🧠 AI Daily BriefAI每日简报

every market we cover, synthesised by AI — in one place我们覆盖的每个市场，由AI综合 — 集于一处

These are AI (LLM) syntheses of the dashboards — they narrate the regime, flag where signals conflict and test the working thesis, with US macro, the dollar, global liquidity and commodities as shared backdrop. RESEARCH CONTEXT ONLY: nothing here feeds a score, signal or allocation, and a brief can be wrong or overconfident — always verify against the underlying panels. Regenerated automatically each day after the close.以下为对各看板的 AI（大模型）综合解读 — 解读周期、标注信号冲突、检验当前假设，并以美国宏观、美元、全球流动性与商品为共同背景。仅供研究参考：此处内容不参与任何评分、信号或仓位，且简报可能出错或过度自信 — 请始终对照底层面板核实。每个交易日收盘后自动更新。

Right now当前状态

as of数据截至 2026-07-15 ?Snapshot of the deterministic dashboards at build time — regime, any signals pointing in different directions, and what liquidity is doing. Research context only.构建时的确定性看板快照 — 周期状态、有无方向分歧信号，以及流动性动向。仅供研究参考。

Regime市场格局 Mixed 混合

Signals信号 One tension worth knowing: bonds and stocks are telling different stories. 值得关注的一处矛盾：债券与股票传递出不同信号。

Money流动性 Liquidity is roughly neutral 流动性大体中性

AI reviewAI复盘 Rate-limited overnight — ran on dashboards alone 隔夜遭速率限制，仅依靠看板运行

What's ahead近期日程

?Scheduled macro events and data releases in the next two weeks. Dates are deterministic — no forecasts on the visible tier. Hover any row for model leans and market-implied reads; all annotated with no accuracy record yet while the scoreboard is empty.未来两周的宏观事件与数据发布。日期为确定性信息，可见层不含预测。悬停行可查看模型倾向与市场隐含读数，评分板为空时注明尚无记录。

2026-07-29 FOMC rate decision 美联储议息会议
2026-07-30 GDP (BEA estimate) GDP 数据
2026-07-30 PCE / personal income (BEA) PCE 物价指数
2026-07-30 PCE PCE
2026-08-07 NFP NFP
2026-08-12 CPI CPI

🇺🇸 Macro · cross-asset synthesis宏观 · 跨资产综合 ?AI synthesis of the macro picture as one cross-asset read — US growth/inflation regime, the dollar, global liquidity, rates, commodities and crypto. RESEARCH CONTEXT only: it narrates the regime, flags conflicting signals and tests the liquidity-rotation thesis; it never feeds a score, signal or allocation. Can be wrong — verify against the panels.AI 将宏观图景作为单一跨资产解读 — 美国增长／通胀周期、美元、全球流动性、利率、商品与加密货币。仅供研究参考：它解读周期、标注信号冲突并检验流动性轮动假设，绝不参与任何评分、信号或仓位。可能出错，请对照各面板核实。

Loading the latest AI brief… it regenerates daily after the close.正在加载最新AI简报…每个交易日收盘后更新。

🇨🇳 China & Hong Kong synthesis中港综合 ?An AI-generated synthesis of the China & Hong-Kong dashboards, with US macro, the dollar, global liquidity and commodities as backdrop. RESEARCH CONTEXT only — it narrates the regime, flags where signals conflict and tests the policy/liquidity thesis (A-shares mean-revert, so leaders read as extended); it never feeds a score, signal or allocation. Can be wrong or overconfident — verify against the underlying panels.AI 对中港看板的综合分析，以美国宏观、美元、全球流动性与商品为背景。仅供研究参考——它解读周期、标注信号冲突并检验政策／流动性假设（A股均值回归，故强势股视为已拉伸），绝不参与任何评分、信号或仓位。可能出错或过度自信，请对照各底层面板核实。

Loading the latest AI brief… it regenerates daily after the close.正在加载最新AI简报…每个交易日收盘后更新。

₿ Bitcoin synthesis比特币综合 ?An AI-generated synthesis of the Bitcoin Vector — cycle, valuation, leverage, on-chain activity and ETF flows, with US macro and global liquidity as backdrop (BTC as the long-duration tail of global liquidity). RESEARCH CONTEXT only — it never feeds a score, signal or allocation, and can be wrong or overconfident. Verify against the Bitcoin Vector panels.AI 对比特币向量的综合分析 — 周期、估值、杠杆、链上活动与 ETF 资金流，以美国宏观与全球流动性为背景（比特币为全球流动性的长久期尾部）。仅供研究参考 — 绝不参与任何评分、信号或仓位，且可能出错或过度自信。请对照比特币向量面板核实。

Loading the latest AI brief… it regenerates daily after the close.正在加载最新AI简报…每个交易日收盘后更新。

The brief's own record简报自身记录

?The AI brief logs a lean for each major call — subject, direction, horizon, and a falsifier. These are scored against realized prices. Below 20 graded calls the scoreboard shows the artifact's own calibration note, not a hit-rate.简报为每次主要判断记录倾向 — 标的、方向、时间窗口与证伪条件，并依据实际价格评分。评分次数不足20次时不显示命中率，仅显示校准说明。

Too few graded calls yet to judge — this is a public track record, not a proven edge. 已评分的判断还太少，暂无法评价——这是公开的判断记录，不是已证明的优势。 ?2 leans scored, hit-rate 1.0 (directional accuracy 1.0). Sample is tiny — the brain's leans are accountability, not yet a validated edge.已评分 2 条，命中率 1.0（方向准确率 1.0）。样本极小 —— 大脑的判断是问责，尚非验证过的优势。

Open leans待验证判断

Leaning away from China equities over the next 20 days. 倾向回避 China equities，未来 20 天。 ?Falsifier: HK peg returns to strong-side and China growth score flips positive. · Check by: 2026-07-24证伪条件：HK peg returns to strong-side and China growth score flips positive. · 截止：2026-07-24

Leaning away from US equities over the next 30 days. 倾向回避 US equities，未来 30 天。 ?Falsifier: HY OAS tightens below 2.7 and EBP turns decisively positive (credit stress abates). · Check by: 2026-08-10证伪条件：HY OAS tightens below 2.7 and EBP turns decisively positive (credit stress abates). · 截止：2026-08-10

Leaning away from China equities over the next 20 days. 倾向回避 China equities，未来 20 天。 ?Falsifier: China quad shifts to Q1 (Goldilocks) or Q2 (Reflation), and HK peg strengthens with HIBOR-OIS spread narrowing. · Check by: 2026-07-29证伪条件：China quad shifts to Q1 (Goldilocks) or Q2 (Reflation), and HK peg strengthens with HIBOR-OIS spread narrowing. · 截止：2026-07-29

Leaning toward Long Treasuries over the next 20 days. 倾向看好 Long Treasuries，未来 20 天。 ?Falsifier: 10y real yield rises above 2.5% and the curve bear-steepens (NTFS turns negative), breaking the real-rate-shock driver and signaling that restrictive policy is reinstating tightening pressure. · Check by: 2026-07-29证伪条件：10y real yield rises above 2.5% and the curve bear-steepens (NTFS turns negative), breaking the real-rate-shock driver and signaling that restrictive policy is reinstating tightening pressure. · 截止：2026-07-29

Leaning toward China equities over the next 20 days. 倾向看好 China equities，未来 20 天。 ?Falsifier: China A-shares reverses and the China fingerprint (HK, copper) stops confirming. · Check by: 2026-08-06证伪条件：China A-shares reverses and the China fingerprint (HK, copper) stops confirming. · 截止：2026-08-06

Leaning away from Growth over value over the next 30 days. 倾向回避 Growth over value，未来 30 天。 ?Falsifier: Growth versus value spread closes to neutral (within 0.5σ) and semis RS recovers above –1σ, or the 10y yield reverses at least 20bp lower. · Check by: 2026-08-27证伪条件：Growth versus value spread closes to neutral (within 0.5σ) and semis RS recovers above –1σ, or the 10y yield reverses at least 20bp lower. · 截止：2026-08-27

Overnight deliberation隔夜AI复盘

?The overnight AI reviewer reads the dashboards with a set of read-only tools. This is AI-generated deliberation — labelled as such. When rate-limited (the usual state), it shows an honest absent note rather than synthetic content.隔夜 AI 审阅者使用一组只读工具读取看板内容。这是 AI 生成的审议结果，已明确标注。遭遇速率限制时（常见状态），显示诚实的缺席提示，不生成合成内容。