the problem for many applications I can think of, but it's fun & hard so that's where the bulk of the attention goes. I love a good cryptography problem as much as the next technologist, but rail against it because the other bits often get neglected.
December 1, 2025 at 7:40 PM
In practice, the cryptography is 2% of the problem for many applications I can think of, but it's fun & hard so that's where the bulk of the attention goes. I love a good cryptography problem as much as the next technologist, but rail against it because the other bits often get neglected.
💯💯💯
That's a big part of what I was pointing at. ❤️ The instinct in a lot of these standards circles is "I slapped some MLS on there, called it the Good Ship Private, job done!"
That's a big part of what I was pointing at. ❤️ The instinct in a lot of these standards circles is "I slapped some MLS on there, called it the Good Ship Private, job done!"
December 1, 2025 at 7:40 PM
💯💯💯
That's a big part of what I was pointing at. ❤️ The instinct in a lot of these standards circles is "I slapped some MLS on there, called it the Good Ship Private, job done!"
That's a big part of what I was pointing at. ❤️ The instinct in a lot of these standards circles is "I slapped some MLS on there, called it the Good Ship Private, job done!"
I've tried to accept it into my heart, but I don't think I'll ever come around to the accessible-but-encrypted model of security (assuming that's what you mean?)
December 1, 2025 at 5:49 AM
I've tried to accept it into my heart, but I don't think I'll ever come around to the accessible-but-encrypted model of security (assuming that's what you mean?)
Spoiler alert: I think we're just re-inventing mailing lists and web forums with better metadata and UX and hopefully maybe interop? 😂
December 1, 2025 at 1:43 AM
Spoiler alert: I think we're just re-inventing mailing lists and web forums with better metadata and UX and hopefully maybe interop? 😂
(this is because email administrators and standards wonks don't give a toss about user-level email concerns; they just want to know that the specifications and protocols that have been written prove what the specs and protocols encoded – this is a net loss for everyone)
December 1, 2025 at 1:31 AM
(this is because email administrators and standards wonks don't give a toss about user-level email concerns; they just want to know that the specifications and protocols that have been written prove what the specs and protocols encoded – this is a net loss for everyone)
Designing specs before having a clear understanding of the expected behaviour is not only unlikely to work, it risks entrenching undesirable systems (e.g., Did You Know that despite SPF, DKIM, DMARC, DANE, etc, etc, etc it's still not possible to tell if a given user sent an email? 🤪🔫)
December 1, 2025 at 1:31 AM
Designing specs before having a clear understanding of the expected behaviour is not only unlikely to work, it risks entrenching undesirable systems (e.g., Did You Know that despite SPF, DKIM, DMARC, DANE, etc, etc, etc it's still not possible to tell if a given user sent an email? 🤪🔫)
A spec doesn't imply interop. My hot take here: Interop means mutually intelligible and compatible behaviour, and doesn't exist without user-level understanding of the product-level intent of that interop.
December 1, 2025 at 1:31 AM
A spec doesn't imply interop. My hot take here: Interop means mutually intelligible and compatible behaviour, and doesn't exist without user-level understanding of the product-level intent of that interop.
But, scoping the community is useful ("who is a member?"). MLS and complicated key management is just way overkill for that until we have good user-centric flows described that we can encode into simple tools.
December 1, 2025 at 1:31 AM
But, scoping the community is useful ("who is a member?"). MLS and complicated key management is just way overkill for that until we have good user-centric flows described that we can encode into simple tools.
There are big open questions about what "privacy" means in these contexts. For us, where we're aiming to have 5-50k member groups, what does MLS even mean in terms of "privacy"? If anyone can join: it means nothing.
December 1, 2025 at 1:31 AM
There are big open questions about what "privacy" means in these contexts. For us, where we're aiming to have 5-50k member groups, what does MLS even mean in terms of "privacy"? If anyone can join: it means nothing.
We're building on atproto primitives but off-network because it's currently not possible to push private/scoped data around the wider atproto network. I hope it will be possible sooner than later!
December 1, 2025 at 1:31 AM
We're building on atproto primitives but off-network because it's currently not possible to push private/scoped data around the wider atproto network. I hope it will be possible sooner than later!
Nice! It looks like docling is probably the answer for local pdf ingestion, could be worth integrating! I might take a poke at some point.
November 27, 2025 at 5:50 PM
Nice! It looks like docling is probably the answer for local pdf ingestion, could be worth integrating! I might take a poke at some point.
One of the four fundamental building blocks of atproto, in addition to the lowlike, the follike, and the likefol. Each attiquark comes in a strange, charmed, up, down, top, bottom, cis, and trans variant and together they combine in many configurations to form what activitypubists call "noticules"
November 20, 2025 at 7:11 PM
One of the four fundamental building blocks of atproto, in addition to the lowlike, the follike, and the likefol. Each attiquark comes in a strange, charmed, up, down, top, bottom, cis, and trans variant and together they combine in many configurations to form what activitypubists call "noticules"
Those of us affected and our neighbours who are at higher risk following the fires have been left to self-organize and are applying for the same streams of funding as everyone else in the area. I'd like to say I'm not cynical about disaster management, but ...
November 20, 2025 at 4:08 PM
Those of us affected and our neighbours who are at higher risk following the fires have been left to self-organize and are applying for the same streams of funding as everyone else in the area. I'd like to say I'm not cynical about disaster management, but ...
Not massively surprising, given that a year on from the Slocan fires, there has been $0 of funding to help support folks who were directly impacted (possibly the first residences directly affected by forest fire in the RDCK, but not the last) and no meaningful policy introspection.
November 20, 2025 at 4:08 PM
Not massively surprising, given that a year on from the Slocan fires, there has been $0 of funding to help support folks who were directly impacted (possibly the first residences directly affected by forest fire in the RDCK, but not the last) and no meaningful policy introspection.
Deeply agree - it's unavoidable to me how impactful this technology will be, in all the possible connotations of "impactful."
I'm deeply grateful to work with very thoughtful people, and this is the AI policy we've settled on for now, which tries to hold the contradictions: newpublic.org/AI-policy
I'm deeply grateful to work with very thoughtful people, and this is the AI policy we've settled on for now, which tries to hold the contradictions: newpublic.org/AI-policy
Our Approach to Artificial Intelligence
We are experimenting with using AI tools to extend our work as a small nonprofit, so that we can focus our time on reinforcing human connections, conversations, and communities that have eroded.
newpublic.org
November 20, 2025 at 5:56 AM
Deeply agree - it's unavoidable to me how impactful this technology will be, in all the possible connotations of "impactful."
I'm deeply grateful to work with very thoughtful people, and this is the AI policy we've settled on for now, which tries to hold the contradictions: newpublic.org/AI-policy
I'm deeply grateful to work with very thoughtful people, and this is the AI policy we've settled on for now, which tries to hold the contradictions: newpublic.org/AI-policy