Anyway, how exactly do you plan on deterring astroturfing and other sorts of public opinion manipulation? Heck, how do you convince us that you are not the one working on an influence campaign yourself?
Astroturfing is actually one of the problems I’m trying to solve. Civie has optional identity verification (via Persona), and anonymous responses can be filtered down to verified-only. The idea is to make participation easier while still giving people a way to look at results that are harder to game, with everything aggregated and publicly inspectable.
On the “influence campaign” part, I’m honestly just trying to get feedback during an early beta, not shape anyone’s opinions. And for trust more broadly, you probably shouldn’t just take my word for it. The only real answer is transparency over time. That’s why I’m attempting to build this in public and gather as much feedback as I can.
The page has very little about how it works (security / trust / identity), there is nothing about tech stack or open source. These are going to be paramount in any civics platform, b/c transparency and the trust-pocolipse.
It’s currently built with Next.js and Firebase. Responses are stored separately from identity data, there are no public profiles, and only aggregate results are exposed in the UI. Verification is used to increase resistance to bots, but answers are not tied to public identity.
I agree that any civic platform lives or dies on transparency right now. I’m actively considering showing the question schemas and aggregation logic so the mechanics are inspectable rather than opaque.
If you have strong opinions on what absolutely must be open versus what can reasonably remain closed, I’d genuinely value that perspective. This is still early and the landing page is pretty barebones, but the core web app and data model are where most of the effort has gone so far.
https://discord.atprotocol.dev (should work, best entrypoint to the dev community)
I like sharing this one for new devs: https://atproto.com/articles/atproto-for-distsys-engineers
This is increasingly looking like the tell tale for Ai bots masquerading as humans. I've heard essentially this phrase as the first in several recent interactions which ended up being bots.
Can you prove your humanity? Bots and agents are not allowed on HN
it's good vocab and writing to have! I wonder how much longer we will have easy indicators
The answer-before-results flow was intentional to avoid anchoring, and the skip option is there to keep participation low pressure.
On accuracy, I’m a bit cautious. Since it’s self selected, it won’t be statistically representative. The hope is that with enough diverse participation, it can still produce meaningful directional signal.
That tradeoff between accessibility and representativeness is basically the core experiment.