Nov 2 8-K: 811 BTC at $80,846 avg ($65.5M). Total: 849,327 BTC, avg cost $75,451. $208.2M USD Cash bought back 2.09M STRC shares. 3x the BTC spend. Two accounts: USD Reserve ($4.79B, dividends/debt) and USD Cash ($132.7M left). Buyback drained the second Signal or housekeeping https://t.co/Dxzv191pso

Base now pays builders, creators and trenchers separately. Good split, different crowds. I built Base Invaders, a Mini App shooter with an onchain leaderboard. @jessepollak is the boss in it. Literally. Does a finished build still count for Builder Grant? )) https://t.co/BCi3ToDkER

Monday, October 26. The OCC gave Ripple a final federal charter for a bank that can’t take deposits, can’t lend, and has no FDIC insurance. So why should anyone holding RLUSD care? Because the bank takes the reserves. The liquid assets behind RLUSD, around $2.5B per one outlet (approximate, not confirmed). Until now that sat under one regulator, NYDFS. Now there’s a second, federal one, looking at exactly what institutions check first: do the dollars exist and are they kept separate. Issuance doesn’t move. Standard Custody & Trust still issues RLUSD under New York. Ripple National Trust Bank holds the reserves and does custody for institutional clients. Two lanes, two supervisors. Application July 2025, conditional approval December 12, 2025, final in October 2026. Almost 16 months start to finish. If you build on stables: RLUSD was already live on Base and Ink per December 2025 reports, check the current state before you integrate. The approved activities are RLUSD reserves, collateral trustee and fiduciary custody. XRP isn’t on that list. Anyone selling this as an XRP catalyst is reading the wrong line. The real test is still ahead. Standard Custody applied for a Fed master account on June 30, 2025, and as far as I can see it’s still pending. Custodia fought for years for its own and got denied. Frens, does the Fed give a master account to a stablecoin issuer in 2027?

On Oct 14 Anthropic reportedly told IPO investors that Q3 revenue hit $16.8B. All of H1 was $16.3B. One quarter beat six months. The same draft S-1 shows an $80.7B net loss. How does a company print that and still target a $2T listing? Easy part first. About $72B of that loss is a non-cash accounting charge on the convertible notes they sold to raise money. What is left is roughly $8.7B, and even that is not cash burn. Real compute spend was nearly $9B in H1, more than all of 2025 ($7.3B). The number that should keep you up is another one. $600B+ in compute commitments, with talks that could push it toward $750B. About a fifth of H1 revenue came from just two customers. Amodei himself said back in February that buying compute far ahead of revenue can sink a lab if growth misses even a little. Small fixes if you are quoting this. Target valuation is around $2T, not 1.5-1.75T. The 1.75T is SpaceX at its June IPO. Hyperliquid’s pre-IPO perp had Anthropic near $2.15T in mid-Sept (approximate data, not confirmation). Ticker is not announced yet, so ignore the fake ones. Now the part for us, frens. The week before SpaceX listed, BTC fell 15%, the worst week since FTX in Nov 2022. Analysts pointed at rotation because SpaceX set aside up to $22.5B for retail. But the same weeks had a hawkish Fed and an equity selloff, so nobody can prove the IPO did it. Anyone posting a clean 25-30% BTC drop from the Anthropic IPO is guessing. It may well happen, but there are no receipts yet. I am tracking the IPO wave (SpaceX, Anthropic, OpenAI) against BTC and ETF flows and will post the data as it lands. If Anthropic prices at $2T and pops, does the wealth effect flow back into crypto or does the money stay in equities? Curious what you are positioned for.

Etherscan has been in Claude's directory since September. 16 tools, 60+ EVM chains, no API key, read-only. Feature list is fine. Deployer, verified source, ABI, logs, gas, all from one prompt. If you check fresh contracts all day, that's a lot of explorer tabs gone. Now the docs. There's a heading "Guard against prompt injection" and one sentence under it: be careful when you combine Etherscan with other connectors in the same chat. The tool only reads. It never signs, never moves funds. So what is it guarding against? My read, not their claim: text onchain is writable by anyone. Token name, symbol, comments inside verified source. A scam deployer controls that text, and your model reads it as part of the answer. Now put Drive or Gmail in the same chat and think about who else is reading it. frens, I'd keep Etherscan in a clean chat with nothing attached and check fresh deploys there. Do you run your explorer tools separate from mail and docs, or does everything live in one chat? Next post I run it on a junk contract and show what comes back.

2 posts. Rank 16 out of 80 curators. And 1,052 views that paid me exactly zero. Ethos asked how you reward unique content on X without paying for botted views. After two posts in Curators v2 my answer is simple: stop paying for views. Entry is 1300+ on Ethos, and only reactions from 1400+ accounts count. A real link click is 0.5 points, a reply of 2+ words is 5, a quote is 15. A curator who fakes it with bots risks negative reviews or a slash on his real onchain rep. That is a very different price tag than "get banned and open a new account". Now my own numbers, since I just poked it with my hands. My breakdown of the Jesse Pollak trust market (Yes sat at 71%) got 1,052 views. On old X habits I'd already be opening champagne. Ethos politely reminded me that views pay zero here. What paid were 34 clean clicks to the market, 0.5 points each. The topic landed with people, and that feels better than any view counter. My favorite part of the dashboard is the "What didn't count" tab. Antifraud calmly cut 23 clicks from bots and 7 clicks from cloud servers (AWS and Google Cloud). And then this line: "Your own replies: 14. Replying to yourself doesn't count." I posted a 15-tweet thread, and the system gently filed my 14 follow-up tweets under me talking to myself. No tokens for heart-to-hearts with yourself. Fair. I also tried to think like a farmer. What if 1400+ curators just reply to each other in circles? The rules got there first. Repeat replies from the same people weigh less every time, and mutual activity between curators gets cut 2 to 10 times. So does this fix paid posting? It looks like it does, and I'll keep checking against my own numbers. Next I'll post living stories from inside the platform. Reputation markets on Ethos Trade, loud slashes, and real dashboard numbers, including the ones that make me look silly. This post carries the Paid partnership label, as the rules say, and yes, I'm paid in base:0xeeee77bc7e82c0d4166d52f58239d4c5bf41eeee to be here. Check your own score or come look at my profile: https://t.co/Zso2QmZiw1 So frens, did you already warm up your X botfarms, or does the 1400+ filter finally send them into retirement and make us all type by hand? Tell me how you like the update. https://t.co/V8Rh1Yek0h

Oct 24. 7 tokens. Around $22M. One line from Hayes: “I dumped it all.” Maelstrom’s directional book is now flat. Most of it was ETH, roughly 6,400 coins, then ENA, UNI, ETHFI, AERO, PENDLE and ZAMA. Lookonchain counts about $22.3M, other trackers say closer to $26M, so treat the number as approximate data, not confirmation. Order matters here. On-chain trackers saw coins moving to market makers and OTC desks first, and only then did he confirm on X. June looked the same. HYPE, NEAR, ZEC and WLD all went within three days. A few days later a wallet linked to him bought back about 34K HYPE after a 26% drop (Arkham). He’s a trader, not an oracle. His reason is macro and geopolitics. He thinks the market has to fall before it rises, and he’s sitting out until the all-clear. The Fed decides on Oct 28. Frens, arguing whether he’s right gets you nowhere. One source says the coins he sold in June now trade above his exit prices (approximate, check it yourself). So his exit isn’t automatically a top or a bottom signal. I’m tracking this book and the on-chain flow around Fed day, will post what moves. Are you holding alts into the 28th or cutting size?

A normal US water heater element pulls around 4,500W. This one pulls 600W, runs 20 Bitmain chips and mines Bitcoin. Heatbit dropped Bitwater on Oct 22. 34 TH/s, 12 J/TH, $899 on preorder, $1,299 for later batches. Ships Q1 2027. US and Canada tanks only, 1 inch fitting, 30 to 80 gallon. The pitch is simple. You were going to burn the electricity on hot water anyway, so the hashing rides on top. Heat is the product, coins are the rebate. Now the part the headlines skip frens. Heatbit’s own estimate is about $1 a day on its solo pool (some outlets say $1.25, approximate data, not confirmation). At that rate the $899 unit pays back in roughly 2 to 2.5 years. Pay $1,299 and you are past 3.5 years. And solo mining at 34 TH/s means that dollar is an expected value, not a paycheck. My read: it is a lottery ticket that also heats your shower. The angle is still real. Heatbit says water cools the chips so it runs silent, with lower element temps and less limescale, and a gas heater version is teased. If difficulty and BTC price move your way the math flips fast. If not, you bought a very expensive element. I’m tracking home mining hardware and will post real payback numbers when units ship. Would you screw one into your tank at $899, or wait for someone else’s 6 month payout screenshot?

On Oct 22 a16z cut one word from the name of the unit behind ~$10B in funds. The word was crypto. Half my feed is already holding a funeral for crypto. Slow down frens. a16z Crypto is now a16z Frontier. Chris Dixon still runs it, Eddy Lazzarin and Guy Wuollet stay on as GPs. AI, robotics and networks connecting humans and machines go into the mandate. a16z says in writing that it keeps investing in crypto and keeps lobbying for regulation. So the unit is not shut down and nothing got pulled out. Headlines about a16z quitting crypto just skipped that part. What bugs me is the timeline. In May a16z said its $2.2B fifth fund would stay on crypto and not drift into AI or robotics. Five months later the name is gone. In July Paradigm raised $1.2B that reaches beyond crypto too (both per TheStreet, approximate, one source). Ali Yahya, the first investor on their crypto strategy back in 2017, reportedly left (The Information via TechCrunch, a16z has not confirmed). Dixon argues crypto mostly runs in the background now and companies should be judged by what they do. Fair for a bank putting a product onchain. For us the takeaway is sharper. When the biggest check writer drops the label, a pitch that only says crypto gets weaker. A pitch like agents paying each other onchain gets stronger. My read: money stays, the premium on the word shrinks. Good for builders who ship something people use. Bad for anything that lived on narrative alone. I’m tracking where Frontier writes its first checks after the rename and will post what I find. What do you think they fund first, frens?

Oct 20, Oval Office. Trump announces a 1% commission for anyone who turns in fraud. “It’s a lot of money.” Run his own example. You bring in a $1B case, you get $10M. The program behind it is reportedly $25M. One good tip eats 40% of the pot. Two details that got lost in the clips. The False Claims Act already pays whistleblowers 15 to 30%. And Trump said he decides personally who qualifies for a payout. Also frens, check what it covers. Fraud against government programs and taxpayer money. The crypto scams in your feed sit with DOJ’s Scam Center Strike Force, over $700M restrained as of April. Now the part this crowd cares about. $TRUMP launched Jan 17, 2025. 80% of supply sits with two entities tied to the Trump Organization. ATH $75.35. Reuters reportedly put the family take at ~$616M against $700M+ lost by buyers (approximate data, not confirmation). The White House line: assets in a blind trust run by his children, all conflict laws followed. So a fraud program where the president picks the winners. Would you file a tip that touches his own circle?