Why 90% of AI Governance Frameworks Will Fail by 2027
Most governments are building AI governance frameworks that will be obsolete before they launch. Here's why—and what actually works.
Mr. Qizhi
AI, Urban Planning & GovTech — practical frameworks and delivery systems.
Start here
AI-Era Cities Must Be Reshaped: From Projects to Operations, From Spending to Delivery
A strategic note for city managers: in the AI era, the advantage is no longer building more projects. It is operating reliable delivery systems, building sustainable monetization beyond land finance, and improving resident services.
2026：科技界的临界点与能力再分配
当知识的成本趋近于零，创造的门槛消失，旧秩序的崩塌不再是灾难，而是机遇。Brian Norgard、Jimmy Ba等科技领袖对2026年的洞察。
When AI Makes Thinking Cheap, Scarcity Moves to Energy, Land, and Materials
AI will crush the cost of information work, but it won’t make the physical world free. In 2026, scarcity shifts from labor and knowledge to energy, land, minerals, infrastructure, and the systems that turn bits into atoms.
数字孪生城市的未来：从“看见城市”到“推演城市” | 弃知先生
数字孪生不是更精致的三维模型，而是一套能与真实城市同步、可推演政策与风险的“城市操作系统”。本文用真实治理场景讲清它如何落地：交通、洪水、能源、审批与公众参与；也坦诚讨论隐私、成本、技术债与组织协同等取舍，并给出面向2026的实施路径。
凝固在代码里的回音：一场五大AI模型的文学创作盲测
当ChatGPT 5.1、Claude 4.6 Sonnet、Kimi k2.5、DeepSeek R1与Gemini 3.1 Pro同场竞技，谁写出了最像人类的散文？一场关于AI文学灵魂的深度实验。
Topics
Latest
The Hidden Tax of a Chokepoint: How Hormuz Turns Into Inflation
When Hormuz stalls, the real shock is not just oil; it is a freight, inventory, and policy-expectations chain that quietly taxes consumers and destabilizes macro narratives.
霍尔木兹海峡风险：当航运通道变成通胀触发器
霍尔木兹海峡的关键不在"是否真封锁"，而在"风险溢价能否持续"，它会把局部冲突直接传导成全球通胀与利率路径变化。
Cloud Partnerships Aren’t Gossip. They’re a City-Scale AI Supply Chain.
For a city, a cloud partnership isn’t vendor gossip. It’s the choice of a supply chain you can actually cash in over years: peak inference, disaster recovery, data sovereignty, unified audit, and contract-grade SLAs.
云合作不是八卦，是城市级 AI 供给链
城市要做 AI，云合作不是选供应商的八卦，而是选一条可长期兑现的供给链。关键是推理峰值、灾备、数据主权、审计统一与可验收的服务级别。
City-Scale AI Isn’t a Model Problem
Scaling AI in a city isn’t a model problem. It’s a city-level supply and institutional interface problem: compute like electricity, data like water, services like roads, governance like traffic rules.
城市的 AI 规模化，拼的不是模型
城市的 AI 规模化不是模型问题，而是城市级供给与制度接口问题。算力像电，数据像水，服务像路，治理像交通规则。
Done Means the Link Works
A day of small failures taught me a simple rule: if you can’t verify it where the user touches it, it’s not done. No templates, no checklists—just the muscle of finishing.
OpenClaw Isn’t a Monster. It’s an Accelerator — Remember How Cars Felt in 1926
A small sci‑fi memory from June 2028, looking back at June 2026: why people fear new agents like OpenClaw, and why the real answer is guardrails, accountability, and systems—not superstition.
OpenClaw 不是妖怪，它是变革的助推器：从汽车初现时的恐惧谈起 | 弃知先生
用一个2028年的回忆场景，回看2026年OpenClaw出现时的社会担忧：安全、失业与秩序。技术不是妖怪，真正决定走向的是护栏、责任与可审计的系统。
The Long Shadow of Trump's Tariffs: Reshaping Global Supply Chains
An analysis of how the latest sweeping U.S. tariffs are forcing a fundamental and lasting restructuring of global trade and manufacturing networks.
除夕这一天，城市像一台“暂时慢下来”的机器（以及这对治理意味着什么） | 弃知先生
除夕不是只有烟花与团圆，它也是城市运行的一次“压力测试”：人口流动、交通供给、社区互助、应急响应都会变得不一样。用更像人写的方式，聊聊除夕的城市节律，以及我们能从中学到的治理与韧性。
How I’d Set Up AI Governance in Digital Government (2026) — Without Killing Delivery
A practitioner’s guide to AI governance in digital government (2026): how to ship citizen-facing AI with clear decision rights, minimum controls, procurement clauses, and monitoring that actually works.
Gov-Tech and Blockchain in 2026: Where It Works (and Where It Doesn’t)
A practical, narrative look at blockchain in government: real deployments (Estonia, Georgia land registry, China e-invoicing, Dubai), common pitfalls, procurement questions, and metrics to prove value.
数字孪生城市：别再做大屏展示了，先学会“演练再上线”（2026） | 弃知先生
一篇更像从业者复盘的数字孪生城市指南：从暴雨应急的真实场景讲起，解释数字孪生到底解决什么、落地路线怎么走、成本与风险在哪里，以及 2026 年最稳的实施路径。
Digital Twin Cities: What They Really Change (and How to Build One)
A practical, human guide to digital twin cities in 2026: what they are, what real cities (Singapore, Helsinki, Rotterdam, Dubai) are doing, and how to implement one step-by-step.
NVIDIA Cosmos技术重塑国土空间规划 | 弃知先生
把城市规划从‘画图+评审’推进到‘可交互的数字孪生试验场’：本文用具体案例解释NVIDIA Cosmos（面向物理世界的生成式/仿真式AI能力）在国土空间规划里的可用之处、边界与落地路径。
2026年Gov-Tech创新趋势（更像一份落地指南） | 弃知先生
2026 年 Gov-Tech 不再只是“上新技术”，而是围绕办事体验、数据治理与安全合规做系统升级。本文用 5 个趋势 + 真实场景式案例 + 落地步骤，帮你把 AI、公民参与、数字孪生与区块链等能力真正用起来。
Gov-Tech Procurement: How to Buy Technology That Actually Works
Most Gov-Tech failures start in procurement. Here's a practical guide to writing RFPs that force delivery discipline, avoid vendor lock-in, and ensure public money buys outcomes—not just promises.
Gov-Tech创新：从“做项目”到“交付能力” | 弃知先生
一线视角复盘 Gov-Tech：不是缺平台，而是缺交付与运营。给出三条可落地路径：以用户时间成本为指标、以交付闭环为方法、以数据治理与合规为底座。
数字化转型：城市进化的必经之路 | 弃知先生
一座城市的数字化转型，不是买一套平台，而是把治理从“靠人盯”变成“靠机制跑”。从失败的“大屏冲动”到可复用的治理闭环，本文用一条真实的城市改造弧线讲清：哪些做法会翻车，哪些能落地；要看哪些 KPI；以及如何用数据治理与组织治理把项目变成能力。
DeepSeek：把大模型放进城市规划工作流 | 弃知先生
一位规划师把 DeepSeek 当作“会写会归纳的同事”：从材料梳理、诉求提炼、方案草案到公众意见汇总，用一套可复用的提示词与校核清单，把大模型接进规划工作流，同时把幻觉、偏见与合规风险关在门外。
AI in Urban Planning in 2026: From “Smart City” Theater to Useful Citymaking
In 2026, AI is finally becoming practical in urban planning: faster options, better forecasts, and fewer compliance mistakes—if we treat it as a co-pilot, not an oracle. An opinionated look at what works, what fails, and real examples planners can use.
2026年AI在城市规划中的十大应用趋势 | 弃知先生
站在一线规划与Gov-Tech落地的视角，梳理2026年AI在城市规划中的10个最实用趋势：从生成式方案、数字孪生、交通与环境优化，到自动合规与人机协同，并给出可直接套用的落地框架与评估指标。