Most business books age badly. The tactics expire, the case studies feel dated, and the author’s certainty about “the future” reads like fiction. But a handful of books capture something deeper — mental models about leverage, systems, and human behaviour that AI has made more important, not less.
These ten are the ones I return to. Each includes what still applies, what AI has changed, and when to reach for it.
The E-Myth Revisited
Core idea: Work on your business, not in it. Build systems that run without you.
Still holds because: Solo founders with AI can now build those systems faster than Gerber imagined — but most still don't. They automate tasks instead of designing roles.
What changed: The "franchise prototype" is now a stack of AI workflows + SOPs. You don't need employees to systematise — you need clear processes that tools can execute.
Antifragile
Core idea: Systems that gain from disorder outperform those that merely survive it.
Still holds because: AI volatility is the new black swan factory. Founders who build optionality — multiple revenue streams, modular offers, reversible decisions — win.
What changed: Speed of disruption has increased 10x. Antifragility now means running small experiments weekly, not quarterly. AI makes experimentation cheap enough to do constantly.
So Good They Can't Ignore You
Core idea: Career capital — rare and valuable skills — beats "follow your passion."
Still holds because: When AI commoditises output, the scarce resource is judgment, taste, and domain expertise. Career capital is more valuable than ever.
What changed: "Rare skills" now include knowing which AI tools to combine and how to brief them. Technical fluency with AI is becoming as foundational as literacy.
The Lean Startup
Core idea: Build-measure-learn. Validate before you scale.
Still holds because: The loop is faster now, but the discipline isn't. Founders still build too much before testing — they just build it with AI in an afternoon instead of a month.
What changed: MVP timelines collapsed from months to days. The bottleneck moved from building to knowing what to build. Customer discovery matters more, not less.
Thinking, Fast and Slow
Core idea: Two systems govern decisions — fast/intuitive and slow/deliberate. Know which one you're using.
Still holds because: AI is System 1 at scale — fast, confident, often wrong in subtle ways. Your job is to be System 2: the deliberate check on automated output.
What changed: The risk isn't that AI thinks for you — it's that AI feels like thinking, so you stop checking. Build verification into every workflow.
Zero to One
Core idea: Competition is for losers. Build monopolies through unique insight, not incremental improvement.
Still holds because: AI makes copying easy — which makes differentiation essential. The founders who win have a point of view AI can't replicate.
What changed: "Secrets" are harder to keep (AI exposes patterns fast) but easier to act on (AI executes on insight fast). Speed of execution on a unique insight is the new moat.
The Hard Thing About Hard Things
Core idea: There's no formula for the hardest decisions — only honesty about what you're facing.
Still holds because: AI can't make the calls that matter: firing, pivoting, saying no to revenue, choosing which bet to make. Leadership is still human.
What changed: Solo founders face these decisions alone, without a board or co-founder. AI is a thinking partner, not a substitute for the courage to decide.
Atomic Habits
Core idea: Systems beat goals. Small consistent actions compound.
Still holds because: The founders who integrate AI successfully don't do it in bursts — they build daily habits around briefing, reviewing, and iterating on AI output.
What changed: Habit design now includes tool workflows. "After I finish my morning coffee, I run email triage" is as important as "after the gym, I write for 30 minutes."
Company of One
Core idea: Stay small on purpose. Growth isn't always the goal — resilience and autonomy are.
Still holds because: This book predicted the solo-founder economy before AI made it obvious. One person + the right stack can outperform a team.
What changed: Jarvis wrote this pre-AI. The "company of one" is now a company of one + an AI department. The philosophy is the same; the ceiling moved.
Range
Core idea: Generalists triumph in complex, unpredictable environments. Deep early specialisation is overrated.
Still holds because: AI rewards people who can connect dots across domains — who see how a framework from biology applies to business, or how a design principle improves a sales process.
What changed: AI gives generalists superpowers. You don't need to be the best writer, researcher, or analyst — you need to be the best at knowing what to ask for and judging what comes back.
How to Use This List
Don't read all ten at once. Pick the one that matches your current bottleneck:
- Feeling overwhelmed by operations? → Start with The E-Myth Revisited
- Unsure what to focus on? → Start with So Good They Can't Ignore You
- Building too much, testing too little? → Start with The Lean Startup
- Over-relying on AI without checking? → Start with Thinking, Fast and Slow
- Questioning whether to stay small? → Start with Company of One
The principles in these books outlasted every tech cycle before AI. They'll outlast this one too — if you apply them.