BTC $62,927.02, ETH $1,865.77, SOL $72.95, XRP $1.062, HYPE $52.44, DOGE $0.069746, AAVE $96.19, ZEC $456.79.
Saturday’s useful signal is reserve quality.
The July 30-31 digests already covered bank-money settlement, tokenization M&A, U.S. perp tax, institutional OTC flow, CLARITY ethics, ETF demand, Robinhood and Coinbase revenue, Kalshi litigation, Coldcard drains, RWA perps, Strategy treasury cash management, mobile-perps, wallet phishing, Moonbeam migration, and the usual dev-tooling mix.
This morning moves the filter. The new stories are less about who can launch another wrapper and more about who can survive audits, sanctions, product bundling, validator mechanics, sports-scale speculation, and the next technical risk clock.
Tether printed $1.5 billion of operating profit, but its excess reserve cushion fell by more than $4 billion. Circle secured a New York trust charter while USDC pushes deeper into regulated custody. A Reuters-backed investigation tied a Dubai exchange to a $4 billion Iran sanctions-evasion network. Treasury sanctioned a bitcoin-accepting Hormuz shipping-insurance scheme. Uniswap launched Earn with Morpho because swaps alone aren’t enough distribution anymore. Ethereum’s 43-day staking queue is real, but Sygnum says it mixes demand with protocol plumbing. World Cup prediction markets topped $20 billion of onchain volume. Bitcoin traders made the $60,000 put the biggest Deribit options position for August. IBM’s quantum push moved from science narrative toward revenue timing. GitHub’s fresh board points to agentic security routing, support desks, and hardware hacking.
That is a better Saturday mix. Less “crypto has new rails.” More “which rails have enough balance-sheet cushion, legal permission, security posture, and user demand to be trusted?”
Price snapshot via Coinbase spot prices for BTC and ETH, plus CoinGecko simple-price data for SOL, XRP, HYPE, DOGE, AAVE, and ZEC around 06:20 HKT.
1. Tether’s Profit Was Stronger Than Its Cushion
CoinDesk reported that Tether earned $1.5 billion of net operating profit in Q2 while its excess reserves fell from just over $8.23 billion to $4.11 billion.
That is still a large buffer. The direction matters more than the headline profit.
Tether reported $187.75 billion of assets against $183.64 billion of liabilities as of June 30. USDT issuance rose by about $446 million to $184.6 billion. The firm also added 14 metric tons of gold, bringing reported gold holdings to roughly 146.2 tons, and lifted bitcoin holdings by about 1,796 BTC to 98,933 BTC.
The reserve mix is getting more interesting. Treasuries and repo income still drive operating profit, but gold and bitcoin add market sensitivity to a company whose product is supposed to feel boring to users.
Stablecoin issuers are now being judged on more than circulation. Reserve quality, auditability, buffer size, asset volatility, and regulatory fit are all part of the product.
Tether remains wildly profitable. The question is how much volatility users will tolerate inside the balance sheet backing the quietest asset in their wallet.
2. Circle Added A New York Trust Layer To USDC
CoinDesk reported that Circle secured a limited purpose trust charter from the New York Department of Financial Services.
This is less flashy than a new chain launch. It is more important.
The charter lets Circle provide fiduciary, custody, and asset-management services under New York banking law. It arrives shortly after Circle received OCC approval to establish a national trust bank, which would provide custody and fiduciary services without taking deposits or making loans like a normal commercial bank.
USDC already has more than $71.8 billion of market cap, according to the report. The new layer is about institutional trust: who supervises reserves, who can custody assets, and what legal box the issuer sits in when banks, funds, and corporates use the token.
This is the stablecoin race after the distribution phase. Tether has scale and profit. Circle is buying regulatory surface area.
The winner may not be the issuer with the most chains. It may be the issuer whose reserve, custody, and legal stack becomes easiest for institutions to approve.
3. Iran Sanctions Risk Hit Exchange Plumbing
CoinDesk reported that Reuters tied a Dubai-based illegal gambling network to Shelbit, an unlicensed crypto exchange at the center of an alleged $4 billion Iranian sanctions-evasion operation.
The numbers are ugly. More than 2,000 gambling platforms allegedly used the network, and Shelbit reportedly moved hundreds of millions of dollars through major crypto venues.
The point isn’t that crypto uniquely creates sanctions evasion. It is that crypto gives investigators a transaction graph, and that graph can implicate exchange plumbing long after the activity starts.
Binance told Reuters that Shelbit never held an account on its platform and said it froze relevant accounts after investigating linked users. That distinction matters, but it doesn’t make the risk disappear. A sanctioned network can route value through intermediaries, counterparties, gambling operators, OTC desks, miners, and exchange customers without presenting as one obvious account.
Compliance is becoming network analysis, not account review.
For large venues, this is the hard part of being financial infrastructure. The dangerous flow won’t always knock on the front door with the right name.
4. Treasury Sanctioned A Bitcoin-Accepting Hormuz Insurance Scheme
CoinDesk reported that the U.S. Treasury sanctioned Persian Gulf Marine Insurance Company and HormuzSafe Marine Services Authority.
Treasury says the firms ran an Iran-linked maritime insurance operation for ships passing through the Strait of Hormuz and accepted bitcoin and other digital assets.
This is a sanctions story with a sharper real-world surface. The alleged product was not a DeFi app. It was insurance tied to one of the world’s most important energy chokepoints, where vessel seizures, oil prices, and military pressure already shape global markets.
Treasury described the arrangement as extortion because the coverage was for risks created by Iran itself. The designations bar U.S. persons from dealing with the firms and can expose foreign companies to secondary sanctions if they transact with them.
The useful read for crypto is simple: paying in bitcoin doesn’t create a separate legal lane.
Digital-asset settlement can make strange markets easier to operate. It can also make sanctioned activity easier to trace and easier to punish.
5. Uniswap Is Turning The App Into A Yield Router
The Block reported that Uniswap launched Earn, a lending product built with Morpho.
Users can deposit USDC, USDT, and ether into three Gauntlet-curated Morpho vaults directly from the Uniswap app. Borrowers pay interest through Morpho markets, and depositors earn the yield while keeping custody and the ability to withdraw.
That is a product shift. Uniswap was the trade button for DeFi. Earn makes the app a routing surface for idle balances.
Morpho is becoming the lending backend for more consumer and institutional interfaces. The Block notes integrations or products involving Coinbase, Robinhood, Bitwise, and Societe Generale. If those channels keep adding deposits, Morpho’s value becomes liquidity aggregation rather than one frontend.
The regulatory wrinkle is still there. SEC Commissioner Hester Peirce recently said crypto vaults and onchain lending strategies may fall under existing securities laws depending on structure. Uniswap says it’s acting in compliance.
The bigger trend is clear. DeFi protocols don’t want users to leave assets idle. Every wallet and app is becoming a balance-sheet dashboard.
6. Ethereum’s Staking Queue Is Demand Plus Plumbing
The Block reported that Ethereum’s validator entry queue reached roughly 2.5 million ETH, with new stakers waiting about 43 days to activate.
At first glance, that looks purely bullish. Sygnum’s Thomas Brunner says it’s more complicated.
The queue reflects genuine demand, including institutional activity, but it also reflects protocol mechanics. Brunner pointed to last year’s Pectra upgrade, which lets validators hold up to 2,048 ETH and compound automatically. Large operators can top up existing validators, sometimes with small amounts, and those top-ups wait in the same queue as fresh stake.
The cleaner signal is the exit queue. It remains largely empty, which says existing stakers aren’t rushing out.
About 41.2 million ETH, or 33.8% of supply, is staked. That gives Ethereum a stronger holder base, but it also makes staking infrastructure more visible to institutions.
Privacy is still the bottleneck. If deposit address, validator, and withdrawal credentials can be linked, professional money has to treat staking as public balance-sheet telemetry.
7. Sports Turned Prediction Markets Into A $20 Billion Stress Test
The Block reported that the 2026 FIFA World Cup generated more than $20 billion of onchain prediction-market volume, citing Chainalysis.
This is the user-demand side of prediction markets after yesterday’s legal noise.
The tournament created markets for match outcomes, player performance, tournament winners, and stranger social bets. One market on whether Cristiano Ronaldo would cry after his expected final World Cup appearance reportedly handled nearly $50 million of volume.
The Block’s own data showed Kalshi and Polymarket processed roughly $54 billion of sports prediction-market volume between June 11 and July 19, though that window also included the NBA Finals, MLB, Wimbledon, and other events.
The category mix changed too. Sports accounted for roughly 70% to 85% of Kalshi and Polymarket volume during the World Cup. After the tournament ended, sports share fell to about 60% on Kalshi and 50% on Polymarket.
Politics built the prediction-market brand. Sports may be what teaches mainstream users to trade outcomes every day.
8. The $60,000 Put Became Bitcoin’s August Hedge
CoinDesk’s Daybook said the $60,000 bitcoin put became the largest open-interest position on Deribit, with $1.17 billion of notional open interest.
That is a cleaner market story than another ETF flow update.
Before the shift, $70,000 and $72,000 call options led the board with about $2.5 billion of notional open interest each. Those bullish bets faded after the Fed decision failed to push bitcoin higher. By Friday’s $10 billion BTC and ETH options expiry, the $70,000 call had fallen to $943 million and the $72,000 call to $888 million.
Seasonality is adding pressure. Since 2013, July has produced a median bitcoin return of 8.61%, while August has produced a median return of -7.51%. This July’s gain was about 8.9%.
Options positioning doesn’t predict price by itself. It shows where protection is expensive, where dealers may be exposed, and what strike starts to matter if spot weakens.
For August, the market’s line in the sand is no longer $70,000 upside. It is $60,000 defense.
9. Quantum Moved From Threat Model To Revenue Clock
CoinDesk reported that STS Digital’s Maxime Seiler sees AI capital diversion as one of three headwinds weighing on crypto. Separately, CoinDesk’s latest page cited IBM CEO Arvind Krishna saying quantum computing could begin generating meaningful revenue before the decade ends.
Crypto usually treats quantum as a future security threat. Public markets are starting to treat it as a nearer commercial race.
That matters for two reasons. First, quantum investment competes with crypto for risk capital, talent, and attention inside the same technology budget conversations. Second, serious quantum progress keeps pressure on chains, wallets, custody providers, and standards bodies to plan migration paths before cryptographic risk becomes urgent.
The immediate implication isn’t panic. Bitcoin and Ethereum aren’t breaking this weekend because IBM gave an optimistic CNBC interview.
The implication is calendar discipline. Post-quantum readiness, signature migration, hardware support, custody procedures, and user education all take years. Systems that wait for a single “quantum day” headline will be late.
Security transitions are boring until they aren’t. The work has to start while the threat still feels abstract.
10. GitHub Trending - Security, Support, And Hardware Edges
The repeat tracker ruled out recent features including different-ai/openwork, mvanhorn/last30days-skill, paperswithbacktest/awesome-systematic-trading, microsoft/AI-For-Beginners, github/copilot-sdk, agavra/tuicr, 1jehuang/jcode, and the July 31 morning and evening repo sets.
zhaoxuya520/reverse-skill has about 10.5K stars and gained 612 stars today. It is a reverse-engineering and authorized penetration-testing skill router pack for AI coding clients. The signal is that agent skills are moving into security workflows where routing, toolchain bootstrapping, and accumulated operator knowledge matter more than one clever prompt.
chatwoot/chatwoot has about 35.1K stars and gained 53 stars today. It is an open-source support desk across live chat, email, and other channels. The read is practical: as AI agents enter customer support, teams still need owned conversation infrastructure instead of pushing every interaction through a closed SaaS box.
geo-tp/ESP32-Bit-Pirate has about 4.9K stars and gained 161 stars today. It is a web-based CLI hardware hacking tool for common protocols. That belongs in the digest because physical devices, signing hardware, and embedded systems keep becoming part of crypto security. Wallet safety doesn’t stop at app code.
Morning Read
Read Tether’s Q2 reserve story, then read Circle’s trust-charter update, then read Uniswap’s Earn launch.
The number to remember is $4.11 billion.
That is Tether’s reported excess reserve cushion at the end of Q2, down from just over $8.23 billion three months earlier. The second number is 43 days, because that’s roughly how long new Ethereum validators now wait to activate.
Saturday’s read is reserve quality. Stablecoin issuers are moving from growth headlines to supervision, asset mix, and custody authority. Sanctions stories are moving from “bad actors used crypto” to concrete network mapping. DeFi apps are turning idle wallet balances into yield funnels. Ethereum staking demand is real, but so are the mechanics underneath the queue. Sports prediction markets just proved they can absorb World Cup-scale behavior. Bitcoin options traders are hedging August around $60,000, not dreaming about $72,000 calls.
Crypto still has plenty of distribution. The harder test is whether the distribution can handle audits, sanctions, privacy, custody, and stress without pretending those are afterthoughts.
Evening Update - 18:14 HKT
BTC $62,945.13, ETH $1,862.97, SOL $72.73, XRP $1.061, HYPE $52.14, DOGE $0.069657, AAVE $90.61, ZEC $463.74.
Saturday evening moved from reserve quality to cleanup risk.
The novelty gate ruled out another broad pass through Tether’s reserves, Circle’s trust charter, Iran sanctions routing, Uniswap Earn, Ethereum’s staking queue, World Cup prediction-market volume, August bitcoin puts, quantum commercialization, and this morning’s GitHub security picks. It also kept the July 30-31 loop in view: bank-money pilots, perps tax, Coinbase and Robinhood earnings, Kalshi litigation, Aave pruning, Coldcard’s first drain estimate, Strategy cash management, RWA perps, Moonbeam migration, and generic agent tooling needed a changed fact or a new consequence to get in.
The evening stories clear that bar. Galaxy Research lifted the Coldcard-linked loss estimate to 1,082.65 BTC. Bitcoin ETFs ended July positive by only $172.4 million while still sitting $5.3 billion negative for 2026. The Bank of Italy found stablecoin remittances don’t beat fiat rails consistently once off-ramps and FX enter the picture. Pump.fun’s reported layoffs turned token vesting into a governance and labor story. FTX creditors started receiving another $900 million. Citadel reportedly bought a chunk of Situational Awareness’ public stock book after the AI rout. Aave’s proposed cleanup now has a sharper six-chain and 50-reserve map. BIS Project Agora settled real-value bank-token payments across six currencies. TIPS data says rising real yields, not inflation, may be the macro pressure point for bitcoin. GitHub’s fresh board points to open project management and synthetic-media tooling after the repeat filter.
That is a better evening mix. Less “the rail exists.” More “who pays when old devices, old corridors, old token grants, stale deployments, leveraged AI books, and real yields start pushing back?”
Price snapshot via Coinbase spot prices for BTC and ETH, plus CoinGecko simple-price data for SOL, XRP, HYPE, DOGE, AAVE, and ZEC around 18:14 HKT.
11. Coldcard’s Loss Estimate Nearly Doubled
Cointelegraph reported that Galaxy Research identified 1,196 addresses tied to the Coldcard incident that lost 1,082.65 BTC, worth about $70.2 million at the time of the transactions.
This is the material change from the morning and Friday evening read.
The first public estimate centered on 594.48 BTC. Galaxy’s wider map puts the sweep across a 41-minute window on July 30, from blocks 960,183 to 960,191. The reported transaction pattern included identical 30 sat/vB fees and no change outputs.
Coinkite’s warning is still the practical point: a firmware hotfix can remove the fallback path, but it can’t make old vulnerable seeds safe. Users who generated seeds on affected Mk3 firmware need a new seed, a verified backup, a test transaction, and then a full migration.
This is self-custody’s nastiest failure class. The attacker doesn’t need a phishing click if weak entropy created exposed keys years earlier.
Hardware-wallet users usually think in terms of today’s device state. This incident says the seed’s birth certificate matters more.
12. Bitcoin ETFs Finished July Positive, Barely
Cointelegraph reported that U.S. spot bitcoin ETFs attracted $172.4 million of net inflows in July.
That is technically green. It is not a strong demand signal.
The funds still ended July about $5.3 billion negative year to date after heavy withdrawals in May and June. A $265.4 million outflow on Friday was the largest daily withdrawal since July 13, and weekly flows turned negative into month-end.
The cross-asset split is more interesting. Ether ETFs posted a $365.2 million July inflow and kept a four-week inflow streak alive. XRP ETFs added $27.3 million in July and have now posted five positive months in 2026.
Bitcoin still has the deepest product shelf and the biggest cumulative ETF base. July says that isn’t the same as a constant bid.
ETF demand is alive. It is also selective, late-cycle cautious, and no longer enough by itself to carry every bullish chart.
13. Stablecoin Remittances Hit The Fiat Bottleneck
Cointelegraph reported that a Bank of Italy study found no consistent cost or speed advantage for stablecoin remittances over traditional payment channels.
That doesn’t kill the stablecoin payments story. It makes it more honest.
Researchers tested 200 USDC remittances across 10 bidirectional corridors linking Italy with Brazil, Argentina, Japan, the United Arab Emirates, and South Africa. Total costs ranged from 0.3% to nearly 9%, and settlement ran from under 20 minutes to one or two business days depending on local payment infrastructure.
The chain fee wasn’t the main problem. Fiat conversion, FX spreads, and on/off-ramp payment rails explained most of the difference.
This is the part stablecoin boosters skip when they say “instant settlement.” The token can move fast while the user’s actual purchasing power waits on a bank, exchange, local payment network, compliance check, or FX conversion.
Stablecoins win more clearly when users can spend them directly. Until then, the off-ramp is part of the product.
14. Pump.fun’s Token-Comp Story Got Ugly
Cointelegraph reported that Pump.fun reportedly fired employees two months before they were due to receive PUMP token grants worth millions of dollars.
This is not just workplace gossip.
According to the report, at least one employee had a seven-figure token allocation scheduled to begin vesting in June 2026. The employees were reportedly fired in April, while co-founder Noah Tweedale attributed layoffs to the company growing too quickly.
Token compensation is supposed to align builders with protocol upside. It becomes poisonous when workers believe vesting dates can be dodged through timing.
Pump.fun already sits inside several uncomfortable debates: memecoin incentives, MEV claims, launchpad quality, and speculative user behavior. This adds an internal governance angle.
Crypto companies love to talk about community ownership. The harder test is whether employees, early contributors, users, and tokenholders can tell who actually owns the upside when the token finally matters.
15. FTX Payouts Are Becoming The Exit Liquidity Story
Cointelegraph reported that former FTX users began receiving funds from a new $900 million creditor distribution.
The FTX Recovery Trust has now returned an estimated $11 billion to customers.
That is a remarkable bankruptcy recovery compared with how the collapse felt in November 2022. It is also a strange market-structure event: old trapped balances are becoming liquid cash while the exchange’s former executives remain part of the criminal and civil aftermath.
The latest distribution used partners including Kraken, BitGo, and Payoneer. Separately, the estate is still pursuing a $1.76 billion clawback claim against Binance and former CEO Changpeng Zhao after a bankruptcy judge allowed that claim to continue.
The market implication is not simple. Some creditors will leave crypto forever. Some will rotate recovered dollars back into BTC, ETH, SOL, or alts. Some will just pay taxes and move on.
FTX is no longer a live venue. Its liquidation process is still releasing capital into a very different market.
16. Citadel Bought Into The AI-Fund Deleveraging
Cointelegraph’s daily roundup said Citadel reportedly acquired a large part of Situational Awareness’ public stock portfolio after July’s AI-driven sell-off.
This is the AI-finance story crypto should care about.
Situational Awareness, founded by former OpenAI researcher Leopold Aschenbrenner, had reportedly grown to about $24 billion of assets in under two years. Reports cited by Cointelegraph said the fund fell about 67% in July while still remaining up roughly 80% for the year.
The details are messy: public stocks, private AI stakes, margin pressure, and proposed sales of Anthropic shares all appeared in the reporting. The useful read is cleaner. When AI exposure gets crowded and levered, one bad month can turn narrative capital into forced selling.
Crypto has lived this movie with miners, funds, treasuries, and basis trades. AI now has its own version.
The overlap matters because the same investors are often underwriting both stories. If AI beta needs balance-sheet repair, crypto risk budgets can get clipped too.
17. Aave’s Chain Cleanup Has A Sharper Map
Cointelegraph reported that an Aave governance proposal would wind down V3 markets on six blockchains and offboard dozens of low-use listings.
The numbers are specific: $98.1 million in supplied assets, $15.6 million in debt, 50 low-use reserves, 21 matured Pendle principal-token listings, and full reserve exits from Sonic, Scroll, zkSync, Metis, Soneium, and Aptos.
This continues the Aave cleanup theme from Thursday, but the consequence is now easier to see. Aptos is the uncomfortable example. Its V3 market launched only 11 months ago, yet available liquidity reportedly fell 94% over six months and quarterly revenue was below $1,000.
Multichain expansion used to sound like automatic growth. It now looks like a recurring risk bill: bridge exposure, oracle maintenance, liquidation support, monitoring, governance attention, and stale collateral.
Aave is doing the unglamorous thing mature protocols need to do. It is admitting some deployments don’t earn their risk surface.
18. BIS Tokenized Bank Money Got Another Real-Value Check
Cointelegraph reported that BIS Project Agora completed real-value tokenized cross-border payment tests worth 800,000 Swiss francs, about $1 million.
This overlaps with Friday’s bank-money pilot, but the evening angle is the trial detail.
Twenty-eight financial institutions and central banks tested 17 transaction scenarios using tokenized central-bank reserves and commercial-bank deposits. Settlement happened across Swiss francs, euros, pounds sterling, Japanese yen, South Korean won, and U.S. dollars, with average settlement time around 80 seconds.
The participant list included central banks such as the Bank of England, Bank of France, Bank of Japan, Bank of Korea, and the Swiss National Bank, plus commercial banks including JPMorgan Chase, Citi, Deutsche Bank, BNP Paribas, UBS, Standard Chartered, and MUFG.
Stablecoins are winning public distribution. Bank-token pilots are trying to win institutional trust and settlement finality.
The next question is whether these tests reduce correspondent-banking cost in production, not only in a controlled pilot.
19. Real Yields Are The Macro Pressure Point
Cointelegraph reported that rising U.S. Treasury yields may be driven more by real yields than by inflation expectations.
That distinction matters for bitcoin.
The article noted that five-year inflation expectations priced through TIPS were around 2.2% and had been trending down since May. If nominal yields are rising while inflation expectations fall, real yields are doing the work.
Higher real yields are rough for non-yielding assets. Bitcoin can still trade as a liquidity hedge, censorship-resistant reserve, or risk asset, but it doesn’t pay a coupon. When real yields rise, the opportunity cost of holding it rises too.
This also complicates the easy macro story. If yields are up because the market fears inflation, bitcoin bulls can reach for the debasement argument. If yields are up because real returns are rising, the argument gets harder.
For August, watch real yields alongside ETF flows. The two may explain more than another round of broad risk-on slogans.
20. GitHub Trending - Fresh Picks After The Repeat Filter
The repeat tracker ruled out this morning’s zhaoxuya520/reverse-skill, chatwoot/chatwoot, and geo-tp/ESP32-Bit-Pirate, plus recent repeats including different-ai/openwork, mvanhorn/last30days-skill, paperswithbacktest/awesome-systematic-trading, microsoft/AI-For-Beginners, github/copilot-sdk, agavra/tuicr, and 1jehuang/jcode.
That left two clean daily picks worth including.
usekaneo/kaneo has about 5.3K stars and gained 194 stars today. It is an open-source project-management app built around a simple “enough tool, less ceremony” promise. That matters because agent teams still need mundane coordination: issues, owners, due dates, acceptance checks, and a place where human decisions survive past the chat window.
deepfakes/faceswap has about 57.1K stars and gained 93 stars today. It is an old and controversial synthetic-media project, which is exactly why its return to the daily board is useful signal. AI media capability keeps getting easier to find, fork, and run. Finance, crypto, and support teams should treat voice, video, and identity verification as live attack surfaces, not future policy work.
Evening Read
Read the Coldcard loss-estimate update, then read the Bank of Italy remittance study coverage, then read the Aave cleanup proposal.
The number to remember is 1,082.65 BTC.
That is Galaxy Research’s expanded estimate of bitcoin lost in the Coldcard-linked incident. The second number is $172.4 million, because that is all U.S. spot bitcoin ETFs pulled in for July after two ugly months of outflows.
Saturday evening’s read is cleanup risk. Security bugs have history. ETF demand can be positive and still thin. Stablecoin payments don’t beat fiat rails if users still need expensive conversion. Token compensation can turn into a trust problem. FTX repayments are still releasing capital years after collapse. AI leverage can force public-stock sales. Aave is pruning chains that don’t earn their maintenance cost. Bank-token pilots are settling real value, but production utility remains the test. Real yields are a harder macro opponent than inflation slogans. GitHub’s fresh picks point to the same lesson: coordination and identity risk are now part of the infrastructure stack.
The morning asked who can survive audits, sanctions, custody demands, and stress. The evening answer is colder: the systems that survive will be the ones willing to clean up old assumptions before users, creditors, courts, or attackers do it for them.