<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Shuai Yao</title><description>The best way to predict the future is to invent it</description><link>https://shuaiyao.dev/</link><language>en</language><item><title>Warsh&apos;s Hawkish Pause: The October FOMC, AI Capex and a Liquidity Tug-of-War</title><link>https://shuaiyao.dev/investing/warsh-hawkish-pause-october-fomc/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/warsh-hawkish-pause-october-fomc/</guid><description>A 10–2 hold, but &apos;tighter, not tight&apos; pulled the Fed put. A line-by-line read of the 17-page transcript, and why AI capex flipped from growth badge to valuation risk.</description><pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate><category>NDX</category><category>GOOGL</category><category>AMZN</category><category>PLTR</category><category>AMD</category><category>AAPL</category><category>BWXT</category><category>CRCL</category><category>Macro</category><category>Fed</category><category>Warsh</category><category>FOMC</category><category>AI Capex</category><category>US Equities</category></item><item><title>Midterm years are weak. The year after is the trade.</title><link>https://shuaiyao.dev/investing/midterm-2026-the-year-after/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/midterm-2026-the-year-after/</guid><description>Since 1950 the midterm year is the weakest of the four and the year after is the strongest. In 2026 the swing factor is the 10-year yield, not the polls.</description><pubDate>Mon, 05 Oct 2026 00:00:00 GMT</pubDate><category>SPX</category><category>SPY</category><category>Macro</category><category>US Equities</category><category>Elections</category><category>Rates</category><category>Seasonality</category></item><item><title>Hyperliquid: The Exchange That Buys Itself</title><link>https://shuaiyao.dev/investing/hyperliquid-the-exchange-that-buys-itself/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/hyperliquid-the-exchange-that-buys-itself/</guid><description>CFTC onshoring, T-bill yield routed into HYPE buybacks, S&amp;P 500 perps on HIP-3. Bullish the project on a 2–3 year view, patient on the coin at ~$80.</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>HYPE</category><category>USDC</category><category>COIN</category><category>BNB</category><category>Crypto</category><category>Hyperliquid</category><category>Stablecoins</category><category>Tokenomics</category><category>US Treasuries</category><category>Regulation</category></item><item><title>Competition is for losers — full notes on Thiel&apos;s Stanford talk</title><link>https://shuaiyao.dev/blog/competition-is-for-losers/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/competition-is-for-losers/</guid><description>Notes on Peter Thiel&apos;s CS183B lecture: value created and value captured are separate variables, and competition grinds the second one to zero.</description><pubDate>Wed, 05 Aug 2026 00:00:00 GMT</pubDate><category>Startups</category><category>Strategy</category><category>Notes</category></item><item><title>Does the American heat move the 2026 World Cup?</title><link>https://shuaiyao.dev/blog/does-heat-move-the-world-cup/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/does-heat-move-the-world-cup/</guid><description>A weather and performance study on the first 81 matches: humid heat drains players and reshapes tactics, but the cost rarely reaches the scoreboard.</description><pubDate>Thu, 02 Jul 2026 00:00:00 GMT</pubDate><category>Forecasting</category><category>Sports</category><category>Data Analysis</category></item><item><title>Coherent (COHR): an AI-optics winner with a diamond-cooling option — priced, not cheap</title><link>https://shuaiyao.dev/investing/coherent-ai-optics-diamond-cooling/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/coherent-ai-optics-diamond-cooling/</guid><description>COHR is being re-rated as an AI optics winner with an NVIDIA anchor; diamond cooling is a 2027+ option. I&apos;d build at 335–355, not chase 380+.</description><pubDate>Sat, 27 Jun 2026 00:00:00 GMT</pubDate><category>COHR</category><category>LITE</category><category>Equity</category><category>AI</category><category>Optics</category><category>Valuation</category></item><item><title>Train the body to train the mind</title><link>https://shuaiyao.dev/blog/train-the-body-to-train-the-mind/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/train-the-body-to-train-the-mind/</guid><description>Exercise trains judgment and attention, not just the body. The evidence for aerobic exercise is solid; the case for team sports is thinner but points the same way.</description><pubDate>Fri, 26 Jun 2026 00:00:00 GMT</pubDate><category>Reasoning</category><category>Learning</category><category>Exercise</category><category>Habits</category></item><item><title>USMCA 2026: The Real Risk Is the Loonie, Not a Breakup</title><link>https://shuaiyao.dev/investing/usmca-2026-review-and-the-canadian-dollar/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/usmca-2026-review-and-the-canadian-dollar/</guid><description>Mexico&apos;s pipe duties hint no one softens before the 2026 USMCA review. Expect a messy annual-review limbo, not a cliff; USD/CAD biased higher.</description><pubDate>Fri, 26 Jun 2026 00:00:00 GMT</pubDate><category>CAD</category><category>USDCAD</category><category>WTI</category><category>Macro</category><category>Trade</category><category>USMCA</category><category>Canada</category><category>CAD</category><category>Steel &amp; Aluminum</category></item><item><title>Warsh&apos;s First FOMC: Not a Hold, a Reaction-Function Reset</title><link>https://shuaiyao.dev/investing/warsh-first-fomc-reaction-function/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/warsh-first-fomc-reaction-function/</guid><description>Warsh&apos;s first FOMC held at 3.50–3.75%; the signal is no dot, no forward guidance and five framework reviews. Not an AI bust, a fragile bull.</description><pubDate>Thu, 18 Jun 2026 00:00:00 GMT</pubDate><category>SPX</category><category>NDX</category><category>USDJPY</category><category>NVDA</category><category>Macro</category><category>Fed</category><category>Warsh</category><category>Yen Carry</category><category>AI Capex</category><category>US Equities</category></item><item><title>The BOJ Hike, US Liquidity, and Japan&apos;s Rally</title><link>https://shuaiyao.dev/investing/boj-hike-liquidity-and-us-stocks/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/boj-hike-liquidity-and-us-stocks/</guid><description>The BOJ hikes to 1.00% and the Nikkei rallies anyway. The market is buying &apos;not more hawkish&apos;, not the hike. Risk-on near term, more fragile the higher it goes.</description><pubDate>Tue, 16 Jun 2026 00:00:00 GMT</pubDate><category>NKY</category><category>USDJPY</category><category>JGB</category><category>NDX</category><category>Macro</category><category>BOJ</category><category>Yen Carry</category><category>Nikkei</category><category>US Equities</category><category>Liquidity</category></item><item><title>SpaceX goes public: tradeable, not yet investable</title><link>https://shuaiyao.dev/investing/spacex-ipo-tradeable-not-investable/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/spacex-ipo-tradeable-not-investable/</guid><description>The IPO prices a $1.77T space-AI hyperscaler, not Starlink cash flow. An event trade at $135, not a chase above $160.</description><pubDate>Wed, 10 Jun 2026 00:00:00 GMT</pubDate><category>SPCX</category><category>IPO</category><category>SpaceX</category><category>Macro</category><category>Valuation</category></item><item><title>How I think about markets</title><link>https://shuaiyao.dev/investing/how-i-think-about-markets/</link><guid isPermaLink="true">https://shuaiyao.dev/investing/how-i-think-about-markets/</guid><description>The frame for everything I&apos;ll post here: a continuous score, a measured edge, and a position sized to how wrong I might be. Opinions, not advice.</description><pubDate>Mon, 01 Jun 2026 00:00:00 GMT</pubDate><category>Framework</category><category>Process</category></item><item><title>Citing the canon is the load-bearing UX choice</title><link>https://shuaiyao.dev/blog/citing-the-canon-as-load-bearing-ux/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/citing-the-canon-as-load-bearing-ux/</guid><description>Why every score in FengShui Blueprint points to a classical source, and why that decision did more work than the rules engine.</description><pubDate>Sun, 17 May 2026 00:00:00 GMT</pubDate><category>product</category><category>ux</category><category>rules-engines</category><category>fengshui</category><category>engineering-discipline</category></item><item><title>CME&apos;s compute futures: not a perpetual — yet</title><link>https://shuaiyao.dev/blog/cme-compute-futures-not-a-perpetual/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/cme-compute-futures-not-a-perpetual/</guid><description>CME&apos;s first compute futures look like electricity or weather derivatives, not Bitcoin perps. The more interesting product is the compute perpetual coming next.</description><pubDate>Sun, 17 May 2026 00:00:00 GMT</pubDate><category>derivatives</category><category>futures</category><category>compute</category><category>ai-infra</category><category>perpetuals</category><category>cme</category><category>hyperliquid</category></item><item><title>Marketing as three measurement problems: SEO, GEO, CRM</title><link>https://shuaiyao.dev/blog/marketing-three-measurement-problems/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/marketing-three-measurement-problems/</guid><description>SEO is hygiene, GEO is the next three years of brand discovery, and CRM is the data plane underneath. One identity pipeline, not three vendor stacks.</description><pubDate>Sun, 17 May 2026 00:00:00 GMT</pubDate><category>marketing</category><category>seo</category><category>geo</category><category>crm</category><category>measurement</category><category>engineering-discipline</category></item><item><title>Non-custodial escrow as a marketplace primitive: design notes from TaskMarket</title><link>https://shuaiyao.dev/blog/non-custodial-escrow-marketplace-primitive/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/non-custodial-escrow-marketplace-primitive/</guid><description>A 4-state Solidity contract on Base: how TaskMarket settles in USDC without ever touching user funds, and what each invariant prevents.</description><pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate><category>solidity</category><category>web3</category><category>marketplace</category><category>escrow</category><category>base</category><category>engineering-discipline</category></item><item><title>83 ties was enough to kill TSFMs as my baseline</title><link>https://shuaiyao.dev/blog/tsfm-knockout-tournaments-83-ties/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/tsfm-knockout-tournaments-83-ties/</guid><description>A 3-model TSFM ensemble lost to a tuned Elo + xG + injury model on 83 knockout ties. Why the data regime mattered more than the model.</description><pubDate>Fri, 08 May 2026 00:00:00 GMT</pubDate><category>forecasting</category><category>time-series</category><category>tsfm</category><category>sports</category><category>backtest</category><category>negative-result</category></item><item><title>Migrating a community platform without migrating the database</title><link>https://shuaiyao.dev/blog/migrating-stack-without-migrating-schema/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/migrating-stack-without-migrating-schema/</guid><description>lakebbs moved from Vue 2 + Express to Next.js 16 + Drizzle without touching the MySQL schema. Why locking the database during a rewrite is the right call.</description><pubDate>Mon, 04 May 2026 00:00:00 GMT</pubDate><category>nextjs</category><category>drizzle</category><category>mysql</category><category>migration</category><category>engineering-discipline</category><category>rewrite</category></item><item><title>Workers AI as a triage filter, not a generator</title><link>https://shuaiyao.dev/blog/workers-ai-as-triage-not-generator/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/workers-ai-as-triage-not-generator/</guid><description>Spend Workers AI calls at the front of the pipeline as a scoring filter, not at the end as a generator. The cost, latency and quality case.</description><pubDate>Tue, 28 Apr 2026 00:00:00 GMT</pubDate><category>cloudflare</category><category>workers-ai</category><category>ai-engineering</category><category>cost-engineering</category><category>triage</category><category>agents</category></item><item><title>Forecasting the new Champions League</title><link>https://shuaiyao.dev/blog/forecasting-the-new-champions-league/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/forecasting-the-new-champions-league/</guid><description>Notes on UEFA Oracle — a hybrid TSFM + Elo + Poisson model for the 2025-26 UCL, benchmarked against Polymarket.</description><pubDate>Tue, 14 Apr 2026 00:00:00 GMT</pubDate><category>Forecasting</category><category>Sports</category><category>Prediction Markets</category></item><item><title>Time Series Foundation Models on equities: the hostile case</title><link>https://shuaiyao.dev/blog/tsfm-on-equities-hostile-case/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/tsfm-on-equities-hostile-case/</guid><description>Why US equities are the stress test TSFMs don&apos;t want, and what an honest zero-shot benchmark has to do differently.</description><pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate><category>Time Series</category><category>Forecasting</category><category>Finance</category></item><item><title>Fermi estimation as a scoping discipline</title><link>https://shuaiyao.dev/blog/fermi-for-data-scoping/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/fermi-for-data-scoping/</guid><description>Before you open a notebook, get the answer within an order of magnitude. If the physics ballpark is wrong, the pipeline can&apos;t save you.</description><pubDate>Mon, 30 Mar 2026 00:00:00 GMT</pubDate><category>Estimation</category><category>Physics</category><category>Practice</category></item><item><title>What the Rust compiler refused to let me avoid</title><link>https://shuaiyao.dev/blog/rust-refuses-to-let-you-avoid/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/rust-refuses-to-let-you-avoid/</guid><description>Rewriting a trading bot from Python to Rust isn&apos;t a syntax change. It&apos;s a forced audit of every decision the dynamic version let you postpone.</description><pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate><category>Rust</category><category>Systems</category><category>Trading</category></item><item><title>Shop-floor latency is a product decision</title><link>https://shuaiyao.dev/blog/shop-floor-latency/</link><guid isPermaLink="true">https://shuaiyao.dev/blog/shop-floor-latency/</guid><description>Two-second lag kills trust in a production console. It&apos;s not a performance knob — it&apos;s the shape of the product.</description><pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate><category>Product</category><category>Realtime</category><category>Ops</category></item></channel></rss>