In the sci-fi series “Stargate SG-1”, the Goa’uld dominate the galaxy by relying on scavenged ancient technology without understanding the principles behind it. This might just be a mirror image of humanity’s future.
01. The Silent Collapse of the “Middle Tier”
A few days ago, I had coffee with a Tech Lead from a major tech giant. Frowning, he complained to me: “The fresh graduates we hire now write code faster than I did back in the day, but the moment they encounter a system-level Bug, everyone’s first reaction is to throw the Error Log into GPT, rather than checking the source code.”
This is actually quite terrifying.
The AI boom we are witnessing now is essentially a brute-force harvesting of humanity’s stock knowledge. Large models act like greedy mining machines, tirelessly excavating the experience humans have accumulated over hundreds of years. Then, business owners discovered—hey, this mining machine is much more useful than junior employees! It doesn’t complain about overtime, doesn’t need social security, and produces output incredibly fast.
Consequently, designers, programmers, junior copywriters, and even medical interns—the “waist strength” or middle tier that once supported the base of the industry pyramid—are being ruthlessly hollowed out.
But this brings about a tricky logical paradox: All senior experts crawled their way out of the rookie pile.
It’s as if, to save trouble, we sawed off the bottom rungs of a ladder because we thought, “Anyway, we’re going to the top floor.” But the question is, how will the young people of the future get up there? Do we expect them to be born as architects or chief physicians?
When AI proxies the process of “honing skills,” it simultaneously confiscates the result of “growth.” Without the days and nights spent rolling around in garbage code, without the painful experiences of being ground into the dust by clients, the so-called “senior intuition” becomes water without a source.
02. The “Inbreeding” of Synthetic Data
Even more darkly humorous is that if the first point holds true, there will be no new human experts to generate new knowledge in the future. So, what will AI eat?
The answer is: It will eat what it produces itself.
There is a specific term for this in academia: “Model Collapse.”
This isn’t just a chart; it’s a “dementia diagnosis” for AI. When AI starts consuming data produced by AI, information entropy begins to increase until all output turns into meaningless babble.
Imagine if all articles were written by AI, all paintings generated by AI, and all code completed by AI. Then, the next generation of AI models uses this data for training. This is like a royal family engaging in frantic inbreeding to maintain blood purity; in the end, all you get is the deformed Habsburg jaw.
A recent paper in Nature has confirmed this: when models are trained on recursively generated synthetic data, they rapidly forget long-tail knowledge, their output becomes homogenized, and they even exhibit severe cognitive distortion.
If humans no longer produce new, flawed but original knowledge, AI’s ceiling is right here today. We would truly end up living like those alien races in Stargate—guarding a pile of god-level technology without knowing how it works, capable only of blind worship and unable to advance a single step further.
03. A “Luddite” Counterattack in Medicine?
Turning our gaze to the medical industry, you’ll find an interesting phenomenon: doctors are far more vigilant about AI than internet professionals.
Is it because they are conservative? Is it because they don’t understand technology?
No, it’s because they understand “fault tolerance” too well.
In the internet industry, if a recommendation algorithm pushes the wrong video to a user, the user just swipes away. But in the medical industry, if AI gives a wrong diagnostic suggestion, it could mean a life.
This chart reveals the hidden worry in the medical field: although AI is low-cost, on certain key metrics, the “intuition” and “sense of responsibility” of human doctors remain a moat that algorithms cannot quantify.
Many top hospitals resist the full-scale rollout of AI diagnosis not to protect vested interests, but to protect “uncertainty.” The essence of medicine is an empirical science; breakthroughs in many intractable diseases rely on the “hand-feel” doctors have honed through massive amounts of non-standard cases. If this generation of young doctors gets used to reading AI-generated diagnostic reports, that acute sense of smell when facing unknown pathologies will completely degenerate.
It’s like having autopilot and not teaching pilots how to make a manual emergency landing. It’s fine when things are normal, but once the system fails, it’s a crash with total fatalities.
04. The Sweet Trap of Cognitive Offloading
Let’s talk about education.
Major tech companies are now pushing “AI in schools,” handing out free computing quotas to students under the banner of “educational equity.” The picture is too beautiful; it resembles that classic metaphor: Giving an electric wheelchair to a child who wants to learn to walk.
This isn’t just about “coping with schoolwork”; this is massive-scale “Cognitive Offloading.”
It looks like it saves you time, but it actually steals your brain. When the process of thinking is skipped, the critical thinking humans are so proud of becomes mere decoration.
When you don’t need to build logical chains in your brain and only need to input prompts, your cerebral cortex will atrophy just like muscles that have been bedridden for too long.
Is this really helping society progress? Or are these giants mass-producing a generation of “perfect consumers”? People who don’t understand principles, don’t question, have no creativity, and only know how to skillfully use tools provided by big tech, contributing subscription fees docilely like penned sheep.
05. Things That Cannot Be Defined by Code
Having said all this, I’m not asking everyone to smash their computers and return to the fields. AI is a tool; that is beyond doubt.
But we must be vigilant: Do not let the tool turn the tables and define our value.
Human dignity often lies not in the perfection of the result, but in the struggle of the process.
It is because you stayed up all night with red eyes to write a piece of code that you understand the beauty of architecture;
It is because you scoured through literature for a diagnosis that you hold reverence for life;
It is because you wandered late at night trying to figure out a philosophical problem that you possess a soul.
If all of this is omitted by AI in the name of “efficiency,” then what remains of us might just be hollow shells, engaging in meaningless interactions with a cursor on a screen.
In this era accelerated by algorithms, “slowing down” has surprisingly become the highest form of resistance.
Retain a bit of “low-efficiency” learning ability; that might be the last ticket humans have before the Stargate closes.
References：
- AI models collapse when trained on recursively generated data – Nature
- Is AI closing the door on entry-level job opportunities? – WEF
- Barriers to and Facilitators of Artificial Intelligence Adoption in Health – PMC
- Educational impacts of generative artificial intelligence on learning – Nature
- Stargate SG-1 Ancient Technology – Fandom
