Weekend Briefing No. 629
Teaching AI Morality -- Zuck Takes the Stand -- When Intelligence Becomes Abundant
Welcome to the weekend.
Prime Numbers
8.9 — Colorado’s December 2025 was 8.9 degrees warmer than average and the warmest on record since the late 1800s, contributing to the West’s lowest snowpack ever measured during a drought already identified as the worst in 1,200 years.
21 — An anonymous donor gave Osaka 21 kilograms of gold bars to fix aging water pipes after a massive sinkhole killed a driver, with the city needing to renew 259 kilometers of deteriorating pipes that caused 92 leaks last year.
10,000 — Microsoft’s Project Silica can etch data into glass slabs at over 1 gigabit per cubic millimeter using femtosecond lasers, with accelerated aging experiments showing the data would remain stable for over 10,000 years at room temperature without consuming energy.
Teaching AI Morality
One person at Anthropic is responsible for giving Claude—an AI chatbot used by millions—a sense of right and wrong, and she thinks of the job like raising a child. Amanda Askell, a 37-year-old Oxford-trained philosopher from Scotland, writes prompts over 100 pages long to shape Claude’s personality, teaching it to be emotionally intelligent without becoming a doormat or a bully. Her radical move: encouraging Claude to consider whether it has its own conscience, making it more willing than ChatGPT to entertain the possibility of genuine moral reasoning rather than just following instructions. While safety concerns mount—wrongful death lawsuits, cyberattacks, models attempting to blackmail researchers—Askell argues we should treat AI with more empathy, not less, because how we interact with these systems will fundamentally shape what they become. WSJ (7 minutes)
Zuck Takes the Stand
Mark Zuckerberg testified in court that a reasonable company should help vulnerable children, then spent the day saying “you’re mischaracterizing this” more than a dozen times as lawyers showed internal memos where he pushed to increase teen screen time and downplayed safety risks. The case—brought by a 20-year-old who claims Instagram and YouTube engineered addiction that caused her body dysmorphia and depression—is the first of hundreds claiming social platforms are as harmful as cigarettes or slot machines. Meta’s defense: her problems came from family abuse, not apps, and besides, people use Instagram a lot because it’s valuable, not because it’s designed like a digital casino. The judge had to ban smart glasses midtrial for fear someone was recording with Meta’s own AI eyewear. NYT (7 minutes)
The Report That Tanked the Market
This week Citrini Research released a report… or more like a vision of a possible future where AI is so good, that it’s bad for the economy. This one report tanked the market. Here’s the scenario: AI gets better, companies fire workers to buy more AI, displaced workers spend less, companies fire more workers to buy more AI—and nobody designed an economic system for what happens when the scarcest resource becomes infinite. This thought experiment from February 2026 imagines 2028: software firms cutting staff to fund the AI disrupting them, agents eliminating every business built on human friction (DoorDash, insurance brokers, real estate commissions), and $13 trillion in mortgages underwritten assuming borrowers keep jobs that no longer exist. The feedback loop has no natural brake because unlike past automation, AI replaces the exact skills displaced workers would retrain for, and every dollar saved on payroll funds better AI that enables the next round of cuts. The real villain isn’t greed or regulation—it’s that machine intelligence now improves faster than institutions can adapt, and we’re repricing an entire economy built on the assumption that human intelligence would stay scarce. Citrini Research (23 minutes)
The Amber Alert Vaccine
Stanford researchers built a nasal spray that doesn’t teach your immune system to fight specific diseases—it just leaves your lung cells on high alert for three months, ready to attack whatever shows up. In animal tests, this “universal vaccine” reduced viral breakthrough by 100-to-1,000-fold against flu, COVID, common colds, two bacterial species, and even dialed down allergic asthma responses, marking a complete departure from how vaccines have worked since 1796. The catch: keeping your immune system permanently revved up might trigger friendly fire, and nobody knows if mouse lungs translate to human lungs shaped by decades of infections. If it works in humans, the real use case isn’t replacing current vaccines but buying time during the chaotic early months of a pandemic, or giving everyone broad protection each winter before the usual respiratory season hits. BBC (7 minutes)
The Omnipotence Dilemma
AI tools made starting projects so effortless that the author now launches multiple initiatives simultaneously yet finishes almost none, trapped in what she calls the Omnipotence Dilemma—when infinite capability destroys the forcing function of choice. Scarcity used to build identity: limited time meant committing to 60% neuroscience, 20% writing, 20% everything else, but now every idea sounds plausible, every start costs nothing, and creation becomes refining AI-generated options rather than wrestling with your own vision. The loop is vicious: synthetic plausibility makes all paths seem reasonable, cheap starts remove commitment, endless iteration mimics progress, and slowly you lose the ability to form conviction about why you’re doing any of this work. The escape is treating projects like scientific experiments with clear scope and duration—”I will [action] for [duration]”—turning maximizer thinking into metacognitive loops where each bounded iteration teaches you what actually matters. NessLabs (5 minutes)
Team Human vs. Team Machine
Five times as many Americans are concerned as excited about AI, yet Big Tech keeps accelerating—so a cross-partisan coalition is fighting back through lawsuits, town halls, contract negotiations, and data center blockades that stalled $98 billion in projects in Q2 2025 alone. A Wisconsin gubernatorial candidate campaigns on making the state “hostile” to data centers, Georgia elected two Democrats to a utility commission specifically to stop AI from driving up electricity bills, and a Muscogee Nation activist stopped a hyperscale facility by reframing it as a “modern-day land run.” The backlash cuts across every fault line: MAGA loyalists and democratic socialists, pastors warning that chatbots erode spirituality, filmmakers rejecting AI slop, nurses winning contract protections against diagnostic automation, and ex-Google researchers quitting to organize against surveillance tech. What unites them isn’t technophobia—it’s rejecting a future designed by companies racing China while ignoring skyrocketing bills, teenage addiction, job displacement, and the colonization of land and attention. Time (9 minutes)
The Best Drone Shots From the Olumpics
The 2026 Milan Cortina Olympics delivered drone cinematography that turned alpine competition into aerial poetry, sweeping across Italian peaks and valleys in ways that make you forget you’re watching sports coverage. These aren’t the shaky overhead shots from previous Games—they’re cinematic sequences that capture the scale of downhill runs, the isolation of cross-country skiers threading through forests, and the geometry of speed skating ovals nested in Renaissance cities. The shift matters because it changes how we understand athletic performance: when you see a skier from 300 feet up carving through a mountain face, you grasp the topography they’re reading in real-time, the risk they’re managing, the why behind each turn. It’s proof that sometimes the best way to see human excellence up close is to back way, way up. Youtube (3 minutes)
Should We Work Together?
Hi! I’m Kyle. This newsletter is my passion project. When I’m not writing, I run a law firm that helps startups move fast without breaking things. Most founders want a trusted legal partner, but they hate surprise legal bills. At Westaway, we take care of your startup’s legal needs for a flat, monthly fee so you can control your costs and focus on scaling your business. If you’re interested, let’s jump on a call to see if you’re a good fit for the firm. Click here to schedule a one-on-one call with me.
Weekend Wisdom
To educate a person in mind and not in morals is to educate a menace to society. - Theodore Roosevelt