🟥 Things I'm looking at today 🟥 Powered by @yeet Subscribe: Crypto lower, Clarity vote today, AI stocks slump Crypto * BTC: 77,090 (-1%) | BTC.D: 58.8% (0%) * ETH: 2,477 (-2%) | BNB: 717 (-1%) | SOL: 101 (-1%) * Fear & Greed: 68 | 24h Liq: $358m * BTC ETFs: +$160M | ETH ETFs: +$121M * Top Gainers: AKE, CAP, LOBSTER, AI, LAPTOP * Crypto lower, ahead of Clarity vote today * Dems reject Clarity draft, send counter * Clarity 2026 odds fall to 16% on Polymarket * 114-page crypto tax package skips mining, staking * 18 state AGs urge Senate to reject Clarity Act * Largest US public union AFSCME opposes Clarity * Gillibrand privately pushes Dems to back Clarity * Bank groups demand strict stablecoin yield limits * Crypto has 'not priced in' Clarity: Bernstein * BTC post-quantum roadmap forms around BIP-360 * Strategy buys $139M STRC, Strive adds $37M BTC * ETH ETFs used as CME futures collateral: Bitfinex * Ethereum L1, Base split on AA standards * Robinhood plans voting, redemptions for stock tokens * Pixelmon to discontinue all game development * Revolut gave hackers data on 680 customers * US SBR bill orders study, not BTC buys * CoinEx to shut down * Etherscan launches Robinhood Chain explorer * WLFI proposes rewards for token lockers * BabyDoge acquires LimeWire for creator push * Kraken adds DeFi yield to tokenised stocks, ETFs * Balancer wind-down proposed as revenue stalls * MetaMask adds new scam protections to wallet * Kalshi launches gold, silver perps in US * DOJ seeks $61M Iran oil cash laundered via Binance * UK BTC DAT Satsuma to delist from LSE * FCA, BoE to publish UK tokenisation roadmap * Poland lost $230M on Venezuela oil deal in USDT * UAE to put Avalanche in UAE PASS digital ID * Stock, crypto interest jumps 96% in South Korea * Argentina to share crypto user data under CARF * Ondo expands to Australian markets Macro * NASDAQ: $26,190 (-0.6%) | Gold: $4,290 (-0.3%) * AI sell-off drives US stocks lower * Odds of Fed hike hit 94% * US 10yr yield hits highest since 2007 * US weighs 'phased' deal with Iran * Oil tops $108 as ME routes threatened * US gas prices up 52% YTD * IEA cuts 2026 oil demand forecast by 940K bpd * Fed to buy $15.6B bills for MBS reinvestment * SC rejects Trump bid to restrict mail ballots * 51% odds of Dems winning Senate, per Kalshi * EPA scraps limits on power plant emissions * DOJ uncovers $245M Covid relief programs fraud * US admits for first time it has weapons in space * Supertanker struck by mines in Hormuz: Iran * US could stay in Iran and keep oil: Trump * Houthis launch more attacks on Saudi * UNSC calls emergency meeting on Bab al-Mandab * Ukraine drones hit Russia's Syzran refinery * NATO jets shoot down drone over Lithuania * Russian frigate fires flares at Danish helicopter * Kim Jong Un vows stronger alliance with Russia * US sanctions Russia's VTB for helping Iran * Newsom won't run in 2028 if Harris runs * Kennedy Center could go broke, shut this week * Paramount threatens California exit * House votes to permanently scrap the penny AI & Tech * AI stocks fall on slowdown calls * CrowdStrike, Palo Alto rally as AI fears lift cyber * Trump rings Jensen on stage, brands AI fears 'hoax' * 'Not going to let' an AI slowdown happen: Jensen * AI doomer predictions are 'made up’: Jensen * China rejects AI slowdown calls * Clinton urges US to 'enlist the Chinese' on AI * Schumer demands classified Senate briefing on AI * Microsoft agrees AI is getting dangerous * Microsoft unveils 'Humanist AI' code of conduct * AI boom needs $110B of new power plants: Moodys * OpenAI contractors read users' ChatGPT chats * OpenAI buys Glass Imaging for $300M * OpenAI enterprise revenue +32% m/m in July * OpenAI cut Luna price 80%, usage rose 10x * OpenAI agents' top priority is automating research * Trump met Altman privately at GOP convention * DeepMind researcher quits, says AI 'could kill us all' * Anthropic launches Claude for Financial Advisors * Anthropic won't delay 2026 IPO: Axios * We’re 'highly profitable' excl. big costs: Anthropic * Rumble surges on reported $13.7B Anthropic deal * Anthropic co-founder calls for AI 'kill switch' * Palantir, Nvidia restrict Anthropic, OpenAI models * Google Cloud claims 2.7x price-performance edge * Oracle announces new layoffs to offset AI capex * Apple accepts 30-40% Samsung memory price hikes * Apple releases iOS 27 with Siri AI overhaul * iOS 27 lets users hide location from chosen people * AI brain implant lets paralysed woman speak again * X, SpaceXAI drop antitrust lawsuit against Apple * Amazon, Netflix, YouTube form streaming alliance @YEET: Crypto's Casino Just passed $3.4B in lifetime volume! gm

September returns for S&P 500 over 25 years: 2025: +3.50% 2024: +2.00% 2023: −4.90% 2022: −9.30% 2021: −4.80% 2020: −3.90% 2019: +1.70% 2018: +0.40% 2017: +1.90% 2016: −0.10% 2015: −2.60% 2014: −1.60% 2013: +2.97% 2012: +2.40% 2011: −7.20% 2010: +8.80% 2009: +3.60% 2008: −9.10% 2007: +3.60% 2006: +2.50% 2005: +0.70% 2004: +0.90% 2003: −1.20% 2002: −11.00% 2001: −8.20% 25-year average: −1.2%

A pipeline with 100% delivery rate, 100% schema validity, and zero retries or errors... yet re-running the exact same request flips the verdict. This paper reports that shocking negative result, fully preregistered with a complete audit trail. Title: Clean Engineering, Unstable Measurement: A Preregistered Reliability Failure of Black-Box LLM Observers on Shared Endpoints URL: 🔧 Highlight 1: Clean engineering does not mean reliable measurement Across 3,312 planned calls, response rate and schema validity both hit 100% with zero retries or errors, yet the repeat-ranking agreement (median Spearman) was only 0.400 — clearly failing the preregistered stability gate (threshold 0.90). 🎲 Highlight 2: Byte-identical inputs still drift Replaying the exact same request 24 hours later gave only a 0.780 exact-ranking agreement rate. Interestingly, comparisons across windows on the same day scored a nearly identical 0.805 median, revealing this isn't "next-day drift" but immediate platform-level nondeterminism. 📉 Highlight 3: More samples don't fix it Scaling observer calls from 8 to 500 (748,000 calls total) still left the stability gate pass rate at 0/500. The root cause: score gaps between candidates were 7-10 orders of magnitude smaller than the noise floor, making most tasks fundamentally unrankable. I think this is a great reminder that before using an LLM as a measurement instrument, you need to measure whether the instrument itself is stable. #LLMEval# #Reproducibility#

Here is every stock currently in the KraneShares Artificial Intelligence & Technology ETF $AGIX: NVIDIA $NVDA 4.34% weighting Meta Platforms $META 4.00% Google $GOOGL 3.79% Microsoft $MSFT 3.70% Apple $AAPL 3.51% Amazon $AMZN 3.21% Snowflake $SNOW 2.84% Taiwan Semiconductor 2.81% Broadcom $AVGO 2.56% Advanced Micro Devices $AMD 2.47% Palantir $PLTR 2.20% Datadog $DDOG 2.18% Tesla $TSLA 2.15% Astera Labs $ALAB 2.10% SpaceX 1.85% SK Hynix 1.83% Anthropic 1.79% Arm $ARM 1.77% ServiceNow $NOW 1.70% Cloudflare $NET 1.70% Palo Alto Networks $PANW 1.65% Micron Technology $MU 1.61% ASML $ASML 1.60% Nebius Group $NBIS 1.54% Nutanix $NTNX 1.50% Shopify $SHOP 1.47% Samsung Electronics 1.46% Eaton $ETN 1.41% MediaTek 1.41% Arista Networks $ANET 1.36% NextEra Energy $NEE 1.36% Hut 8 $HUT 1.35% Roblox $RBLX 1.30% Oracle $ORCL 1.28% Texas Instruments $TXN 1.28% Flex $FLEX 1.25% Riot Platforms $RIOT 1.20% Tower Semiconductor $TSEM 1.20% Salesforce $CRM 1.18% Netflix $NFLX 1.14% Samsara $IOT 1.09% Apptronik 1.06% GlobalFoundries $GFS 1.06% Applied Digital $APLD 1.04% TeraWulf $WULF 1.03% Sanmina $SANM 1.03% Tempus AI $TEM 1.01% Cadence Design Systems $CDNS 0.99% Twilio $TWLO 0.97% SAP $SAP 0.96% SpaceX (pre-IPO shares) 0.96% Zscaler ZS 0.95% Teradyne TER 0.95% Vertiv VRT 0.91% CoreWeave CRWV 0.90% Renesas Electronics 0.79% Circle CRCL 0.74% TE Connectivity TEL 0.73% Ciena CIEN 0.73% Ayar Labs 0.68% Nuro 0.55% Nitto Boseki 0.51% Polymarket 0.28% Disclosure: KraneShares is a WOLF Financial Partner.

On 8090's go to market: We could have sold into easier, forgiving customers and posted a nice logo wall. Instead, we did the opposite. We went straight at the hardest, most regulated buyers we could find: healthcare, insurance, life sciences, aerospace, energy, manufacturing, financial services, and the United States government. The logic was simple. If AI can be trusted to do mission-critical work inside the industries with the least tolerance for error and the most oversight, it can be trusted with serious work anywhere. Some recent 8090 wins: - A publicly traded insurer ran a workflow its existing systems could never accommodate and took years of cost out of their system. - A life sciences customer cut its authoring-to-approval cycle by more than half, so the team spends its time on the science instead of the paperwork. - A manufacturer brought tens of thousands of parts under automated handling, and the people who used to do manual intake now work the exceptions that actually need a human. Overwhelmingly all of our customers create negative churn for our business: a small customer quickly becomes medium, a medium customer becomes large. Our large customers are becoming extra-large. And every engagement makes the platform smarter, because of how have engineered our Software Factory. A network effect in Enterprise Software is quite novel and we may be onto one.

1 week out from our @EYnews x 8090 Modernization webinar. It will cover: ---> Discovery and Portfolio Analysis ---> Large-scale delivery ---> Practical methods for reverse engineering legacy ---> Where AI agents fit ---> How people approve/audit every step One hour with live Q&A. Reserve your spot:

To CEOs, CFOs, CIOs and Directors: Have you read the White House Accord on Super Intelligence? What will you do to safely deploy Super Intelligence within your Enterprise? EY and 8090 are building a new solution - KYSI (Know Your Super Intelligence) designed to create a verifiable record of AI-enabled decisions by capturing the context and evidence, and spotlighting controls, across enterprise environments. The solution helps organizations strengthen governance while keeping pace with innovation. It does this by providing: - End-to-end traceability - Policy-to-risk mapping - Auditable evidence packages Together, we are helping organizations move beyond AI experimentation toward AI they can trust, govern and scale enterprise-wide with confidence. To learn more, reach out to your EY Partner or ping us at

This is a really good development. As model costs converge, the next likely step will be the use of independent 3rd party harnesses/control planes (8090 Software Factory, Palantir, Nous’ Hermes etc). A 3rd party harness + the ability to use any model quasi-interchangeably is something all enterprises will ask for within 36 months as they realize the importance of data/IP leakage and keeping complete sovereignty of their encoded judgement.

For those tracking the America closed source vs China open weight model debate: In an initial pilot on modernizing an application from PHP to Next.js, Opus 4.8 with 8090’s Software Factory was simultaneously 1.4× cheaper and 1.5× faster than Opus 4.8 alone. Pairing our Software Factory with the cheaper GLM 5.2 model cut costs 16.4×, though it ran 3× slower than Opus 4.8 alone. Results are directional (n=1 per arm), and next steps are to rerun with proper controls and 10–15 buyer-relevant legacy modernization tasks developed with our Sales and GTM teams. More to come but the important question it raises is why any American public company with shareholders burning money on closed source models when the open weight ones are so much cheaper? We will rerun this on American open source models next (aka Nvidia)…