天气块报
屏幕、定制区块链网络、定制软件、定制电路、霓虹灯、电脑、树莓派微型单板电脑、钢架
2022/2025
围绕气象数据的争议近年来反复出现：2008年，美国驻京大使馆发布的自测空气质量数据因与官方数据相悖，引发了关于监测权与数据解释权的长期讨论；2019年，美国“记号笔门”（Sharpiegate）事件中，官方飓风预测图被黑色记号笔手工篡改，试图以行政意志修正科学预测；2022年夏天，中国多地遭遇极端高温，手机 App 显示的数值与官方通报之间的落差，使气象数据在社交媒体传播与官方辟谣之间反复拉扯。这些争议揭示了气象数据在自然指标之外的社会属性：它会直接触发一系列面向现实的阈值判断：风险预警是否启动、公共措施是否生效、责任与补偿如何被认定。这些冲突的核心在于：当同一时间、同一地点存在多个版本的数据时，谁有权决定哪个版本进入记录？而进入记录、被采信的那一个，往往就成了制度与记忆里的事实。
《天气块报》以一座形如天线塔的装置呈现，由三组计算设备构成的区块链系统、屏幕与霓虹灯组成。作品在每个整点从三个不同的数据源抓取同一城市的气象信息，而得到的三份气象信息往往并不一致。装置内部运行一套采用权益证明（PoS）的区块链机制：每到整点，系统依据 PoS 的随机抽取逻辑在三个节点中选出当轮写入者；被选中的节点所抓取的那一份气象信息将被写入链上并永久留存，并在屏幕上显示，同时塔顶对应的天气符号霓虹灯随之点亮。作品并不试图比较不同版本数据的准确性，而是把争议的重心从“数据准确性”推向“事实化机制”：气象数据作为不可复验的时间性数据，事后难以“回到当时”重新核验，多版本并存时，关键往往不在于反复追问哪个版本更“准确”，而在于哪个版本会被记录、被追溯、被引用，并在治理与责任判定中被正式采信——从而成为行动与责任的依据。由此将‘记录’转化为一种历史书写。
区块链以去中心化与共识记录为特征，常被看作一种可验证的记录方式。但在现实中，数据可及性与算法处理路径存在结构性差异：不同平台能接入的底层数据不同，采用的同化、修正与预报算法也不同；由此生成的不同版本在发布渠道与传播路径中的权重亦不相同，其中某些版本更容易进入权威口径、被平台分发，并成为默认引用。作品由此追问：当区块链的数据记录功能在特定领域被引入为技术基础设施时，所谓的可信记录机制是否会把偏差固化为长期生效的事实，并为数据集权提供新的正当性？
《天气块报》作品首版由毕昕于北京现代汽车文化中心展览《飞出个未来：区块链中的时间多重性》（2022）委任创作。 其后受大馆当代美术馆支持，于展览《保持在线：云中游荡》（2025）中进行了区块链网络软件与装置硬件结构升级。
软件开发： 陈立立
特别致谢： 毕昕， 姚翔， 原语里弄
The Weather Consensus
Screens, Custom blockchain network, Custom software, Custom circuits, Neon lights, Computers, Raspberry Pi, Steel frames
2022/2025
Disputes over meteorological data have resurfaced again and again in recent years. In 2008, the U.S. Embassy in Beijing released its own air-quality measurements that conflicted with Chinese official figures, prompting long-running debates over the right to monitor and the right to interpret data. In 2019, during the U.S. “Sharpiegate” incident, an official hurricane forecast map was manually altered with a black marker—an attempt to bend scientific prediction to administrative will. In the summer of 2022, amid extreme heat across multiple regions of China, gaps between readings shown in mobile weather apps and official bulletins produced a recurring tug-of-war between social-media sharing and official rebuttals. These disputes reveal meteorological data’s social character beyond natural indicators: it can trigger real-world threshold decisions—whether alerts are issued, whether public measures take effect, and how responsibility and compensation are determined. What these conflicts share is a single question: when multiple versions of data exist for the same place at the same time, who gets to decide which version enters the record? And the version that is recorded and accepted often becomes the “fact” within institutions and collective memory.
The Weather Consensus takes the form of an antenna-like tower housing a three-node blockchain system, screens, and neon lights. At the top of every hour, the work retrieves weather information for the same city from three different sources—yet the three sets of readings often do not match. The tower houses a blockchain mechanism based on Proof of Stake (PoS): each hour, the system uses protocol-generated randomness under PoS to select one node as the writer. The selected node’s weather data is then recorded on the chain and permanently stored; it is displayed on the screen, while the corresponding neon weather icon at the top lights up. Rather than adjudicating accuracy, the work shifts the dispute from “data accuracy” to the mechanics of fact-making. When multiple versions coexist, the key issue is not which is “truer,” but which version will be recorded, remain traceable, be cited, and be formally accepted in governance and liability determinations—thus becoming the basis for action and responsibility. In this way, recording is turned into a mode of historical writing.
Blockchain, characterized by decentralization and consensus-based recording, is often framed as a verifiable way to keep records. Yet in practice, data accessibility and algorithmic processing paths differ structurally: platforms draw on different underlying datasets and apply different assimilation, correction, and forecasting algorithms. The resulting versions also carry unequal weight across channels of publication and pathways of circulation; some are more likely to enter authoritative narratives, be distributed by platforms, and become default references. The work therefore asks: when blockchain’s data-recording function is introduced as technical infrastructure in specific domains, can this so-called trusted-record mechanism solidify bias into long-term effective “facts”—and provide new legitimacy for data centralization?
The first version of The Weather Consensus was commissioned for BI Xin’s exhibition Time After Time: The Polychronicity in Blockchain (2022) at Hyundai Motorstudio Beijing. It was later supported by Tai Kwun Contemporary and underwent upgrades to its blockchain network software and installation hardware structure for the exhibition Stay Connected: Navigating the Cloud (2025).
Software development: CHEN Lili
Special Thanks to BI Xin, YAO Xiang, and Primitives Lane.
--
Initial version of the work from 2022.