Reading up on organizational-health

Managers vet AI-produced work less carefully, and AI models favor AI-written content and exhibit rational biases, undermining productivity gains from workplace AI.

Rather than asking for permission, offer colleagues a deadline to veto your planned action, which reduces their cognitive load and keeps projects moving.

Eric Ries argues that companies succumb to 'financial gravity'—structural incentives that corrupt their mission over time—and that certain governance structures can resist this decay.

Meta has reassigned 30-50% of engineers from core teams to data labeling for AI, sparking morale collapse and skepticism about leadership's strategy.

Intent debt—the missing externalized rationale for why a system is built—is the most expensive kind of debt because AI agents cannot generate it, and its cost compounds as teams rely on more agents.

As AI speeds up code creation, trust lags behind; Kent Beck argues that software teams must deliberately manufacture trust through practices like testing, pairing, and continuous deployment.

John Cleese argues that creativity requires a calm, playful, and open mental state, which management cultures often suppress through urgency and interruption.

Japanese companies diversify into many industries because lifetime employment and indifference to shareholder pressure force them to create jobs for employees.

Harvard faculty voted to approve a cap on the number of A's per course in an effort to curb grade inflation, despite sharp student backlash.

The article argues that AI amplifies the gap between perceived and actual productivity by enabling workers to produce voluminous artifacts without the competence to evaluate them, leading to organizational dysfunction.

Meta's AI push, including token tracking, layoffs, and pressure to adopt AI tools, is making its employees miserable, the article reports.

A Microsoft engineer in Boca Raton escalated a TAB-key dispute to his manager, who refused to overrule him, leading to an IBM VP opposing the choice and a deadpan reply ending the debate.

Meta is installing software to capture employee mouse movements, keystrokes, and screen content for training AI agents to perform work tasks autonomously.

Probabilistic engineering shifts code from known correctness to believed correctness as agent output outpaces human validation, requiring new organizational structures and review discipline.

Every layer of review multiplies wall-clock time by 10; AI coding accelerates only the first step, so reducing reviews and building quality into systems is the only path to speed.

A Cornell study finds that employees who are impressed by corporate buzzwords perform worse on tests of analytic intelligence.

Jeff Bezos, dissatisfied with losses at The Washington Post, is pushing the newsroom to double productivity with half its budget.

Individual AI tools boost personal productivity but fail to improve firm-level outcomes until organizations redesign their processes and coordination structures around AI.

Former employees allege that René Redzepi, chef at Noma, subjected staff to years of physical and psychological abuse.

Engineering cultures reward complex over simple solutions in interviews, design reviews, and promotions, creating incentives that penalize engineers who avoid unnecessary complexity.

Organizational failures stem from locally rational decisions within poorly designed incentive systems, not from bad people.

The Horological Society of New York's welcoming atmosphere is cited as a key factor in its 160-year survival and ongoing relevance.

Anthropic operates as a chaos-driven hive mind with more work than people, a model Steve Yegge argues will define successful companies as AI accelerates development.

The author proposes defining a company's structure, roles, and processes as declarative code, enabling versioning, testing, and automation.

Apparent contradictions like OpenAI selling ads don't disprove imminent superintelligence because AI labs must fundraise and hedge against normal-tech scenarios to win the race.

Early-stage startup founders should avoid formal engineering management and instead focus on hiring motivated people, staying flat, and using boring, lightweight practices until the team reaches 20+ engineers.

AI-assisted coding doesn't boost most dev teams' productivity; only teams that already eliminated process bottlenecks see gains.

AI coding agents reveal that development environments are brittle, non-standard, and reliant on human intuition, forcing teams to standardize and harden their systems to leverage agents effectively.

In a 2025 Ask HN thread, developers share that low-effort tech jobs exist in government, defense, and internal tools, though many warn such roles are soul-crushing and suggest addressing deeper unhappiness instead.

A boss's aversion to sick day requests highlights workplace tensions and the impossible demands placed on middle managers.

Pursuing what you are good at creates self-reinforcing success and engagement, rather than following passion, and balancing work with life is key to longevity.

AWS is fine but faces real challenges: the AI gap is closing, talent attrition threatens operational excellence, and 2026 will test execution.

Effective software design in large existing systems requires deep codebase familiarity, making generic design advice largely useless for practical problems.

Large software products are so complex and rapidly changing that even engineers often cannot answer basic questions about them without investigative research.

Creating mini-frameworks on top of existing shared tech stacks causes more harm than good, leading to complexity, maintenance issues, and cognitive load.

OpenAI researcher John Schulman discusses reinforcement learning scaling, LLM usefulness milestones, and lessons for building AI research teams.

Workers anonymously confess to bosses that burnout, performative care, and profit-at-all-costs leadership have destroyed trust, leaving them disengaged and self-protective.

Hacker News commenters debate how technical debt and system failures often stem from human factors like communication, politics, and organizational silos.

Meta's AI elite clash with longtime Zuckerberg lieutenants, creating an us-versus-them tension inside the company.

Estimates become harmful when treated as deadlines rather than tentative guides, creating tension between developers and product owners.

HN commenters largely agree with slide-deck advice on declining useless meetings but cite cultural resistance and lack of enforcement as barriers.

Frequent team rotations and overloaded senior engineers cause big tech companies to write surprisingly sloppy code.

Large software companies sacrifice efficiency for legibility to control scale and enable enterprise deals, despite slowing engineering velocity.

Interruption rate, recovery time, and minimum focus block size mathematically determine whether a workday yields deep work, with simulations showing small parameter changes drastically shift productivity.

As organizations grow beyond 100 people, feedback becomes overwhelming noise because personal relationships don't scale, requiring structured systems like proxy relationships and working groups to replace direct trust.

Good execution is defined by quiet, low-noise operations where clear direction, stable context, clear ownership, and trust enable teams to deliver without friction or management theater.