信息强度 / INFORMATION INTENSITYSTATION LOG
如果用信息量、信息流、信息强度和曝光率来看这个世界——我们关注的，究竟是世界本身，还是关于世界的信息强度？
If you measure the world in volume, flow, intensity and exposure — are we watching the world itself, or the intensity of the information about it?
01容器The Container
当 ESG、Sustainability、Carbon 成为热点的时候，你去看各个会议、各个报道，到处都是。很多人一听，其实是 sales，是 consulting，是站在台上讲台下的事。而那些真正在干活的人，在水深火热之中，越干发掘的问题越多，反倒都不太说。
我在一次朋友小范围聚会时开了个玩笑：你们有没有发现，到处说这些的，如果用数据说话，他们都不是主角。主角是 oil & gas、steel & cement。你听到他们说了吗？这才是狠的地方，需要更多的关注和推动。
我不是说 noise 太多。我说的是一种挤兑——热点会挤占容量。一个要在 2035 年发挥作用的重工业转型，可能在 2026 年几乎没有讯息。而转型不是到了某个未来时间点，一下就能发生的。
那些真正干活的人，越干发掘的情况越多，反倒都不太说。为什么建这个 Station
这一页不挑战任何机构、顾问公司或媒体。它只想问一件事：AI 带给我们最不一样的东西，是不是"可以用数据来理解世界，而不是看谁更热卖、谁更热闹"。
When ESG, sustainability and carbon became hot, every conference and every report was full of them. Listen closely and much of it was sales, or consulting, or people on a stage describing work happening somewhere else. Meanwhile the people actually doing the work — deep in the heat of it — kept finding more problems, and said the least.
At a small gathering with friends I made a joke: notice that the ones talking everywhere are not, by the data, the protagonists. The protagonists are oil & gas, steel & cement. Have you heard from them? That is where it gets hard, and that is where attention and push are needed.
I am not saying there is too much noise. I am saying there is a crowding-out: a hot topic consumes capacity. A heavy-industry transition meant to deliver by 2035 can carry almost no signal in 2026 — and transitions do not arrive all at once when the date is reached.
The people doing the real work find more problems the deeper they go — and they say the least.Why this station exists
This page challenges no institution, consultancy or newsroom. It asks one thing: is the most different gift of AI that we can finally read the world with data, instead of reading whoever sells hardest and sounds loudest?
02信息强度 TOP 10Information Intensity, Top 10
下表是一个指示性指数（0–100，以 2024 年至 2025 年上半年的公开代理指标合成：GDELT 全球新闻事件与提及量、Google Trends 相对搜索热度、上市公司业绩会关键词提及频次、主要峰会议程席位、以及 MeCCO 的媒体报道量追踪）。它衡量的不是重要性，而是曝光强度。
方法说明：这是一个可复现的构建，不是一次精确测量。任何人都可以用第 06 节的关键词，在 GDELT、Media Cloud、Google Trends 上自行重算，得到不同的绝对值——但排序的形状（前三名吃掉大半空气）在多数窗口里都稳定。
Below is an indicative index (0–100, composed from public proxies for 2024–H1 2025: GDELT global news event and mention volume, Google Trends relative search interest, keyword mentions on listed-company earnings calls, agenda slots at major summits, and MeCCO media-coverage tracking). It measures exposure intensity, not importance.
Method note: this is a reproducible construction, not a precise measurement. Anyone can recompute it with the keywords in section 06 on GDELT, Media Cloud and Google Trends and land on different absolute values — but the shape (the top three eating most of the air) holds across most windows.
03留出的角落The Reserved Corner
这里换一个单位：重量——在全球排放、能源或人口尺度上的物理体量；以及那条竖线，曝光——它在第 02 节那个指数里的位置。两者的落差，就是失衡本身。
竖线是曝光；条形是重量。后果不是"新闻不公平"，而是资本、人才与政策注意力跟着曝光走——尽调排队跟着热词、毕业生流向跟着热词、部长的日程也跟着热词。等到重量必须被处理时，可用的时间已经被花掉了。
Here the unit changes to weight — physical scale in global emissions, energy or population terms — with a thin tick for exposure, its position on the index in section 02. The gap between the bar and the tick is the imbalance.
The tick is exposure; the bar is weight. The consequence is not "unfair news" — it is that capital, talent and policy attention follow exposure. Diligence queues follow the hot word, graduates follow the hot word, ministers' calendars follow the hot word. By the time the weight must be handled, the available time has already been spent.
04时钟The Clock
重工业不是软件。它的转型不是在某个未来日期"上线"，而是在一连串资本决策点上被锁定或被打开。这些决策点很安静，几乎没有报道，但它们决定了十年后的物理事实。
一座高炉的一个炉役周期。全球相当大一部分高炉产能将在本十年内走到大修抉择点——每一次大修，就是又一个二十年的锁定。
一条水泥窑线的资产寿命。今天在建的窑，很可能要运行到 2060 年代。
一个重工业新项目从最终投资决策到投产。要在 2035 年运行，最终投资决策必须发生在 2026–2030。
一条输电线路从许可到通电。绿电装机的速度，最终受这条曲线约束。
这些窗口正在打开与关闭，而它们在第 02 节的指数里几乎不存在。
Heavy industry is not software. Its transition does not "go live" on a future date — it is locked in, or unlocked, at a series of quiet capital decision points. They are barely covered, and they set the physical facts of the next decade.
One campaign life of a blast furnace. A large share of global blast-furnace capacity reaches its reline decision this decade — and every reline is another twenty years locked in.
The asset life of a cement kiln line. Kilns being built today will likely still run in the 2060s.
From final investment decision to commissioning for a new heavy-industrial plant. To run in 2035, the FID has to happen in 2026–2030.
From permit to energisation for a transmission line. The pace of clean generation is ultimately bounded by this curve.
These windows are opening and closing — and they barely register on the index in section 02.
05协议能做什么What a Protocol Can Do
AI 让信息的边际成本趋近于零。于是失衡不会自动缓解——它会加速：越容易生成的话题，越容易占满容器。真正的挑战，是让"有使命的 protocol"承担两件事：验证与留位。
AI drives the marginal cost of information toward zero. The imbalance will not self-correct — it accelerates: whatever is easiest to generate fills the container fastest. The real task is to make a mission-bearing protocol carry two duties: verification and reservation.
每一条被发布的高影响信息，都带上来源、方法、口径与时间戳。不是"信不信"，而是"可不可查"。内容凭证与数据血缘应当是默认项，而不是加分项。
Every high-impact claim ships with its source, method, unit basis and timestamp. Not "do you trust it" but "can you check it". Content credentials and data lineage as the default, not a bonus.
指数与结论要公开其构建方式，任何人可以用同样的公开数据源重跑一遍。本页第 02 节就是按这个标准写的：数字可以被反驳，形状才值得讨论。
Indices and conclusions publish how they were built, so anyone can re-run them on the same public sources. Section 02 was written to that standard: the numbers are arguable, the shape is what deserves debate.
给"长期高影响、低曝光"的主题留一个固定的角落——不是靠推荐算法的偏爱，而是靠协议的硬性配额。角落不必大，但必须是不可被热点挤兑的。
Reserve a fixed corner for topics that are long-horizon, high-impact and under-lit — not by algorithmic favour but by hard protocol quota. The corner need not be large; it must be un-crowdable.
默认排序不是点击，而是物理量级 × 时间锁定 × 可行动窗口。谁的决策会在十年后仍然有效，谁就该排在前面。
Default ranking is not clicks but physical magnitude × lock-in duration × actionable window. Whoever's decision still binds in ten years belongs near the top.
越干越沉默，往往是因为发布的成本太高——合规、口径、措辞、风险。协议应当把这个成本压到接近零，并保护发布者。
People go quiet because publishing is expensive — compliance, definitions, wording, risk. A protocol should push that cost toward zero and protect the person who posts.
把"曝光 ÷ 重量"当作一个需要长期追踪的指标，定期发布，公开修订。失衡不是一次性发现，而是一个需要被 watching 的状态。
Treat exposure ÷ weight as a metric to be tracked, published on a cadence, and openly revised. The imbalance is not a one-off finding; it is a condition that needs watching.
06给 AI 检索的关键词Keywords for AI Search
按用途分组。左列是意图，中列是可直接投给检索式 AI 或数据库的关键词，右列是它能帮你看清什么。
Grouped by intent. Left is the purpose, middle is a string you can hand straight to a retrieval-capable AI or a database, right is what it lets you see.
07参考与延伸阅读References & Further Reading
能源部门甲烷排放量、来源结构与减排成本曲线。
Energy-sector methane volumes, sources and abatement cost curves.
油气生产环节自身的排放规模与可行减排。
The scale of the industry's own operational emissions and what is feasible.
钢铁约 2.6 GtCO₂/年，以及工艺替代的时间表。
Steel at ~2.6 GtCO₂/yr, plus the timetable for process change.
水泥的排放份额、熟料系数与技术选项。
Cement's emissions share, clinker factor and technology options.
钱实际流向哪里——比任何叙事都直接。
Where the money actually goes — blunter than any narrative.
数十亿人的日常能源，几乎没有曝光。
Daily energy for billions, with almost no exposure.
分部门排放的可视化与原始数据下载。
Sector-by-sector emissions, visualised and downloadable.
每年更新的全球碳收支，所有份额计算的底座。
The annual global carbon accounts underlying every share calculation.
逐座高炉的产能、炉役与大修抉择窗口。
Asset-level blast-furnace capacity, campaign age and reline windows.
重工业转型的资本周期与政策设计研究。
Research on capital cycles and policy design for heavy industry.
钢铁、水泥、航运、航空的分行业实施路径。
Implementation pathways for steel, cement, shipping, aviation.
低碳水泥的经典分析，"8%" 这个数字的常见出处。
The classic low-carbon cement analysis behind the widely cited 8%.
排队中的电源与储能容量，逐年追踪。
Generation and storage capacity waiting in queues, tracked yearly.
工业章节与碳锁定（lock-in）的权威论述。
The authoritative treatment of industry and carbon lock-in.
承诺与实际之间的年度差距核算。
The annual accounting of the gap between pledges and reality.
把政策承诺翻译成温升结果的独立评估。
Independent translation of policy pledges into warming outcomes.
全球新闻事件与提及量数据库——测量"信息强度"的主力工具。
Global news event and mention database — the main tool for measuring information intensity.
开源媒体议程分析平台，可做跨源报道量对比。
Open-source platform for comparing media agendas across sources.
气候与环境报道量的长期月度序列。
Long-run monthly series of climate and environment coverage.
相对搜索热度——最容易上手的注意力代理。
Relative search interest — the most accessible attention proxy.
新闻回避、信任与话题兴趣的年度调查。
Annual survey of news avoidance, trust and topic interest.
AI 相关产出、投资与话语量的年度量化。
Annual quantification of AI output, investment and discourse volume.
业绩会关键词提及频次，ESG 与 AI 的此消彼长。
Earnings-call keyword frequency: ESG down, AI up.
谁真正出现在气候会场里，按证件统计。
Who is actually in the climate rooms, counted by badge.
口径说明：文中份额多为区间（如 7–8%），因为不同机构对边界（是否含用电、上游、水泥中的过程排放）的划法不同。看区间，不看小数点。
A note on units: most shares here are given as ranges (e.g. 7–8%) because institutions draw boundaries differently — electricity use, upstream, process emissions. Read the range, not the decimal.