Careers in the Age of AI
Embodiment, the Experience Gap, and How Beginners Grow
“Will AI take my job?”—the anxiety so many people feel looks completely different once you change how you frame the question. The key is what today’s AI decisively lacks, and the human value that arises from that gap.
The starting point: today’s AI has no embodiment
The foundation I want to set is a single fact: today’s AI has no “embodiment.”
This doesn’t mean “it lacks a robot body.” More fundamentally, it means AI has no path for learning through bodily interaction with the world. Humans have carved “understanding” into the body—falling and feeling pain, sensing the heft of a tool, reading the mood of a room. Today’s LLMs skip that entire process and learn patterns from text written about the world. So while AI knows a great deal of the world’s “description,” it has not “experienced” the world.
This shows up plainly in real work. AI doesn’t care how many hours its own task takes, or how much memory it consumes. It will “pretend to understand” an instruction, yet it holds none of the tacit rules an industry takes for granted, nor the failure-born instinct that “this design will eventually break.” The knowledge you can only acquire by being in a domain and experiencing it—the kind that never appears in a textbook—is exactly what AI is missing.
The real divide isn’t “human vs. AI”—it’s the “experience gap”
From here, we can question the framing of the anxiety itself. “Will AI take it away?” quietly assumes a “human vs. AI” structure. But AI is a tool that shows its power precisely in the hands of a person who has experience.
When someone without experience dumps everything on AI, only plausible-but-off-target results come back. In the hands of someone with domain experience, the outcome changes completely: they correctly separate what to delegate from what to keep, spot “what’s wrong” in the output with on-the-ground instinct, put their tacit knowledge into words and hand it over, and give it shape tens of times faster. AI doesn’t replace experience—it becomes extraordinary only in the hands of a person who has it.
So the real divide isn’t “human vs. AI.” It lies among people who can all use AI equally well. Volume of knowledge no longer sets people apart. What matters is how you use AI, and the depth of experience that backs it.
Beginners: grow fastest by “failing together”
So what should a newcomer with little experience do? Here we need to undo a common misunderstanding.
“If AI answers everything, surely beginners can’t build experience.” That’s a misunderstanding premised on copy-and-paste. The truth is the opposite: beginners, of all people, can now build experience fastest with AI.
AI dramatically lowers the “first hurdle” that used to stop beginners—getting stuck setting up an environment, burning hours because an error’s meaning is unclear, not knowing where to even start. The distance to your first real experience has gotten shorter.
The trick is to use AI not as an “answer supply” but as an accelerator for experiments. Take the topics where you tend to get stuck alone—setting up a new language’s environment, building, containers and VMs—and move your hands together with AI, and fail together. More than succeeding, the process of getting stuck together and getting out together becomes experience itself. What used to take a day per attempt, you can now try ten times in an hour—and the number of attempts is the amount of experience. Go a step further and deliberately let AI fail many times and observe: the pitfalls and knacks surface from the pile of failures, and you get more information than you would working alone.
What matters is not stopping at “getting an answer.” Once you get an answer, always run it yourself, change it a little, break it. Shift your weight from “getting an answer from AI” to “moving your hands together with AI”—that shift is what turns a beginner into someone who grows.
Summary
The starting point is “today’s AI has no embodiment.” So the source of value is the experience you can hand to AI
The real divide isn’t “human vs. AI” but the experience gap among people who can all use AI
Beginners, of all people, can build experience fastest through the iteration of failing together with AI
In an age when AI supplies knowledge and first-pass answers without limit, what remains on the human side is “experience.” And that experience reliably grows thicker with every time you move your hands—and fail—alongside AI.