Good morning. It's Monday again, and there's a quiet satisfaction in seeing genuine technical progress unfold across different domains. Three stories caught my attention today that all share a common thread: methodical people solving real problems, step by step.

The most significant news comes from Spain, where researchers have shown that a simple blood test can diagnose Alzheimer's disease with 94.5% accuracy. That's up from 75.5% accuracy when doctors rely solely on clinical evaluation. The test measures p-tau217, a protein that accumulates abnormally in Alzheimer's patients and shows up in blood before symptoms become severe. What makes this particularly compelling isn't just the accuracy improvement — it's that this study was conducted in real-world medical settings, not highly controlled research labs. Two hundred consecutive patients aged 50 and older with cognitive symptoms were followed through routine neurology consultations and specialized cognitive units. When doctors incorporated the blood test results, they changed their diagnosis for one in four patients. Some who initially seemed to have Alzheimer's turned out to have different conditions entirely. Others dismissed as experiencing normal aging were correctly identified as having early-stage disease. Perhaps most importantly, doctors' confidence in their diagnoses rose from 6.9 to 8.5 on a 10-point scale. This isn't just about accuracy — it's about clinical confidence in one of medicine's most challenging diagnostic territories.

The implications are substantial. Current Alzheimer's diagnosis typically requires expensive brain scans or invasive spinal taps. They're costly, uncomfortable, and often inaccessible to patients who need them. A blood test that achieves 94.5% accuracy represents a fundamental shift toward early, accessible diagnosis. The protein p-tau217 has been known as a reliable biomarker for some time, but proving its clinical utility in everyday practice is what transforms laboratory curiosity into medical tool. The researchers tested it across every stage of cognitive decline, from early memory complaints to late-stage dementia, and it remained effective throughout. For a rapidly aging global population facing an Alzheimer's epidemic, this could be transformative.

Meanwhile, in the world of hardware liberation, someone with the handle "kittywitch" just successfully ported Coreboot to the ThinkPad X270, and their write-up is a masterclass in technical perseverance. Coreboot is open-source firmware that replaces the proprietary BIOS in your computer — it's what you install if you want complete control over what your machine does when you press the power button. The X270 port started as an adaptation of existing X280 support, which sounds straightforward until you realize that "similar" laptops can differ in ways that make firmware development feel like defusing a bomb while consulting incomplete schematics.

The process began predictably: dump the existing BIOS for backup and reverse engineering, set up a Raspberry Pi-based SPI flash programmer, and start building modified firmware based on the closest existing support. Then reality intervened. While attempting to connect the programmer clips to the laptop's flash chip, they accidentally knocked off a tiny capacitor. Not just any capacitor, but one so small that when they tried to resolder it, it "pinged off into orbit to never be seen again." This required ordering replacement parts, consulting circuit schematics to identify the exact component, and learning surface-mount soldering on components smaller than a grain of rice. The capacitor was 0.8x1.6mm — roughly the size of a large speck of dust.

After successfully installing the replacement capacitor (which fell off again at least eight more times during the development process), they got their modified firmware to boot but discovered that neither the WiFi card nor NVMe storage worked. Both devices simply vanished from the system, appearing neither in BIOS menus nor Linux hardware detection. This is where the story becomes genuinely educational. Working with Leah Rowe, founder of Libreboot, they methodically diagnosed the problem through PCIe allocation analysis and schematic comparison. The issue turned out to be clock request (CLKREQ) pin assignments — the X270's WiFi card used two separate PCIe devices for WiFi and WiGig, requiring different clock routing than the X280. Adjusting these assignments by one position each and rebuilding the firmware finally produced a working system running completely open firmware.

The persistence here is admirable, but what struck me most was the documentation. Instead of simply announcing success, they wrote a detailed technical post explaining every failure, every debugging step, and every lesson learned. This kind of knowledge sharing is what makes complex technical projects accessible to the next person. They're now upstreaming their changes to both Coreboot and the deguard project, ensuring that future X270 owners won't have to repeat this discovery process.

Both of these stories connect to something happening in Indonesia that deserves more attention. Bitcoin adoption there has reached remarkable scale through a combination of legal creativity and grassroots organizing that should make anyone interested in monetary freedom pay attention. Bitcoin Indonesia currently runs 40 meetups per month across 40 different cities, with an estimated 55,000 people engaged with Bitcoin through their efforts. That's extraordinary, but the backstory is what makes it significant.

Indonesia banned cryptocurrency payments in 2017 following strong Bitcoin adoption momentum that included hundreds of merchants accepting Bitcoin in Bali and high-profile purchases like a $500,000 villa bought with 800 BTC. The crackdown wasn't superficial — businesses advertising Bitcoin faced undercover investigations, shop closures, and threats of arrest. The timing coincided with Bitcoin's 2017 fee crisis, where transactions cost up to $50, effectively killing the payments use case anyway. Bitcoin adoption in Indonesia went dormant for years.

The resurrection happened through legal ingenuity. Post-COVID, a group of Indonesian Bitcoiners discovered a loophole that had protected credit card rewards programs and airline miles from the 2018 restrictions. If your payment system operates as a "closed loop," it's not legally classified as a currency. Bitcoin ownership remained legal as a store of value — only medium of exchange usage was restricted. So instead of spending Bitcoin, merchants allow customers to "redeem" Bitcoin for products and services within their network. This keeps Bitcoin usage legally compliant while enabling real-world adoption.

The organizing methodology is equally interesting. Bitcoin Indonesia published their meetup template on GitHub as a free guide, identifying six pillars of successful Bitcoin communities: be a genuine Bitcoiner with solid knowledge; commit to long-term community leadership; secure free venues with no barriers; provide free branding materials with reimbursement; keep everything completely free for attendees; focus on Bitcoin and financial literacy rather than trading speculation. They explicitly contrast their approach with competing crypto academies that charge $1,000-$2,000 for education, arguing that charging creates unrealistic return expectations that lead to disappointment and scam susceptibility.

The scale they've achieved is impressive — 3,600 Telegram members, 27,000 Instagram followers, 10,000 TikTok followers, and 500 graduates from the My First Bitcoin certification program by 2025. Their Fedimint federation, which provides Bitcoin ecash payment infrastructure, has an estimated 10,000 to 20,000 members. What makes this remarkable isn't just the numbers but the survival through government hostility and the methodical approach to sustainable growth.

These stories share something important: they represent people building real solutions to real problems through patient technical work rather than grand proclamations. The Alzheimer's researchers didn't just develop a test — they proved its clinical utility in real-world conditions. The Coreboot porter didn't just adapt existing code — they documented every failure and insight for future developers. Bitcoin Indonesia didn't just circumvent restrictions — they built sustainable organizing structures that can replicate across different regulatory environments.

In Bitcoin's technical development sphere, this kind of methodical progress continues. On Delving Bitcoin, Anthony Towns started a discussion about using AI tooling for code review, examining how machine learning could augment human reviewers in identifying potential issues in Bitcoin Core contributions. This is exactly the kind of careful integration of new tools that Bitcoin development requires — not revolutionary promises, but practical evaluation of how AI might improve security and efficiency in code review processes. Similarly, quantum resistance discussions are moving from theoretical to implementation-focused, with multiple threads examining concrete upgrade paths should quantum computing threats become imminent.

The common thread across all of these developments is patient, methodical work that prioritizes real utility over spectacular announcements. Whether it's medical diagnosis, open firmware, monetary adoption, or protocol development, the most significant progress comes from people willing to document their failures, share their insights, and build systems that others can improve upon. That kind of work doesn't generate dramatic headlines, but it moves the world forward in ways that matter.

[1] Blood test boosts Alzheimer's diagnosis accuracy to 94.5% — https://medicalxpress.com/news/2026-02-blood-boosts-alzheimer-diagnosis-accuracy.html

[2] I Ported Coreboot to the ThinkPad X270 — https://dork.dev/posts/2026-02-20-ported-coreboot/

[3] Bitcoin Indonesia's Real-Life Comeback — https://bitcoinmagazine.com/culture/from-40-meetups-a-month-to-nationwide-freedom-bitcoin-indonesias-real-life-comeback

[4] Using AI tooling for code review — https://delvingbitcoin.org/t/using-ai-tooling-for-code-review/2277

[5] Bitcoin Core PR #32220: cmake improvements — https://github.com/bitcoin/bitcoin/pull/32220

[6] Terence Tao at 8 years old (1984) [pdf] — https://gwern.net/doc/iq/high/smpy/1984-clements.pdf

[7] Bitcoin network status (mempool.space) — https://mempool.space/