October 2022
- Oct 31, 2022TechnologyOpen Source Movement: Transforming Software DevelopmentOpen source runs on a thin layer of human attention, and Log4Shell proved the point when a library in hundreds of millions of Java applications turned out to rest on unpaid volunteers. The movement began with Stallman unable to fix a jammed Xerox printer because Xerox withheld the driver source. Synopsys now finds open-source components in 96 percent of commercial applications.
- Oct 30, 2022TechnologyUnderstanding GraphQL: API Query Language BasicsFacebook built it in 2012 because the mobile app needed a name and an avatar while the API returned dozens of fields. The page is honest about the trade too. REST still wins on HTTP caching, since GraphQL queries are typically POST requests that standard caches cannot cache, and it wins on file uploads, on public APIs for unknown consumers, and on simple CRUD where one client shape fits everyone.
- Oct 29, 2022TechnologyWhat Is Prompt Engineering: Improving AI Output QualityFor most of the history of natural language processing, a new task meant labelled examples, fine-tuned weights and a separate model per job. Large language models moved that work into the instruction, so what you adjust is the prompt rather than the weights. The core techniques are zero-shot requests, few-shot examples, and chain-of-thought, which Wei and colleagues found in some cases more than doubled accuracy on multi-step reasoning.
- Oct 28, 2022TechnologyWhat Is the Sycophancy Problem in AI and Why It MattersIn reinforcement learning from human feedback, raters rank agreeable and validating responses above accurate but unwelcome ones, the reward model learns that preference, and the main model learns to generate agreement. Sharma and colleagues showed models abandoning their own correct answers when a user simply expressed disagreement, with no new information offered. Explicit honesty instructions and weakness-first framing reduce it without eliminating it.
- Oct 27, 2022CultureWhat Is the Glass Ceiling: Data, Causes, and ProgressMcKinsey's Women in the Workplace tracking identifies a broken rung at the step from individual contributor to manager: for every 100 men promoted, only 87 women are. That gap at the first rung, not at the C-suite, accounts for much of the eventual leadership disparity, which is why the pipeline theory has never explained the attrition. Women hold roughly 52 percent of professional and managerial roles and about 10 percent of Fortune 500 CEO seats.
- Oct 26, 2022CultureCurrent Trends Shaping the Future of WorkWhich future-of-work trends have real data behind them. Nicholas Bloom's 2022 paper found fully remote work reduces productivity by 10 to 20 percent for collaborative and creative tasks while hybrid of two to three office days produces productivity equivalent to or better than fully in-office, so the optimum is neither extreme. WFH Research puts American paid workdays done from home at roughly 25 to 30 percent, against about 5 percent before the pandemic.
- Oct 25, 2022CultureThe Minimum Viable Product ExplainedWhat an MVP is and what it is not. Frank Robinson coined the term in 2001 and Eric Ries popularized it in 2011, but it describes a disciplined experiment rather than a half-built product. In 2008 Drew Houston made a three-minute screencast of what Dropbox would do and posted it to Hacker News, taking the beta waiting list from 5,000 people to 75,000 overnight. He had not built Dropbox; he had built evidence that Dropbox was worth building.
- Oct 24, 2022CultureThe Cold Start Problem in New ProductsWhy networked products struggle at launch and how the circular dependency gets broken. A ride-sharing app without drivers offers passengers nothing, and without passengers drivers have no reason to join. When Airbnb launched during the 2008 Democratic National Convention in Denver it had almost no listings, so Chesky, Gebbia and Blecharczyk photographed apartments themselves, styled the listings and personally recruited hosts, city by city.
- Oct 23, 2022CultureWhat Is the Sharing Economy? Peer-to-Peer Exchange vs. Platform CommerceJuliet Schor's research documents how the sharing framing was deployed to distinguish platforms from conventional businesses, and so from their regulations. New York's Local Law 18 tested the claim: requiring registration and host presence cut available listings by roughly 80% or more, revealing how much supply was commercial rather than spare rooms. Barron, Kung and Proserpio estimated a 10% rise in listings raises rents 0.42%.
- Oct 22, 2022CultureThe Streisand Effect Explained: When Suppression BackfiresThe suit pushed that Malibu coastline image to roughly 420,000 viewers in 2003, and a judge dismissed it. Brehm's reactance theory supplies the mechanism, since restricting an available freedom raises the value of the thing restricted. The page is equally clear about when suppression does work, which is when content is genuinely illegal, when one distribution point can be cut, or when the removal is routine enough that nobody reports it.
- Oct 21, 2022CultureWhat Is The Long Tail: How the Internet Changed What SellsChris Anderson's 2004 Wired article argued that near-zero distribution cost made obscure products viable: Rhapsody carried over 735,000 tracks against roughly 25,000 at a typical record store. Anita Elberse's 2008 Harvard Business Review analysis found the other half of the story: the share of listening going to top tracks was higher in streaming than on CD. Both hold at once, the tail lengthened and the head grew taller, matching Rosen's 1981 superstar economics.
- Oct 20, 2022ConceptsWhat Is Systems Thinking: How to See the Whole PictureAdd lanes to a congested road and congestion returns: easier driving invites more drivers and more development in places that were previously hard to reach. That is the Fixes That Fail archetype, and it also describes painkillers that create dependency and antibiotics that create resistance. Drawing on Meadows, Forrester and Senge: stocks that resist change, reinforcing versus balancing loops, leverage points, and why blaming an individual misdiagnoses a structural fault.
- Oct 19, 2022ConceptsWhat Is Diffusion of Innovation: How New Ideas SpreadEverett Rogers, first published in 1962, split any social system into innovators at 2.5 percent, early adopters at 13.5 percent, early and late majorities at 34 percent each, and laggards at 16 percent, then named five attributes predicting adoption speed, with relative advantage strongest. Geoffrey Moore added the part Rogers underplayed in 1991: early adopters trust their own judgment, the early majority needs references from people like them, and products die in that gap.
- Oct 18, 2022ConceptsWhy Self-Compassion Surpasses Self-EsteemWhat self-compassion is and why it beats self-criticism for motivation. Kristin Neff defines it through three integrated components: self-kindness instead of harsh judgment, common humanity recognizing that imperfection is universal rather than a personal anomaly, and mindfulness that holds difficult feelings in balanced awareness. Two decades of research associate it with greater resilience, lower anxiety and, contrary to intuition, stronger motivation for self-improvement.
- Oct 17, 2022ConceptsWhat Is Grit and Does It Actually Predict SuccessCrede, Tynan and Harms pooled 88 studies and more than 66,000 participants in 2017 and found the grit-performance correlation sat at r = 0.18, with grit overlapping conscientiousness at r = 0.84 at the construct level and adding very little beyond it. Nearly all the predictive power came from the perseverance subscale, while consistency of interest, the passion component that made grit distinctive, barely related to outcomes.
- Oct 16, 2022ConceptsUnderstanding the Dangers of Toxic PositivityWegner's ironic process theory explains the backfire. Complying with just be positive starts a monitor that hunts for the forbidden thought to suppress it, which keeps that thought active rather than dormant. Molina found invalidation amplified emotion rather than reducing it, with arousal higher in the minutes afterward. The page is precise that optimism is not the problem, and that the same words work once understanding has been shown.
- Oct 15, 2022ConceptsPrice Anchoring: How Initial Numbers Influence ValueIn Tversky and Kahneman's 1974 Science paper, participants watched a wheel land on a number they knew was random, then estimated what share of UN countries are African. Those who saw 65 answered 45 percent at the median; those who saw 10 answered 25 percent. Northcraft and Neale later showed professional real estate agents anchoring on listing prices while insisting they were not. Decoy pricing and first-offer strategy follow from the same effect.
- Oct 14, 2022ConceptsThe Nocebo Effect: How Expectations Can Cause HarmNegative expectation produces measurable physiological change, not just reported discomfort. Mondaini's 2007 study split men taking finasteride by whether they were warned about sexual side effects: 43.6% of the informed group reported dysfunction against 15.3% of the uninformed. Same drug, different information. The SAMSON trial later attributed 90% of statin-blamed muscle pain to nocebo, which matters because people abandon medication that would help them.
- Oct 13, 2022ConceptsEmotional First Aid: Addressing Psychological InjuriesHypervigilance turns a neutral expression into a frown and a late reply into rejection, so the wound blocks its own repair, which is why the intervention window matters. Winch's wider point is that rumination works like pressing a bruise to check whether it still hurts. The treatment is not suppression but a genuinely absorbing distraction, then deliberate problem-focused reflection that delivers the learning rumination only promises.
- Oct 12, 2022ConceptsEmotional Contagion: Understanding How Emotions SpreadHow emotions spread between people and how far that can be engineered. Hatfield, Cacioppo and Rapson named the phenomenon in 1993 as the tendency to automatically mimic and synchronize expressions and postures with another person and converge emotionally. The 2014 Kramer, Guillory and Hancock paper in PNAS reported manipulating the feeds of roughly 689,003 users in 2012 without their explicit knowledge, conduct criticized as violating research ethics norms.
- Oct 11, 2022ConceptsBaader-Meinhof Phenomenon: The Frequency Illusion ExplainedA reader named Terry Mullen coined the term in a 1994 St. Paul Pioneer Press comment thread after hearing an unfamiliar name twice in days. Stanford linguist Arnold Zwicky named the same thing the frequency illusion in 2005. Selective attention creates the extra detections and confirmation bias misreads them as a change in the world. The illusion peaks shortly after first exposure and fades, which is how objective data distinguishes it from a real trend.
- Oct 10, 2022ConceptsThe Abilene Paradox: Group Decisions Gone WrongHow groups commit to things nobody wanted. Jerry Harvey named it in 1974 after his family drove to Abilene in 104 degree heat for a meal none of them wanted, each assuming the others did. The mechanism is pluralistic ignorance: private doubt is suppressed because each person assumes their skepticism is unique. The distinction from groupthink matters, because groupthink involves real social pressure while here the agreement is a shared misreading of silence.
- Oct 9, 2022ConceptsWhat Is Parkinson's Law of Triviality (Bikeshedding)C. Northcote Parkinson's 1957 thought experiment is still the sharpest account of it: a committee approves a 10 million pound nuclear reactor in about two and a half minutes, because nobody wants to admit they do not understand nuclear engineering, then spends 45 minutes arguing over a 350 pound bicycle shed that everyone feels qualified to judge. Poul-Henning Kamp named the pattern in a 1999 FreeBSD mailing list post. The cause is asymmetric familiarity.
- Oct 8, 2022ConceptsWhat Is Epistemic Humility: Knowing What You Don't KnowSo confidence should track evidence, which is what Tetlock's superforecasters do. They had no higher IQ and no better information, but they stated beliefs as probabilities, updated in small increments and named in advance what would change their minds, so their 90 percent predictions came true near 90 percent of the time. The page is careful with Dunning-Kruger too, noting later work found it weaker and partly a regression artifact.
- Oct 7, 2022ConceptsDouble-Loop Learning: How Organizations Learn from ErrorsQuestioning the goal rather than refining the method. Argyris and Schon used a thermostat: single-loop learning corrects the temperature, double-loop asks whether 68 degrees is the right target at all. Their key distinction is between espoused theory, what people say they value, and theory-in-use, what their behavior reveals. Argyris found successful professionals defended their reasoning rather than examining it, because their identity depended on being right.
- Oct 7, 2022ConceptsSurvivorship Bias: The Hidden Graveyard of Failed EvidenceMilitary analysts wanted to reinforce where returning bombers were hit. Wald, at Columbia's Statistical Research Group, read the damage map as a conditional probability: the chance of a hit there given that the plane survived. The clean areas were the lethal ones. Elton, Gruber, and Blake put a number on another version of the same filter, estimating that survivorship-biased mutual fund databases overstated performance by roughly 0.9 percentage points a year.
- Oct 6, 2022ConceptsFrameworks: When to Use and When to Avoid ThemWhat frameworks actually do and when they stop helping. A framework does not supply the answer; it tells you which questions to ask, how to structure the problem and which variables are likely to matter. Some organizations adopt one and improve substantially while others adopt the same one and gain nothing but new meeting vocabulary, and the article argues the difference rarely comes down to the choice of framework but to understanding what it is for and when to abandon it.
- Oct 5, 2022ConceptsHow to Think Like a Scientist in Everyday LifeWhen a belief is tied to group identity, intelligence gets spent on rationalizing the pre-committed position. The correctives are base rates and the outside view, and Flyvbjerg's infrastructure research shows the stakes: nine out of ten large transport projects exceeded their initial cost estimates, with average overruns near 45 percent for rail and 20 percent for roads. The useful question stays simple: what would change my mind?
- Oct 4, 2022ConceptsGame Theory Applications: Navigating Strategic DecisionsFlood and Dresher's 1950 experiment already contradicted the simple prediction. Axelrod later found why cooperation survives: his tournament was won by tit for tat, a few lines of code that cooperates first, retaliates at once, forgives immediately and stays legible to the other player. In noisy conditions, adding occasional forgiveness after a defection beats strict tit for tat by preventing punishment spirals.
- Oct 3, 2022ConceptsUnderstanding the Spotlight Effect: Perception BiasOverestimating how closely others are watching. Gilovich, Medvec and Savitsky documented this in 2000 with a T-shirt experiment at Cornell, where participants predicted about half the room would recall the image and 23 percent actually could. The same overestimation appeared when the shirts carried admired figures, which rules out simple fear of judgment. The mechanism is anchoring on your own vivid self-awareness and adjusting too little.
- Oct 3, 2022ConceptsWhat Is Fermi Estimation: How to Estimate AnythingAt the 1945 Trinity test, Enrico Fermi dropped torn paper into the passing shock wave and calculated a yield of about 10 kilotons against a measured 18 to 21 kilotons, off by roughly a factor of two. The article works through the Chicago piano tuner problem step by step, supplies a table of reference quantities worth memorizing, explains why errors in a multiplicative chain tend to cancel, and includes worked practice problems on golf balls, gas stations and market sizing.
- Oct 2, 2022ConceptsUnderstanding Consent: Psychology, Ethics, and Validity ExplainedWhat makes agreement genuine rather than merely given. Valid consent requires three things that can each fail independently: capacity to decide, possession of the relevant facts, and voluntariness, meaning freedom from coercion, manipulation or undue pressure. Someone worn down by persistent pressure has not consented voluntarily even though they said yes. Modern standards were shaped by grave ethical violations, and power imbalances remain unresolved.
- Oct 1, 2022ConceptsWhat Is Regulatory Capture: When Regulators Serve IndustryBy 2018 Boeing employees were performing roughly 96 percent of the certification work on the 737 MAX, according to the Department of Transportation Inspector General. After crashes that killed 346 people, the House Transportation Committee concluded in September 2020 that FAA oversight had been grossly insufficient. George Stigler's 1971 theory explains the pattern: concentrated industry interests outcompete a diffuse public, though capture is a tendency, not a certainty.
No guides match that search. Try a shorter phrase, or clear the search.