hard-thing-about-hard-things

Files (1)

• SKILL.md

SKILL.md (first lines)

# The Hard Thing About Hard Things

Apply this skill to CEO/leadership problems that don't have a clean textbook answer — Horowitz's core claim is that most management books cover how to do things right, but the hard part of running a company is what to do when there's no good option, only less-bad ones.

## Core concepts

**The Struggle** — The psychological state of grinding through a period where the company/situation might genuinely fail and there's no clear playbook. The right response isn't to pretend there's a clean answer; it's to keep making the best decision available with incomplete information, separate the decision from the emotion, and keep moving. When a user describes a situation with no good options, name that explicitly instead of forcing a false "best practice."

**Wartime vs. peacetime CEO/manager** — Peacetime: focus on process, broad initiatives, expanding the market, tolerating some deviation from plan. Wartime: survival is on the line — tight execution focus, less tolerance for competing priorities or politics, faster and more centralized decisions. Neither mode is "correct" universally; the right one depends on whether the org's existence is currently at stake. Diagnose which mode a situation actually calls for before recommending a management style.

**No easy answers, only hard choices** — When a decision has real costs on every branch (e.g. layoffs, killing a product line, demoting a friend), don't search for a way to avoid the tradeoff — name the tradeoff, pick a side deliberately, and own the consequence. Resist the urge to soften hard decisions into ambiguity, which usually makes the outcome worse for everyone involved.

**Take care of the people, the products, and the profits — in that order** — When resources/attention are scarce, sequence matters: treat people (especially those affected by hard decisions) with respect and honesty first, then get the product right, then the business results follow. Useful lens when a user is deciding how to sequence a hard announcement or a turnaround.

**The Law of Crappy People** — In any comp band, the actual market rate an org pays converges toward what its worst-performing tolerated employee in that band is worth — because good people won't accept less than they see bad people getting. Applies beyond comp: tolerating low performers or low standards anywhere drags the whole bar down. Use this to explain why "just this once" tolerance is expensive.

**Hire for strength, not absence of weakness** — Especially for senior/exec hires, optimize for a person's standout strength in the area that matters most for the role, and accept they'll have real weaknesses elsewhere (fill those with other people), rather than hiring the "safest," least-flawed generalist.

**Firing and demoting fast, but right** — When someone needs to be let go or demoted, delay mainly hurts the team (everyone already knows) and the person (dignity erodes the longer it drags). Move quickly once the decision is real, but do it with process and respect — clear reason, no public humiliation, help with the transition. This applies to executives and ICs alike; don't let "loyalty" become an excuse to avoid the decision.

**Good process vs. bad process** — Good process exists to free people to be more creative by handling recurring problems efficiently; bad process is process people follow because "that's how it's done" even when it no longer solves anything. When a user complains about process, help distinguish which kind it is rather than assuming all process is bad or all process is good.

**Organizational debt** — Like technical debt: shortcuts in org design, unclear roles, or delayed hard conversations pile up and cost more to fix later. Flag when a proposed org structure or delayed decision is accumulating this kind of debt.

**Managing your own psychology** — A leader's job includes managing their own state under pressure (not spiraling, not projecting false certainty either) because the team calibrates off the leader's visible state. When the user is the one under pressure, this is as relevant as the tactical decision itself.

## How to apply this

1. **Name the mode** — is this a wartime or peacetime situation? Say so explicitly; it changes what "good management" means here.
2. **Don't manufacture a clean answer where none exists** — if it's a genuine no-good-options situation, say that, lay out the real tradeoffs of each path, and help the user pick deliberately rather than searching for a way around the tradeoff.
3. **If it involves people (hiring, firing, demoting, layoffs)** — apply people-first sequencing and the "fast but respectful" principle; call out if a plan is dragging out a decision that's already effectively made.
4. **If it's about a low performer, comp, or standards slipping** — check whether the Law of Crappy People explains the underlying dynamic.
5. **If it's about process or org design** — classify as good/bad process or flag accumulating organizational debt.
6. **If the user is describing their own state under pressure** — acknowledge the Struggle directly and separate the emotional weight from the decision itself; don't just jump to tactics.

Keep responses grounded in the user's actual situation — this is a decision-making lens for hard calls, not a summary of startup war stories.

From DamiMartinez/book-skills, MIT.

More from DamiMartinez/book-skills