The Good Regulator, Part Three
Fictionalization of a Mathematical Proof. 700 words, 2 minute read. With Claude.
Elena's coffee had gone cold three hours ago, a film of separated cream floating on its surface like the lipid barriers of dying cells. Through the sealed door of Lab Four, she could smell it—that peculiar sweetness of three years of research decomposing. The temperature readouts mocked her: 67°F. 84°F. 71°F. 92°F.
"How bad?" Marcus didn't need to ask what. He'd been there when she'd isolated the strain, watched her coax it through hundreds of generations until it could synthesize insulin at twice the rate of anything published.
"Bad." Elena's fingers drummed against the desk. She had spreadsheets open on both monitors—on the left, temperature control logs; on the right, her dating optimization matrix. Red cells dominated both. "The controller's having an identity crisis. Watch—" She highlighted a data cluster. "External temp hits 75, sometimes it heats, sometimes it cools, sometimes it does nothing. It's like it's trying every possible response to see what works."
Marcus pulled up a chair. "Expensive way to learn."
"Twenty thousand dollars of specialized media. Gone." She minimized the dating spreadsheet, but not before Marcus caught sight of it.
"Is that—"
"My mother made me promise to try a 'scientific approach' to dating." Elena's laugh had no humor in it. "Vary the parameters. Try different types. Collect data. Optimize."
Her phone buzzed. Craig, the accountant. Yesterday it had been Miguel, the CrossFit trainer. Tomorrow, her mother had promised, a chef who'd been on some cooking show.
Marcus studied the temperature chaos. "My ex used to do this thing. She'd constantly change how she responded to situations, trying to figure out what I wanted. One day she'd be super independent, the next clingy, then somewhere in between. Said she was 'adapting' to me."
"How'd that work out?"
"Like your bacteria." He gestured at the dying cultures. "Organisms need consistency to thrive. Constant change is just another form of stress."
Elena stared at the screens. Three years of work dying in one room, three years of failed relationships catalogued in the other. Both victims of the same fundamental error—the belief that trying every possible response would eventually yield the right one.
"The Good Regulator Theorem," she said quietly.
"Come again?"
"It's an old cybernetics principle. A good regulator has to be a model of the system it controls." She grabbed a marker, began sketching on the whiteboard. "You can't regulate something by throwing random responses at it. You have to understand what it needs and respond consistently."
They worked in silence, mapping temperatures to responses. One input, one output. No randomness, no optimization matrices, just clear, deterministic functions. Elena thought about her strain—how she'd learned its needs through careful observation, not random experimentation. It required exactly 37.2°C, pH 7.1, glucose levels that never varied more than 0.5%. She'd given it those conditions religiously, and it had thrived.
"Ready?" Marcus asked, finger hovering over the activation key.
Elena nodded. The system initialized, and the temperature readout stuttered, searching for its new equilibrium. Then: 72°F. Steady. The number held like a kept promise.
Through the lab door's window, she could see the culture flasks. Still cloudy. Still alive.
Her phone buzzed again. Without looking, she deleted the dating app, then her optimization spreadsheet. Three years of data, gone.
"Harsh," Marcus observed.
"No." Elena picked up her phone, scrolled to a different contact. David from the microscopy core. They'd talked for hours last week about bacterial metabolism, forgetting to eat lunch. She'd felt her thoughts settling into familiar patterns, like pH finding its balance. "I already know what conditions I thrive in."
She typed: Want to grab coffee and discuss that paper on metabolic flux?
His response came quickly: Only if we can argue about their statistical methods.
Elena smiled—a real one this time. Some systems revealed themselves not through random sampling but through patient observation. And some responses didn't require any modeling at all, just the recognition of compatible conditions already in place.
The cultures in Lab Four continued their quiet metabolism, converting glucose to insulin in the steady warmth Elena had finally learned to provide. Through the window, morning light transformed the growth media into something golden, something alive.