Daniel

Electrical crimping tool

GPS is in 6 billion devices. Almost nobody knows how it actually works. Here's the full story, including the 5 problems engineers had to solve first. [1/7]

Most engineering projects fail because of questions nobody asked at the start. Here are the 8 that matter most, before a single line is drawn: 1. WHAT DOES SUCCESS ACTUALLY LOOK LIKE? Not the spec. The outcome. "A bridge that carries X vehicles per day" is a spec. "Economic connectivity between two communities" is the outcome. Design for the outcome. 2. WHO HAS THE AUTHORITY TO CHANGE THIS? Identify every stakeholder who can alter requirements mid-project. Get them aligned before you start. Surprises from power centres you didn't map will kill your schedule. 3. WHAT IS THE ONE THING THAT WOULD MAKE THIS FAIL? Name your single biggest risk. Now test it first, not last. 4. WHAT ARE WE ASSUMING THAT WE HAVEN'T VERIFIED? List your top 10 assumptions. Rank by consequence if wrong. Address the dangerous ones before committing to the design. 5. WHAT DOES THIS LOOK LIKE WHEN IT BREAKS? Define failure modes before you design. You'll make different choices. 6. HOW WILL THIS BE MAINTAINED? Design for the person maintaining it in 10 years, not the person building it now. They won't have your knowledge. They'll have a manual. 7. WHAT WILL WE DO IF THE BUDGET DROPS 30%? Forces you to identify which 70% of scope is essential. Decisions made under pressure produce worse outcomes than decisions made in advance. 8. IS THIS THE RIGHT PROBLEM? The hardest question. Solve the right problem badly and you've wasted nothing. Solve the wrong problem brilliantly and you've wasted everything. Ask these once. You'll save months.

Your phone contains 15 billion transistors. Each one switches between 0 and 1 billions of times per second. Here's how that actually works — from sand to signal. [1/7]

Remote work is quietly destroying engineering culture. Not productivity. Culture. The best engineering I've seen happened in rooms where: • A junior engineer overheard a senior one thinking out loud • Someone walked to a whiteboard mid-sentence • A bad idea got killed in 10 seconds because the right person was three desks away None of that happens on Slack. Mentorship doesn't scale asynchronously. Tacit knowledge, the stuff that's never written down, transfers in person or not at all. The engineers who thrived remote were already great. The engineers who needed to become great? Many of them didn't. I'm not saying remote work isn't real work. I'm saying that the craft of engineering requires proximity in ways we pretend it doesn't. The industry optimised for comfort. We'll spend 10 years figuring out what we lost.

90% of engineering disasters trace back to 6 failure modes: 1. Normalization of deviance Small violations of protocol that don't cause failure. They get normalized. Then one day the margin runs out. (See: Challenger, Columbia) 2. Organizational pressure overriding engineering judgment Engineers knew. Management decided anyway. (See: Challenger, 737 MAX) 3. Single point of failure One component. No backup. Catastrophic when it fails. (See: Tacoma Narrows, Chernobyl control rods) 4. Cumulative damage Nothing dramatic. Just slow degradation nobody tracked. (See: every bridge that "suddenly" collapsed) 5. Interface failure Each component works perfectly. They fail where they connect. (See: Mars Climate Orbiter – metric vs imperial) 6. Design outside tested parameters The system was tested at normal conditions. Reality wasn't normal. (See: Deepwater Horizon, Tenerife disaster) Know these. They're recurring.

ldk = I don’t know Imo = in my opinion Brb = be right back F = ma pV = nRT V = IR p1V1 = p2V2 E = mc²