Morning digest: October 13, 2026
Bitcoin slid from about $84,200 to roughly $81,400 on Monday evening as about $771M in leveraged positions were force-closed, mostly bets on higher prices. It climbed back above $83,000 early Tuesday, and ether dipped to about $2,425. The main crypto story finishes the Ledger reseller saga: the implant was a hidden cellular chip that read the recovery phrase off the screen during setup. Aptos turned on an encrypted mempool, the Financial Stability Board warned about cross-border stablecoin runs, and Morgan Stanley pushed tokenized stocks to 2027. In health, masters athletes held steadier muscle control under mental load than their peers, and a four-year trial found supervised training improved cognition in healthy older adults. At home, Prince William talks FY2028 taxes today, Fairfax holds a hearing on its legislative agenda, and CPI is out tomorrow.
Blockchain: how the Ledger implant worked, Aptos hides pending trades, stablecoin run risk
The tampered CryptoBilis wallets used a hidden cellular chip to steal recovery phrases. This updates the drains we covered Saturday and Monday. In a Tuesday update, Ledger said an implant inside the case of some devices sold by CryptoBilis, a Southeast Asian reseller, recognized the 24-word recovery phrase while it was displayed during setup. The implant then sent the phrase to the attackers over an LTE connection with an eSIM. It never touched Ledger’s secure chip, firmware or transaction signing. With the phrase in hand, attackers simply rebuilt the wallets on their own devices. On-chain researchers put losses at up to $93.2M from more than 300 wallets, and Tether has frozen $10M of stolen USDT. Ledger says every confirmed loss so far involves a CryptoBilis device, CryptoBilis has stopped selling them, and Ledger is checking reports of implants elsewhere. No software check can catch this, so the fix is still a device bought directly from Ledger and a brand-new recovery phrase. (CoinDesk on the chip; CoinDesk on the freeze)
Aptos switched on an encrypted mempool, so pending trades can be hidden from bots until the block is final. The mempool is the waiting room for transactions, and bots watch it to front-run or sandwich trades. On Aptos, users can now opt in per transaction. The payload is encrypted on the device with a key the validators jointly create each epoch, the block leader orders it without seeing it, and validators decrypt and execute it only after consensus finalizes the block. Nobody below the chain’s two-thirds stake threshold can decrypt early, so users don’t have to trust anyone beyond the validators they already rely on. Other chains lean on private builders, hardware enclaves or outside key committees instead. It costs about 20 ms per block in benchmarks. The sender address and fee stay visible, and keyless accounts can’t use it yet. Petra wallet supports it, and the Decibel perps exchange plans to use it for private orders. (Aptos)
The Financial Stability Board warned that stablecoins could suffer cross-border runs even with enough total reserves. The FSB is the G20’s financial-stability watchdog. Its Oct. 12 report looks at “multi-issuer” stablecoins, where affiliated companies in different countries issue the same token. Holders in one country could demand redemption while the reserves sit in another and are hard to move in a crisis. Round-the-clock token trading against reserve markets that keep business hours makes it worse, and nobody knows whose insolvency law would apply if an issuer failed. The FSB calls for cross-border information sharing and crisis planning, with no timetable. The roughly $306B stablecoin market has never been through a severe crisis, and G20 finance ministers take up the findings on Oct. 15. (CoinDesk; FSB)
Morgan Stanley now expects tokenized stocks on its trading venue in Q1 2027, not late 2026. Amy Oldenburg, its head of digital asset strategy, called the new date more realistic. She cited DTCC’s tokenization service, due at the end of this month, and the GENIUS Act stablecoin law taking effect in January as catalysts. The bank is also building wallets that hide key management from clients. This is a roadmap, not a launched service. (The Block)
USDGO, a stablecoin issued by a U.S. national bank, got its first DeFi venue. Anchorage Digital Bank, which holds an OCC charter, issues USDGO, and Hong Kong-listed OSL distributes it. About $1.27B of the token sits on Solana, spread across only a few hundred wallets. A trading pool went live Oct. 12 on Jupiter Lend’s AMM. Lending markets are promised without a date, and the pool’s depth and fees weren’t disclosed. Bank regulation covers the issuer, not the DeFi contracts, so this is a test of how the two mix. (Solana Compass)
Jeonbuk Bank became the first South Korean bank to lend against bitcoin. It offers corporate clients won loans with bitcoin as the only accepted collateral. A registered custodian holds the coins, and on default the bank recovers its money through a court-ordered sale under rules revised Oct. 1. One loan has been made so far, and rates and loan-to-value limits weren’t disclosed. (CoinDesk)
Briefly: Blur, once the biggest Ethereum NFT marketplace, closes its website next month. Blend, its NFT-backed lending product, stops matching new loans on Nov. 12, so borrowers need to repay or manage their loans before then. The on-chain contracts keep running (The Block). A London court ruled that Chinese law governs fraud victims’ claims to the 61,000 bitcoin (about £3.8B) seized in the Zhimin Qian case. That makes it more likely the U.K. Treasury keeps most of it, and the victims are appealing (CoinDesk). Stripe-incubated Tempo activates its T12 upgrade today at 10 a.m. ET, and nodes not on v1.16.0 will fall out of sync (release notes). Standard Chartered’s RLUSD custody, covered yesterday, is live, but the LMAX collateral plan still depends on testing and approvals (Standard Chartered). No new Ethereum Foundation or LayerZero posts, and no new Swift shared-ledger news.
Health: athletes’ steadier nerves, years of training for the mind, and a sitting cue
Masters athletes kept steadier muscle control under mental load than their peers. Researchers tested 56 older adults: 19 masters athletes with at least 10 years of continuous competitive sport, 22 non-sarcopenic controls and 11 people with sarcopenia (age-related muscle loss). Each person held a steady force while doing a mental task at the same time. Dual-tasking made force less steady and motor-unit control worse in sarcopenia. The masters athletes were the only group whose motor-unit firing actually became more regular under the extra load. Endurance and power athletes responded alike. This is an exceptional cohort, and the lever is long-term, varied competitive training, which seems to protect how the brain and muscles work together when attention is split. That matters for falls and everyday multitasking. The limits: it’s a small, single-time-point study, and the athletes were younger (about 70, against about 80 in the sarcopenia group). Age-adjusted analyses held up, but cause can’t be shown, and the summary gives no single headline effect size. (Scienmag; study)
Four years of supervised training improved cognition in healthy older adults. This single-blind randomized trial assigned 882 community-dwelling older adults (674 analyzed) to 48 months of supervised moderate-to-high-intensity aerobic, resistance and stretching training, or to their usual habits. Healthy participants gained on the MMSE screening test compared with controls, and on memory, attention and reasoning tests. People with subjective memory complaints showed a similar gain, and the mild-cognitive-impairment subgroup was too small to judge. The design is strong and the duration rare, but the effects were small (partial η² about 0.03), the MMSE is a coarse measure, and about 24% dropped out. These were ordinary older adults, not already-fit people, so it says little about what more training adds for someone who already trains. (Frontiers in Aging)
A free-living study found thigh-muscle activity fades about 25 minutes into a sitting bout. Using textile sensors in clothing, researchers tracked 7,684 sitting bouts in 84 adults (mean age 44). Detectable muscle activity fell below 1% after about 25.7 minutes. People whose usual bouts ran at least 25 minutes had about 6.6 minutes of muscle activity per sitting hour, against 10.4 minutes for those who got up sooner. More frequent and more vigorous interruptions went with about 5 extra minutes of activity per hour. The practical cue is cheap: stand or move briefly every 25 to 30 minutes. But the study measured muscle activity, not glucose, heart health or cognition, and its three cohorts were general adults, women with overweight or obesity, and people with type 2 diabetes. So it’s a plausible reason to break up sitting, not proof of benefit or a precise cutoff. (study; MedicalXpress)
Also today, from general or impaired populations only: In 72 sedentary adults with memory complaints (average age 62), 24 weeks of water walking three times a week improved an executive-function test (Trail Making B) compared with controls. Land walking didn’t, and the group-level results were inconsistent (study). A meta-analysis in people with mild cognitive impairment found multicomponent exercise (aerobic, strength, and balance or coordination work) improved global cognition with a large but very uncertain effect (g = 1.06), with high heterogeneity (study). A meta-analysis of 9 trials (1,814 adults aged 60 and older) pairing multicomponent exercise with nutrition programs found a modest cognition gain (g = 0.31) and a larger drop in depressive symptoms (g = −0.51). Longer programs did better, but the authors call the evidence limited (BMC Geriatrics).
General / local: truce still unconfirmed, EU sanctions final, PWC tax talk, Fairfax hearing
Moscow still hasn’t confirmed Trump’s “energy ceasefire.” Kremlin spokesman Dmitry Peskov said only that “contacts are continuing.” Lavrov said Russia would press on, and Zelensky said Russia hit six Ukrainian energy facilities on Monday. Trump says the truce is holding. No terms have been published. (Al Jazeera)
The EU sanctions listing is now final. The Council’s official release confirms 1,569 listings (826 Russian companies and 743 of their executives), plus 77 people tied to the Duma elections in occupied Ukraine. The legal acts are in the Official Journal, and the EUR-Lex annex is where to screen counterparties. There’s no new OFAC action. (Council of the EU; EUR-Lex)
Economy: Bonds reopen today after the holiday. September CPI is out Wednesday at 8:30 a.m. ET, and the forecast is 3.6% from a year earlier. The Beige Book follows at 2 p.m., and PPI and retail sales come Thursday.
Prince William and Manassas: The Board of County Supervisors meets from 2 to 7 p.m. at the McCoart Building. The work session includes the first FY2028 tax policy discussion, covering tax rates, spending priorities and five-year-plan balances. At 6:30 p.m., the Board and the School Board meet the General Assembly delegation at Old Hickory Golf Club, where PWCS presents its 2027 legislative priorities. There’s no citizen comment time at that meeting. No Manassas or Manassas Park city items. (PWC agenda; PWCS)
Fairfax County: The Board of Supervisors meets at 9:30 a.m. A 3 p.m. public hearing takes resident comments on the county’s 2027 General Assembly agenda, with a vote on Dec. 8. Chair Jeff McKay wants the state to pay more of the cost of core services, with housing and tax reform among the priorities. The board will also consider advertising a hearing on changes to the taxation chapter of the county code, and conveying land on Richmond Highway to VDOT for Bus Rapid Transit. Separately, a developer has applied to build a 312,000-square-foot data center in Chantilly powered by fuel cells. (Fairfax County; FFXnow)
Bottom line
The Ledger drains came from a cellular spy chip in resold devices, so the fix is still a device bought from Ledger and a new recovery phrase. Aptos now offers trust-minimized protection from front-running, and the FSB is putting cross-border stablecoin runs on the G20 agenda this week. In health, the exceptional-cohort signal is that years of varied competitive sport keep muscle control steady when attention is split. A four-year trial supports long-term structured training for the mind, though the effects were small in ordinary older adults. Getting up every 25 to 30 minutes is a cheap, unproven extra. Watch today’s Prince William tax discussion and Wednesday’s CPI.