Thank you for not writing the easy promotional version of this. You took on the hard question, and the piece is better for it.
The "hollow" reaction to my broadcast is fair, and it's the right instinct. A fully automated summary is a starting line, not a finish line. The version of CivicDigest I want to exist isn't a robot anchor replacing a reporter — it's plumbing. It does the translation busywork of turning 80-page PDFs into plain English so the people who do the real work (e.g. local journalists, civic organizers, the ones who actually show up to the meeting) get their time back for judgment, presence, and accountability. The machine handles legibility. People handle meaning.
Richard Young names the open problem exactly: these tools can convey information, but they can't build trust or do the hard work of institutional reform on their own. I agreed, and that's what I'm building toward, not pretending I've solved.
I'm genuinely glad to be part of this conversation alongside SeeGov, Aware, CivicLex, and folks like Beth Noveck. The question I keep coming back to: where does automation end and human editorial judgment have to begin? Curious where others here land on it.
My fall ballot will field candidates for 80-100 local judicial races; I suspect that most all votes are for a party, not a candidate, as there is little information available for most all of them.
Indeed! We democrats should also notify people of where to vote so that what happened in Dallas this year doesn't happen again. The capital-D party is based in Dallas; its senior leadership is from there, as was one of the candidates. The voting procedures were changed in late December. Yet no one got the word out. (I did get something in the mail from a Houston group a week after the election. But I credit the thought.)
Thank you for not writing the easy promotional version of this. You took on the hard question, and the piece is better for it.
The "hollow" reaction to my broadcast is fair, and it's the right instinct. A fully automated summary is a starting line, not a finish line. The version of CivicDigest I want to exist isn't a robot anchor replacing a reporter — it's plumbing. It does the translation busywork of turning 80-page PDFs into plain English so the people who do the real work (e.g. local journalists, civic organizers, the ones who actually show up to the meeting) get their time back for judgment, presence, and accountability. The machine handles legibility. People handle meaning.
Richard Young names the open problem exactly: these tools can convey information, but they can't build trust or do the hard work of institutional reform on their own. I agreed, and that's what I'm building toward, not pretending I've solved.
I'm genuinely glad to be part of this conversation alongside SeeGov, Aware, CivicLex, and folks like Beth Noveck. The question I keep coming back to: where does automation end and human editorial judgment have to begin? Curious where others here land on it.
thank you Brandi! The "plumbing vs replacement" framing is a very useful distinction. Looking forward to where CivicDigest (and you!) go next.
Add that to the catalog. Want to be the catalog keeper?
The Athenians did pretty ok without it.
My fall ballot will field candidates for 80-100 local judicial races; I suspect that most all votes are for a party, not a candidate, as there is little information available for most all of them.
This seems like a solvable problem in age of AI. We should make a catalog of these solvable problems
Indeed! We democrats should also notify people of where to vote so that what happened in Dallas this year doesn't happen again. The capital-D party is based in Dallas; its senior leadership is from there, as was one of the candidates. The voting procedures were changed in late December. Yet no one got the word out. (I did get something in the mail from a Houston group a week after the election. But I credit the thought.)