Become A Tinkerer
AI collapsed the skill gaps that made organizational specialization rational, but companies still operate like it's 1913—and the tinkerers who ignore job titles to own complete problems are eating everyone's lunch.
Read Original Summary used for search
TLDR
• Ford's assembly line made sense when cross-training a welder to paint took months; software companies copied the model but AI eliminated those skill gaps in months while org charts stayed frozen
• Fast-growing startups now hire tinkerers who do it all—marketers building landing pages in Claude, engineers creating launch videos, founders shipping customer research → design → code → launch solo because "I had all the context"
• The real divide isn't generalist vs specialist, it's ownership vs dependency: a rough version you ship with full context beats whatever arrives six weeks later after three handoffs stripped out everything that matters
• Companies pay more in tokens than new hires now, letting teams keep startup scrappiness at 100-200 people instead of fragmenting into specialized stations
• Most people won't cross lanes because titles become identity and performance reviews punish it, but companies that don't fix this will bleed talent to ones that already have
In Detail
The author argues that organizational specialization—PMs writing specs, designers mocking up, engineers building, QA testing—was a rational response to real skill gaps, just like Ford's assembly line made sense when training a welder to paint took months. Agile didn't fix the core problem; it just made the handoffs faster while keeping all the stations intact. But AI collapsed those skill gaps in months, and the org chart stayed exactly where it was.
The evidence is in how fast-growing startups now operate: they hire tinkerers, not specialists. The marketer builds landing pages in Claude, writes ad copy, and launches campaigns with no tickets or handoffs. The engineer ships features, creates launch videos, spins up agents to review PRs, and owns outcomes. A founder walked the author through a feature he shipped solo—customer research, design, code, launch—because "why would I hand it off? I had all the context." These teams pay more in tokens than they'd spend on new hires, maintaining startup scrappiness at scale.
The friction is real: titles become identity, crossing lanes feels wrong, performance reviews punish it, and you'll be bad at first. But a rough version shipped with full context beats whatever gets delivered six weeks later after handoffs strip out everything that matters. The real question isn't generalist versus specialist—it's ownership versus dependency. Companies that don't give people space to cross stations will lose their best talent to ones that already have, because someone without your title, permission, or experience is shipping the thing you've been waiting to hand off.