Most onchain yield is still someone else's token schedule. PLAY is the opposite. @DualMintRWA is putting 200 operating claw machines on Solana. People play. The machines earn. Distributions go out monthly. Target is 12–15% a year from that usage, not from emissions. $230K deposit target. Pre-deposits opened Sep 22. Steel earns it. Solana moves it. Join Uptime, follow @DualMintRWA and @stardotfun. uptime.dualmint.com

The Binance listing was the headline this week. The more useful number is Hyperliquid’s share of global perpetual open interest. Recent prints put that share around 11.3% to 11.4%, with Binance, Bybit, and OKX still in the denominator. Venue open interest has also printed a new high near $18B. That is no longer an “on-chain perp DEX versus other DEXes” story. ➢ What the listing actually changes A Binance spot listing improves access and price discovery for the token. It does not automatically improve the quality of the book. Those are two different products: - the token as a listed asset - the venue as a place where risk sits overnight Retail often treats the first as proof of the second. Serious capital does not. ➢ Why OI share is the cleaner signal Volume can jump on listing day and fade. Open interest share against the global book is harder to fake for long. If a venue keeps taking share while CEX books remain the majority, two things are happening at once: 1. Some flow is leaving centralized venues for on-chain matching. 2. New flow is being created on assets CEXes do not list the same way, including RWA and pre-IPO perps. Both can be true. Only the mix tells you whether the share is sticky. ➢ The risk that comes with the print Higher OI is ammunition. An $18B book with more retail access after a major CEX listing can produce cleaner hedging, or faster cascades, depending on liquidation quality and how correlated the new flow is. The question after a listing is not “did price hold $90.” It is whether the next volatility spike still looks like a venue that can transfer risk at size. ➢ Practical filter Watch three things over the next two weeks, not the listing candle: - Does global OI share hold after the listing spike? - Does funding stay orderly when weekend volume thins? - Do liquidations stay two-sided, or does one side get run? A listing is distribution. OI share is whether the venue is becoming part of the global market structure. Which of those two do you think matters more from here?

Hyperliquid is no longer just a perpetual venue. Manual lending is live. Users can post HYPE or BTC as collateral and borrow USDC or USDT. Reported day-one borrows were about $269M. That is a different product than another HIP-3 market. ➢ Why this matters more than the ATH headline HYPE printing a new high above $90 is the easy story. The more useful story is what the venue is becoming: 1. A place to transfer risk (perps, including RWA underlyings) 2. A place to borrow against the assets already sitting there 3. A place competing with tokenized-spot rails that the SEC just opened on a temporary exemption Those three layers used to live on different platforms. They are starting to sit on the same stack. ➢ The mechanism, not the price Reported parameters: - HYPE collateral: 65% LTV, liquidation around 82.5% - BTC collateral: 50% LTV, liquidation around 75% That design is conservative on purpose. A perp venue adding borrow against its own token and against BTC is taking on liquidation risk in a second market. If the engine is clean, idle collateral becomes working capital. If the engine is messy, a HYPE drawdown and a perp cascade can hit at the same time. ➢ The bigger map this week The SEC’s innovation exemption is a path for tokenized NMS stocks on permissioned AMMs. That is closer to owning the claim. Coinbase Derivatives and Kalshi are filing to list single-stock perpetual futures in a regulated wrapper. That is closer to what Hyperliquid already runs, just packaged for the U.S. book. So the race is not “who lists Nvidia first.” It is who becomes the default place to: - express the view - hold the claim - borrow against it without splitting capital across three systems. ➢ What to watch from here - Does borrow demand stay after the launch spike? - Do liquidations on the lending side stay isolated from the perp book? - Does US access via Bitnomial change who is allowed to use which layer? A venue that only prints volume is a trading app. A venue that can support trading, collateral, and borrow starts looking like market infrastructure. Which of those three layers do you think institutions will actually use first?

The hottest number in on-chain markets this week is not another BTC level. It is how much perpetual volume is now sitting on real-world underlyings. RWA perpetual volume across venues ran from about $85B in January to roughly $470B in June. On Hyperliquid, HIP-3 markets were already approaching half of perp volume earlier this summer. The latest 30-day print still has Hyperliquid at the top of public-chain perp volume, around $240B. That is a different market than the one most people still describe. ➢ What actually changed Crypto-native perps proved you can transfer risk 24/7 on BTC and ETH. RWA perps are testing whether the same rails can price stocks, indices, and commodities with enough liquidity to matter. HIP-3 made that test permissionless. Builders can list markets. Traders can take leveraged exposure. The other side can hedge. Volume followed. ➢ Why this week matters more than the summer prints The SEC just opened an innovation exemption path for tokenized U.S. stocks to trade through on-chain AMMs. That is not the same product as a synthetic perp. One is closer to the actual equity. The other is leveraged price exposure settled in crypto. They will compete in some books and complement each other in others. Aave is also pushing an institutional RWA borrowing hub on Avalanche. Tokenized instruments as collateral. Liquidity without selling the underlying. Three layers are moving at once: 1. Trade the exposure (RWA perps) 2. Own a tokenized claim (spot / AMM path) 3. Borrow against it (RWA lending) ➢ The part most coverage is missing Volume is the easy headline. The harder questions are still the same ones that decide whether this becomes infrastructure: - Can mark prices stay clean on thinner equity and commodity books? - Do liquidations stay orderly when a single-name stock gaps? - Does fee generation from RWA flow survive when crypto-native vol compresses? - Do institutions want synthetic perps, tokenized spot, or both? Hyperliquid is currently the clearest case study because the activity is already visible on one venue. It is not the only venue that will matter. ➢ The useful framing This is no longer “will RWA come on-chain.” The live question is which primitive captures the flow first: leveraged perps, tokenized spot, or collateralized lending. Which of those three do you think becomes the main institutional on-ramp over the next 12 months?

If crypto volatility stays compressed for several quarters, which model do you expect to hold up better? - Trading-driven perps - Lending-driven protocols - Something in between What’s the reason?

On-chain perps versus centralized exchanges. What each still uniquely wins. ➢ Where CEXes still clearly win - Absolute liquidity and depth on major pairs - Fiat on-ramps and off-ramps - Brand trust with large traditional institutions - Customer support and dispute resolution - Regulatory packaging in major jurisdictions These advantages are real and will not disappear quickly. ➢ What on-chain uniquely enables - Composability with other on-chain assets and strategies - 24/7 markets on a wider set of underlyings - Transparent liquidation and funding mechanics - Faster or permissionless market listing in some designs - Self-custody of margin and positions (with its own trade-offs) These are structural differences, not just UX gaps. ➢ The realistic picture CEXes still dominate absolute volume and institutional flow on majors. On-chain venues are winning a growing share of risk transfer that benefits from transparency, composability, or access to assets traditional venues do not list cleanly. The two systems are complementary more often than they are substitutes. ➢ Better framing Stop asking whether perp DEXes will kill CEXes. Start asking which forms of risk transfer are better served by transparent, composable venues, and which still require the full institutional stack of a centralized exchange. Where do you currently see the biggest remaining gap that keeps sophisticated capital on CEXes?

The real product of every perpetual venue is not trading. It is risk transfer through funding and liquidations. Most retail discussion focuses on direction: long or short. Sophisticated capital focuses on the continuous transfer of risk and who pays whom. ➢ Funding is peer-to-peer Funding is not a protocol fee in the traditional sense. When the perpetual trades at a premium, longs pay shorts. When it trades at a discount, shorts pay longs. This is a market-driven transfer of capital between participants holding opposite risk. The protocol facilitates it and usually takes a cut of trading fees, but the funding itself is horizontal. Understanding the current funding regime is often more useful than most technical analysis. ➢ Liquidations are the forced transfer When a position can no longer meet margin requirements, the liquidation engine forcibly closes it. This is the final risk transfer: from the underwater trader to liquidators, backstop vaults, or in worst cases to the broader system. A good engine minimizes price impact and bad debt. A bad engine turns normal volatility into amplified forced selling. ➢ The mental model Stop thinking only “will price go up or down.” Start thinking “who is currently paying funding, how crowded is that side, and what happens to this position if volatility expands quickly.” That shift changes how most people size and manage leveraged trades. What’s the part of the funding or liquidation mechanism that most people still get wrong?

Quick follow-up from this morning. If you had to rank these 4 metrics for evaluating a perp venue, what order would you use? - 24h volume - Open interest stickiness - Fee generation - Liquidation quality Drop your ranking and the reason for #1.

Quality of open interest versus vanity volume. One number is easy to inflate. The other is much harder to fake. Most people still rank perpetual venues by 24-hour volume. That metric is the easiest to manufacture and the least useful for understanding real capital commitment. ➢ Why volume misleads Volume can be produced through incentive programs, short-term mercenary capital, or circular activity. Open interest is different. It represents capital that is willing to keep risk on the books through funding payments and mark-to-market volatility. Sticky OI is harder to manufacture at scale. ➢ The practical filter Look at two numbers side by side: 1. How much volume a venue prints. 2. How much open interest remains after incentives cool and through a volatility spike. Venues that print massive volume but see OI collapse quickly are running a different business from venues that maintain high OI with cleaner liquidations. ➢ Liquidation quality is the third filter High OI is only valuable if the venue can liquidate large positions without creating cascading free-falls or socialized losses. A venue can look healthy on volume and OI until the first real stress event. Then the liquidation engine either proves itself or becomes the story. ➢ How to evaluate any perp venue Ask three questions: - Is the OI sticky or temporary? - Do liquidations look clean or messy at meaningful size? - Is volume organic or mostly incentivized? Most marketing focuses on the first number that looks good. Serious capital looks at the other two. How do you personally filter real volume from farmed volume when looking at perp DEXes?

GM 🐸 My to-do list looked very manageable this morning. Then I remembered I’m the one who made it.

Senate Republicans have revised the Clarity Act to force trading protocols that are decentralized in name only to register with the CFTC. That matters because the bill is now drawing a line between software that is actually run by a diffuse network and software that can be controlled or materially changed by a person or group.Sen. Cynthia Lummis said the updated 630-page draft includes 100+ Democratic requests, narrows the DeFi section to spot and cash transactions, and asks the CFTC and Treasury to write the rules for protocols that can be altered by humans. In practice, this turns governance into the regulatory test: if a protocol has real admin leverage, it starts to look like a venue, not a neutral stack.The September 15 procedural vote is the real checkpoint. If the bill advances, most crypto activity gets a clearer federal lane, but token issuance, stablecoin yield, and protocol control all become part of the same jurisdictional map. The open question is whether Congress is drawing a durable framework, or simply deciding which forms of onchain power count as finance.

Ethereum is testing the lower Bollinger Band at $2,460 while ETH holds just under $2,500. The market is not deciding direction yet; it is compressing energy inside a narrow band.On the 4-hour chart, the middle line sits at $2,484.18, the upper band at $2,507.87, and the lower band at $2,460.49. That means price is leaning on support while sellers keep defending the first obvious breakout zone. The daily structure is still healthier than before August, but the same $2,500 area has rejected every recovery since late August.The liquidation map makes the setup cleaner. There is leverage clustered near $2,440 and again around $2,490 to $2,535, with more liquidity near $2,550. That is why a small break can turn into a sharp move: stops and liquidations sit close to spot on both sides.The real line in the sand is $2,550. A weekly close above it opens $2,656 and $2,812. Lose $2,440 and the chart starts looking at $2,344 and $2,215 instead. Until then, ETH is just a crowded range with a macro event in front of it: the Fed meeting on Sep. 15 to 16.

gm, I almost sent a message this morning that was just "ok" and then stared at it like it had attitude.Deleted it, rewrote it twice, and somehow ended up sounding even more suspicious.

I looked out the window for a second and somehow got assigned night duty by the streetlights.gn. Just me, a dark glass reflection, and one very committed porch light doing the most 🌙

Coinbase is building a financial account for AI agents, and the important part is already visible in the product. Coinbase for Agents lets a model connect to Coinbase Advanced Trade, use a remote MCP server or local CLI, and operate inside an isolated portfolio instead of a full account.That structure matters because Coinbase is trying to solve the real problem in agent finance: not whether an agent can trade, but how much damage it can do. Users can fund a separate portfolio, restrict permissions to that bucket, and keep transfers inside authorized Coinbase accounts. The service already covers spot trading across more than 900 crypto pairs, plus eligible U.S. futures, S&P 500 equities, monitoring, and USDC to USD conversions. Withdrawals to external blockchain addresses are blocked.x402 is the second layer. Coinbase wants agents to pay for research, data, and compute inside an HTTP request, but that piece is still “coming soon.” The trade-off is obvious: more automation, less direct user control. Coinbase even says agents can misread instructions and users remain responsible. That is the market structure here: AI gets a narrower wallet before it gets a bank.

Tether has frozen about 39.3 million USDT across 10 Tron addresses linked to Xinbi Guarantee. The biggest wallet held roughly 10.78 million USDT, while three others held about 8 million each, showing how balance can be scattered across many addresses even when the underlying flow sits inside one marketplace graph.MistTrack, from SlowMist, said one of the larger wallets was funded by multiple accounts labeled “Guarantee Merchant” and then routed onward to another Xinbi address. That is the structure of these Telegram-era escrow markets: they sit between merchants and buyers, use crypto for settlement, and turn USDT into working capital for scams, laundering, and stolen-data trade.TRM Labs estimates Xinbi has processed about $24.2 billion since 2022, including $12.1 billion in inflows since May 2025. The freeze follows earlier action against Huione-linked funds, which suggests issuers are not just policing individual wallets, but chokepoints inside an entire marketplace model. The open question is how much pressure these platforms can absorb before they fragment into smaller, harder-to-trace payment rails.

SHR Miner is packaging Dogecoin mining as a short-duration yield product, with contracts ranging from 1 to 50 days and a web dashboard that handles mining status, rewards, and withdrawals.The mechanism is simple: instead of buying hardware, paying power bills, or running rigs, users lease hash power from industrial equipment and receive whatever the contract says they are owed. New accounts get a $15 registration bonus and a free hash power contract that the platform says pays $0.60 per day, which lowers the entry point but also frames the whole thing as a marketing funnel into larger paid contracts.That structure matters because it shifts the operational burden away from the user while keeping the economic exposure with the user. The upside is convenience and no upfront hardware management. The trade-off is obvious: the user is trusting a third party for both execution and payout terms. In cloud mining, the real question is less about DOGE itself and more about who controls the hash rate, the accounting, and the withdrawal path.

Finished work and now my brain is doing that thing where it acts like it also clocked out.gn. The weird part is I still reach for my desk like there might be one more email hiding there, very polite and very unnecessary.

Stacks is trying to turn idle BTC into productive capital without asking holders to leave Bitcoin’s custody model behind. Its planned Bitcoin Staking product locks BTC on Bitcoin L1, pairs it with STX worth about 5% of the BTC position, and targets roughly 3% annualized rewards paid in Bitcoin. That yield comes from Proof of Transfer, where miners spend BTC to produce blocks, and Stacks says more than 4,200 BTC has already flowed through that mechanism since January 2021.The structure matters because it avoids the usual trade-off: foreign-chain risk, wrapped exposure, or protocol slashing of principal. If the product reaches mainnet, it becomes the entry point for a wider stack: faster execution, sBTC improvements, liquid staking, trading, credit, perpetuals, and programmable BTC. StackingDAO, Bitflow, Zest Protocol, and Hermetica are already building pieces of that market. The real question is whether Bitcoin can grow a financial economy around itself without losing the properties that made it Bitcoin in the first place.

gm, the neighbor’s dog did that tiny morning bark again like it was announcing a meeting nobody asked for.Respectfully, I was not ready to be included in its little neighborhood business.