#general
2026-06-01
@munusshih has joined the channel
@d.rathwell has joined the channel
我想有多餘 AI token 的人是存在的，只是貢獻的pipeline需要順暢
可以試著
1. 將如何貢獻寫在專案的README.md中
2. 把近端開發的流程友善化。通常是不需要部署，可以在近端一鍵跑動localhost來開發。
3. 議題切細，利於貢獻者發PR。
可以試著
1. 將如何貢獻寫在專案的README.md中
2. 把近端開發的流程友善化。通常是不需要部署，可以在近端一鍵跑動localhost來開發。
3. 議題切細，利於貢獻者發PR。
Replied to a thread: 2026-05-31 22:08:30
先把腦袋裡的東西實作了一個版本出來：https://g0v.hackmd.io/@Pno233SAS8G5UfL5OvSRmA/Bk0uRboxzg
HackMD
# TokenHunt: 你多餘的 token 可能某個專案的解方 > 把你多餘的 AI Token，貢獻給需要的公民科技專案。 --- ### 想要貢獻專案？ - 歡迎在共筆上加入討論，或者
如果要用閒置 token 產生 markdown 相對風險小，但假如要外包寫 code 這件事，要擔心有人表面上貢獻 token 解 issue 寫 code ，實際上偷埋後門的風險，而且這些不一定是靠 review 就可以抓的出來的
外包 token 寫 markdown 其實也還是有風險，就是他也有可能竄改 prompt 或是 input data ，來刻意產生有偏差的結果
所以要做群眾外包有個前提是這些東西要有簡易的方式可檢驗
所以要做群眾外包有個前提是這些東西要有簡易的方式可檢驗
了解！我來想想看可以如何設計機制
@2017xiao06 has joined the channel
2026-06-02
@isaac105 has joined the channel
Hello everyone,
I would like to present a project that I have dedicated to the *PUBLIC DOMAIN*, which I believe could help us make civic participation tools more scalable. As the author, I outline my reasoning below. Thank you.
"In political elections, candidacies that do not monopolize the media do not exist for the electorate and do not win, resulting in the entrenchment of a tiny elite. In the proposed system, however, each candidate is given the same initial visibility, preventing favorites from monopolizing votes from the outset."
Polivoz is a *PUBLIC DOMAIN* design that combines pairwise comparison (A vs. B) with a minimalist algorithm (without AI) to identify consensus in large groups.
Pairwise comparison involves repeatedly making binary judgments (A vs. B) where voters express their preference for one option over another (A or B). This technique allows for the efficient and fair ordering of thousands or millions of options, without requiring voters to be informed about all of them at once (typically, they are only asked about around 30).
If the algorithm grants the same initial visibility to each proposal, inequalities stemming from variables external to the voting process—which are highly difficult to control—are significantly reduced.
Furthermore, by rewarding options that garner more consensus with higher visibility, while minimizing the exposure of less appealing ones, high-quality rankings are achieved in the top tier.
In computer simulations, the minimalist Polivoz algorithm demonstrates its potential to identify the highest-consensus proposal among 100,000 options, with success rates ranging from 45% to 77% depending on the research scenario.
In the context of civic technology, I believe that this pairwise comparison (A vs. B) functionality with a meritocratic matching algorithm could significantly increase a platform's capacity to identify the highest-consensus options in extreme use cases involving thousands of proposals.
https://doi.org/10.5281/zenodo.18703650 [Spanish]
I would like to present a project that I have dedicated to the *PUBLIC DOMAIN*, which I believe could help us make civic participation tools more scalable. As the author, I outline my reasoning below. Thank you.
"In political elections, candidacies that do not monopolize the media do not exist for the electorate and do not win, resulting in the entrenchment of a tiny elite. In the proposed system, however, each candidate is given the same initial visibility, preventing favorites from monopolizing votes from the outset."
Polivoz is a *PUBLIC DOMAIN* design that combines pairwise comparison (A vs. B) with a minimalist algorithm (without AI) to identify consensus in large groups.
Pairwise comparison involves repeatedly making binary judgments (A vs. B) where voters express their preference for one option over another (A or B). This technique allows for the efficient and fair ordering of thousands or millions of options, without requiring voters to be informed about all of them at once (typically, they are only asked about around 30).
If the algorithm grants the same initial visibility to each proposal, inequalities stemming from variables external to the voting process—which are highly difficult to control—are significantly reduced.
Furthermore, by rewarding options that garner more consensus with higher visibility, while minimizing the exposure of less appealing ones, high-quality rankings are achieved in the top tier.
In computer simulations, the minimalist Polivoz algorithm demonstrates its potential to identify the highest-consensus proposal among 100,000 options, with success rates ranging from 45% to 77% depending on the research scenario.
In the context of civic technology, I believe that this pairwise comparison (A vs. B) functionality with a meritocratic matching algorithm could significantly increase a platform's capacity to identify the highest-consensus options in extreme use cases involving thousands of proposals.
https://doi.org/10.5281/zenodo.18703650 [Spanish]
- 👀2
Hello everyone,
I would like to present a project that I have dedicated to the *PUBLIC DOMAIN*, which I believe could help us make civic participation tools more scalable. As the author, I outline my reasoning below. Thank you.
"In political elections, candidacies that do not monopolize the media do not exist for the electorate and do not win, resulting in the entrenchment of a tiny elite. In the proposed system, however, each candidate is given the same initial visibility, preventing favorites from monopolizing votes from the outset."
Polivoz is a *PUBLIC DOMAIN* design that combines pairwise comparison (A vs. B) with a minimalist algorithm (without AI) to identify consensus in large groups.
Pairwise comparison involves repeatedly making binary judgments (A vs. B) where voters express their preference for one option over another (A or B). This technique allows for the efficient and fair ordering of thousands or millions of options, without requiring voters to be informed about all of them at once (typically, they are only asked about around 30).
If the algorithm grants the same initial visibility to each proposal, inequalities stemming from variables external to the voting process—which are highly difficult to control—are significantly reduced.
Furthermore, by rewarding options that garner more consensus with higher visibility, while minimizing the exposure of less appealing ones, high-quality rankings are achieved in the top tier.
In computer simulations, the minimalist Polivoz algorithm demonstrates its potential to identify the highest-consensus proposal among 100,000 options, with success rates ranging from 45% to 77% depending on the research scenario.
In the context of civic technology, I believe that this pairwise comparison (A vs. B) functionality with a meritocratic matching algorithm could significantly increase a platform's capacity to identify the highest-consensus options in extreme use cases involving thousands of proposals.
https://doi.org/10.5281/zenodo.18703650 [Spanish]
I would like to present a project that I have dedicated to the *PUBLIC DOMAIN*, which I believe could help us make civic participation tools more scalable. As the author, I outline my reasoning below. Thank you.
"In political elections, candidacies that do not monopolize the media do not exist for the electorate and do not win, resulting in the entrenchment of a tiny elite. In the proposed system, however, each candidate is given the same initial visibility, preventing favorites from monopolizing votes from the outset."
Polivoz is a *PUBLIC DOMAIN* design that combines pairwise comparison (A vs. B) with a minimalist algorithm (without AI) to identify consensus in large groups.
Pairwise comparison involves repeatedly making binary judgments (A vs. B) where voters express their preference for one option over another (A or B). This technique allows for the efficient and fair ordering of thousands or millions of options, without requiring voters to be informed about all of them at once (typically, they are only asked about around 30).
If the algorithm grants the same initial visibility to each proposal, inequalities stemming from variables external to the voting process—which are highly difficult to control—are significantly reduced.
Furthermore, by rewarding options that garner more consensus with higher visibility, while minimizing the exposure of less appealing ones, high-quality rankings are achieved in the top tier.
In computer simulations, the minimalist Polivoz algorithm demonstrates its potential to identify the highest-consensus proposal among 100,000 options, with success rates ranging from 45% to 77% depending on the research scenario.
In the context of civic technology, I believe that this pairwise comparison (A vs. B) functionality with a meritocratic matching algorithm could significantly increase a platform's capacity to identify the highest-consensus options in extreme use cases involving thousands of proposals.
https://doi.org/10.5281/zenodo.18703650 [Spanish]
2026-06-03
chewei 哲瑋 _
8/8 週六 COSCUP _ AI 開放治理、AI x Civic Tech 議程概況
https://pretalx.coscup.org/coscup-2026/schedule/
【AI 開放治理】Room_TR411_9:30-12:00
[9:30-10:20] AI x Open Source: Global Trends & Policy Landscape｜人工智慧與開源的對話：國際趨勢與政策發展｜Speaker: 數位發展部 Ministry of Digital Affairs
[10:20-10:50] Open-Source AI: Definition, Risk and Governance｜開源人工智慧的定義、風險與治理｜Speaker: Cui Jia Wei
[10:50-11:00] Break｜休息時間
[11:00-11:30] What did the UN say about Open Source x AI?｜關於開放原始碼 x 人工智慧，聯合國怎麼說？｜Speaker: RR
[11:30-12:00] Towards Trustworthy AI: Essential Open-Source Tools for Agent Bias Detection and Mitigation｜邁向可信任 AI：開發者必備的 Agents 偏見檢測與緩解框架｜Speaker: Jing-Tian Sung
[12:00-13:00] Break｜休息時間
【AI x Civic Tech】Room_TR411_13:00-16:30
[13:00-13:30] Open source, open voice: the power of FOSS in supporting citizen participation｜Speaker: Lucía Luzuriaga
[13:30-14:00] Engineer Cafe's OSS Reception System: A Reproducible Model for Local-Global Collaboration｜Speaker: Kosuke Shimada (Engineer Cafe Community manager)
[14:00-14:30] AI Sovereignty in Practice: From Taiwan Benchmark to Production Payment Agents｜Speaker: 蕭新晟
[14:30-14:40] Open Discussion｜綜合交流
[14:40-14:50] Break｜休息時間
[14:50-15:20] Citizen Development Governance in the Era of Generative AI: From Business Innovation to Corporate Management｜生成式 AI 時代的公民開發治理：從業務創新到企業納管｜Speaker: Teemo
[15:20-15:50] Returning Rescuers to the Frontline: AI-Assisted Development and Open Public-Private Collaboration for Fire Safety Equipment Smart Scheduling System｜讓人力回歸前線：智慧排程系統的 AI 輔助開發與開源公私協力實錄｜Speaker: Josh (杜佳憲)
[15:50-16:20] Building an Open-Source Labor Observatory: Using n8n and LLMs to Track Skill Drift and AI Employment Risks｜打造開源勞動觀測站：用 n8n 與 LLM 解析 AI 時代的職能轉變與就業衝擊｜Speaker: 林源隆（Wilson）
[16:20-16:30] Open Discussion｜綜合交流
8/8 週六 COSCUP _ AI 開放治理、AI x Civic Tech 議程概況
https://pretalx.coscup.org/coscup-2026/schedule/
【AI 開放治理】Room_TR411_9:30-12:00
[9:30-10:20] AI x Open Source: Global Trends & Policy Landscape｜人工智慧與開源的對話：國際趨勢與政策發展｜Speaker: 數位發展部 Ministry of Digital Affairs
[10:20-10:50] Open-Source AI: Definition, Risk and Governance｜開源人工智慧的定義、風險與治理｜Speaker: Cui Jia Wei
[10:50-11:00] Break｜休息時間
[11:00-11:30] What did the UN say about Open Source x AI?｜關於開放原始碼 x 人工智慧，聯合國怎麼說？｜Speaker: RR
[11:30-12:00] Towards Trustworthy AI: Essential Open-Source Tools for Agent Bias Detection and Mitigation｜邁向可信任 AI：開發者必備的 Agents 偏見檢測與緩解框架｜Speaker: Jing-Tian Sung
[12:00-13:00] Break｜休息時間
【AI x Civic Tech】Room_TR411_13:00-16:30
[13:00-13:30] Open source, open voice: the power of FOSS in supporting citizen participation｜Speaker: Lucía Luzuriaga
[13:30-14:00] Engineer Cafe's OSS Reception System: A Reproducible Model for Local-Global Collaboration｜Speaker: Kosuke Shimada (Engineer Cafe Community manager)
[14:00-14:30] AI Sovereignty in Practice: From Taiwan Benchmark to Production Payment Agents｜Speaker: 蕭新晟
[14:30-14:40] Open Discussion｜綜合交流
[14:40-14:50] Break｜休息時間
[14:50-15:20] Citizen Development Governance in the Era of Generative AI: From Business Innovation to Corporate Management｜生成式 AI 時代的公民開發治理：從業務創新到企業納管｜Speaker: Teemo
[15:20-15:50] Returning Rescuers to the Frontline: AI-Assisted Development and Open Public-Private Collaboration for Fire Safety Equipment Smart Scheduling System｜讓人力回歸前線：智慧排程系統的 AI 輔助開發與開源公私協力實錄｜Speaker: Josh (杜佳憲)
[15:50-16:20] Building an Open-Source Labor Observatory: Using n8n and LLMs to Track Skill Drift and AI Employment Risks｜打造開源勞動觀測站：用 n8n 與 LLM 解析 AI 時代的職能轉變與就業衝擊｜Speaker: 林源隆（Wilson）
[16:20-16:30] Open Discussion｜綜合交流
單純看到資訊分享
提供Github repo 有機會申請 OpenAI 送 6 個月ChatGPT Pro會員～（如果有發錯頻道再跟我說）
https://www.koc.com.tw/archives/644591
提供Github repo 有機會申請 OpenAI 送 6 個月ChatGPT Pro會員～（如果有發錯頻道再跟我說）
https://www.koc.com.tw/archives/644591
單純看到資訊分享
提供Github repo 有機會申請 OpenAI 送 6 個月ChatGPT Pro會員～（如果有發錯頻道再跟我說）
https://www.koc.com.tw/archives/644591
提供Github repo 有機會申請 OpenAI 送 6 個月ChatGPT Pro會員～（如果有發錯頻道再跟我說）
https://www.koc.com.tw/archives/644591
@az31402 has joined the channel
2026-06-06
g0v-digiresi.kktix.cc
只要你對台灣網路韌性的議題感興趣，都歡迎來參加！ 一起來動腦動手，提案協作來提升台灣的網路韌性吧！
- 💡1