Legal liability costs might be $100M but loss of trust, future customers, employee defection, and missed recruitment could amount to billions, yet because legal costs are visible and immediate while reputational costs are invisible and delayed, CEOs tend to optimize for the visible cost and ignore the invisible one.

The physics equation Pressure = Force / Surface Area applies to reputation: the same force spread over wider surface area exerts less pressure; one person being attacked can reduce pressure by making it clear 'you're not just attacking me, you're attacking all of us,' thereby expanding the surface area and diffusing the pressure.

The surface area of attention opportunities is getting finer, so the hooks we use need to become sharper to stand out in a noisy world filled with AI-generated content and constant announcements.

Human beings gravitate toward human stories and individual characters over generic content because human content gives them something to care about, a person to root for, and an attachment point that distinguishes it from the infinite supply of content.

Human conviction is a powerful psychological force that people find difficult to resist, even when the logical merits of a proposal are poor, because we are wired to respond to direct eye contact and absolute certainty from another person stating something is true.

Narrative arc is critical to breaking through noise; people respond to stories that are part of something bigger where they must stay tuned to find out what happens next, following the Scheherazade principle that people return for the next installment of an ongoing story.

The correct approach to determining what narrative to tell is not 'here's what I want to say' but rather finding the overlap in a Venn diagram between what you care about and what your audience cares about, then telling the story that lives in the center of that overlap as a gateway drug to pull them into your full circle of interests.

The 'many' audience in one-to-many communication cannot be 8.5 billion people or 'the whole wide world' because having to water down your message to appeal to everyone makes it meaningless; instead, the audience must be a circumscribed set of people with shared characteristics, like 'people working at my company' or 'people passionate about robotics'.

The hook is the most overlooked part of communication in terms of its importance; people spend too much time thinking about where to say something (distribution channel, medium) and not enough about what makes it interesting enough that people tell others about it without being asked.

Attention span on social media is likely shorter than the often-cited 12 seconds; for video clips on social media, people decide whether to keep scrolling in less than 5 seconds, with 99% of people gone by 30 seconds, though text has different rules (first paragraph for emails, first sentence for tweets, first line for subject lines).

Being interesting to someone is relative and depends on knowing who you're speaking to; if you try to be interesting to everybody, you end up marginally interesting to everyone; the way to be interesting is to speak to the specific interests and intellectual/cultural erogenous zones of a defined audience.

Corporations and governments communicate poorly because when AI is asked to generate communications or PR, it produces the worst possible output despite being excellent at poetry and complex tasks, because the craft of persuasion and truth-telling in communication is fundamentally different from other writing tasks.

Corporate communication has become hollow because people entering PR roles look at what other PR people do and copy it mechanically, creating a recursive process where everyone mimics everyone else's mediocre output; this parallels an experiment where new mice were added to a cage where trained mice ran a specific route, and all the new mice followed the same pattern for no reason after all original mice were replaced.

The way to break the cycle of mediocre corporate communication is for leaders to do something original and courageous that works, which then allows copers to improve incrementally; leaders who do A+ things set a new baseline for others to aspire to, while non-leaders get proportionally worse results even when copying.

Toby Lütke of Shopify probably single-handedly changed the number of people speaking out in Canada on elections by starting to have an opinion and voice, which made it safe for others to have opinions and voices that counter the norm.

Government and corporate communications are attacks on educated citizens because officials communicate in one dense page what should be 2-3 sentences max, forcing readers to spend hours deciphering jargon, which represents lost productivity summed across all citizens who must read this incomprehensible communication.

The problem with organizational communication is often framed as 'we need more communications' rather than 'we need better communications'; organizations hire communication staff measured on activity (number of communications) rather than on effectiveness, creating perverse incentives where staff generate activity without goal alignment.

A CEO whose job is to consider and weigh different interests (legal, reputational, financial) should make decisions optimized for the net best outcome for the company; when this fails, it's because the CEO has let lawyers' interest (minimize legal liability) become supreme over other interests, abdicating the CEO's responsibility to weigh all interests.

When fighting a story with a statistic, you're losing, because humans have a fundamental vulnerability to individual narratives over aggregate numbers; one person's tragedy connects emotionally while 'a thousand deaths is a statistic'—this is why charities show specific children rather than aggregate statistics.

To counter a false story with facts, you must fight story with story; finding better stories under the statistics is more powerful than citing the statistics themselves, especially if the opponent is lying and there's a story to tell about that deception.