The Most Human Technology Ever Made
a16z Anish a
Core thesis: AI is less an efficiency tool and more a paintbrush — a technology that expands what humans can express and make, not just what they can consume faster.
"Most people aren't trying to save time. They're trying to spend it."
Key ideas
- Making stretches you across past/present/future — you build toward something while drawing on everything you've loved. Consuming parks you in the "now."
- Slop isn't just machine-generated — it's what happens when everyone consumes and nobody creates. AI is unusually good at lowering the cost of making.
- Real examples: a master electrician in Kentucky built a load-calculation tool that sells for $12.99 and replaces a $500 service call. A plumber canceled a $40k consulting contract after an afternoon with OpenClaw.
- The "tax" of work (politics, meetings, permission-seeking) gets eaten by AI, leaving more of the actual play part where breakthroughs live.
- "The next big thing will start out looking like a toy." — Chris Dixon
- Individuality at scale: the thing that used to be a luxury (being a distinct person) becomes the point of the work itself.
Why it matters: Directly maps to the agent/tool/making space. It's not another "AI will 10x efficiency" piece — it's arguing the opposite direction is the interesting one.
How to Survive the Abyss
Andrew Yeung Lessons from Mark Pincus (Zynga)
Core thesis: The "Abyss" is the disorienting in-between period (post-exit, career pivot, fired, sold company, no clear "what's next"). It's not a failure to escape — it's a necessary generative phase.
The five-step framework
- Stop calling it failure. Don't rush to get out. Steve Jobs' wilderness years (NeXT + Pixar), Sara Blakely selling fax machines, Reid Hoffman between PayPal and LinkedIn — the list is long.
- Become a flâneur. Wander without a destination. Pincus obsessed over perfecting Bolognese sauce during his first Abyss — not to become a chef, but to train taste/pattern recognition. Sample new disciplines via advising, fractional work, fellowships.
- Stay in a state of readiness, not control. Sahil Bloom doesn't plan more than a year ahead. Improv actors who pre-plan their next line kill the scene. Grandmasters don't make premature moves.
- Shed your shell. Lobster analogy: you have to cast off the old identity and sit soft/exposed while the new one forms. Adam Grant: hold your identity loosely as a hypothesis, not a truth to defend.
- Reject the B+. The biggest danger is leaving too early and falling in love with a mediocre idea just because it ends the uncertainty. Wait for the A. "Negative capability" — Keats' term for remaining in uncertainty without reaching desperately for the nearest answer.
"The Abyss isn't the thing standing between you and the work. The Abyss is where the work happens."
Why it matters: Extremely resonant with where a lot of founders (and people in the AI/agent/tool space) are right now — post-hype cycles, rapid capability shifts making old playbooks obsolete. The "flâneur + shed your shell + reject B+" combo is useful operating advice.
How they rhyme
The two essays complement each other. The a16z piece argues why the making/expression side of AI matters more than pure efficiency. The Abyss piece is practical advice for the exact career/identity uncertainty that often precedes or accompanies big shifts in what "making" looks like. One is the why, the other is the how to navigate the transition.