AI Built the Bomber, But the Workplace Still Needs You to Be That "Hummingbird"
“Man, I’m really about to lose it lately.” I stirred my cup of cold coffee, staring at the screen on my phone stuck on some big tech company’s news page about their latest inference model—all in English, and my eyes were killing me. “Look at how fast things are changing; it’s not giving anyone a break. I feel like there’s new stuff to learn every day, and I can’t keep up. But if I don’t, I’m scared I’ll get left behind by the folks using AI and end up out of a job.”
Across from me, Lao Bai took a deep drag from his e-cig, the vapor drifting in the dim cafe light. He’s been in the internet game for nearly twenty years—saw the live-streaming wars, the O2O bubble, all of it. A real industry vet.
He listened, then smirked and chuckled. “Kid, you’re giving the people in the office way too much credit. You think everyone’s out there revolutionizing things with AI? Come on. Most companies haven’t even scratched the surface.”
“But my feed’s full of people showing off how they use AI to make viral videos or write code… They seem to be killing it.”
“That’s survivor bias,” Lao Bai said with a shrug, his voice even. “Most employees and bosses have no clue about what these models can really do. They still think ChatGPT is just a fancy search engine or writing assistant. At best, it’s replying to emails, summarizing meetings, or looking up info. Real deep research? They haven’t even found the door. Like I heard one team griping about Notebook LM: ‘We’ve been messing with this for months, digging up new features every week!’”
He leaned in, eyes locked on mine. “It’s like handing someone the internet in 2000, and they just use it for email. They’re thrilled, no idea that search, e-commerce, and social media are about to flip the world upside down. You’re panicking because you’re ahead of the curve; but your coworkers are still swatting mosquitoes with a cannon.”
“So, if I just learn a bit more and use it a bit more, I’ll be fine?” I pressed, my voice urgent. “But won’t AI eventually crush us all, become the ultimate worker, and leave us in the dust?”
Lao Bai shook his head, wagging a finger. “Crush us completely? That’s off base. I like using physicist Feynman’s ‘artificial flight’ analogy for this.”
He blew a smoke ring and went on. “‘Flight’ is just a property. Birds fly, planes fly, but totally differently. Birds flap down for lift; planes thrust forward and use wing pressure differences to rise. Not the same at all.”
“What’s that got to do with AI taking our jobs?” I scratched my head, not quite getting it.
“It’s huge,” Lao Bai said, narrowing his eyes, his tone serious. “Planes are amazing—they changed the global economy—but they can’t do what birds do. No plane’s as efficient as an albatross, as agile as a bat, or can hover and pivot on a dime like a hummingbird. Sure, planes carry more and go faster.”
He paused, looking at me. “AI right now is that plane, and you’re the bird. It beats you hands down at sifting through massive messy data, building complex toolchains, or doing math reasoning. But it doesn’t get the nuances of human interactions, cultural trends, or how to push and grind for a goal like a person. Get it?”
“So our edge is being the ‘hummingbird’—flexible, intuitive stuff?” My eyes lit up as the fog cleared a bit, and I couldn’t help but smile.
“Exactly.” Lao Bai snapped his fingers. “Without human intuition, it can cause big messes in key areas. Let me tell you a real story from the medical field.”
“Medical? That’s life and death.”
“Yep. At Mount Sinai’s cath lab last year, they rolled out an AI called Sofiya to replace nurses calling patients about pre-op stuff. Saved over 200 nursing hours in five months.”
“Sounds helpful, saves time.”
“Helpful? If you were a nurse, would you think so?” Lao Bai snorted. “At a New York City Council meeting last year, a 37-year veteran nurse called it out, saying Sofiya’s info had to be double-checked by nurses to be reliable. And UC Davis Health’s BioButton for monitoring vitals—it was alarming nonstop in wards. Nurses were fed up. Execs hyped it for predicting hemorrhagic strokes early, but it was just heart rate, temp, stuff like that. By the time those change, the patient’s in trouble. Piloted for a year, then yanked.”
I sat there stunned. “So the ‘plane’ hauls a ton, but can’t ‘hover and turn’ in a hospital room?”
“Right, it lacks on-the-ground human instinct.” Lao Bai leaned back. “For now, patients’ lives depend on doctors’ gut feelings. In the AI era, a nurse’s value might not be following orders, but overriding them. Think about it—not just nurses, but every job. What saves your spot is the guts to tell AI, ‘That’s wrong.’”
“Telling AI no…” I muttered, my mind racing.
Lao Bai took a sip of water, slowing down. “Chatted with a Freshworks exec the other day—she said she once rushed and had AI write design principles, slapped them on slides. Mid-presentation, she realized it was all off, total embarrassment. The cost of blind trust. If AI spits out 50 ideas, you need the eye to pick the winners. Taste, process, structure—tools can’t give you that.”
He stubbed out his e-cig, eyeing me. “Lots of folks want to be Iron Man, automate everything in one go. Suicide mission. Treat AI like a super intern. You manage it, don’t depend on it.”
“So what do we do? Can’t just ignore it.”
“Use it, experiment, but don’t ditch the basics.” Lao Bai stood, patting my shoulder as he headed out. “What worries me most is five years from now—no one’s doing the foundational work anymore. It’s like climbing while burning the ladder below. When it’s gone, if AI glitches or goes dumb, we’re the ones falling.”
The cafe door swung open, and Lao Bai stepped into Beijing’s early spring night. I stared at my phone’s blinking cursor, suddenly feeling less anxious. At least for now, the one holding the keyboard is still me.
Source:
#197: Something Big Is Happening, Claude Safety Risks, AI fo
Every AI Company’s Safety Plan is ‘Use AI to Make AI Safe’.
AI Is entering health care, and nurses are being asked to tr
I’ve helped lead AI adoption at both PwC and Freshworks. Her