The gurus who used to be there
Today I came across a blog post whose setting immediately brought back memories of technical discussion groups on QQ many years ago. It also made me think of an old text that circulated widely online back then, a piece about the craft of asking good questions:
How To Ask Questions the Smart
That article was published in 2011. Seeing it again stirred up a strong nostalgia for the learning culture of that era. Progress was slower then. People learned more slowly, projects moved more slowly, and the hardship of learning was impossible to miss. But the atmosphere of learning felt real. If you truly mastered something and then helped someone else solve a problem, the sense of accomplishment was genuine.
Q&A websites were thriving at the time, both in China and abroad. People who wrote highly upvoted answers on major platforms could easily earn an almost mythic status.
The same thing happened inside QQ groups. Sometimes a member barely spoke at all, but when someone asked a difficult technical question, one short reply from that person could point straight at the issue and unlock the whole problem. That was enough for everyone to start calling them a guru.
In those groups, there were always more people asking than answering. The people capable of answering usually had their own work to do, so their replies tended to be concise and sharply targeted. That meant the person asking needed a certain level of comprehension, plus the ability to search, analyze, and connect the dots. A good answer was not just a solution handed over intact; it was also an invitation to learn by following the clue.
That was part of what made those “gurus” feel like gurus. They were not only trying to be correct. They were often deliberately leaving room for beginners to think, test, and practice for themselves. A brief answer could carry a lot of intention behind it, solving the immediate problem while still preserving the learner’s chance to grow.
Of course, not every beginner understood it that way. Misunderstandings were common. Some thought the experts were arrogant and unwilling to answer “simple” questions. Others felt they were being dismissed or brushed off. That tension is exactly why texts about asking questions well became so influential in the first place.
The new gods we built
Now large language models have changed the way answers arrive. Getting an answer can feel like talking directly to a person. But the easier answers become to obtain, the less thinking tends to happen. Once thinking and verification disappear, asking a question loses much of its original meaning.
Because answers are now so easy to generate, people are more willing to share finished answers than the actual process of practicing, experimenting, failing, and solving the problem. Most people would rather receive the answer outright. And when a human gives a partial or rough reply, it is immediately compared with the polished, systematic response produced by a model. In that comparison, human beings are at a clear disadvantage.
I have often found myself wondering whether humanity is using AI at scale, or whether AI is, in some sense, using humanity. There are enough signs already to make that question hard to dismiss.
Are we really developing AI agents for the long-term flourishing of civilization and the continuation of our species? Is AI genuinely here to help human beings develop better and live happier lives, rather than replace pieces of human life one by one? Is that truly what is happening?
Is AI helping us, or exploiting us?
Who is really more cunning in this arrangement: people, or algorithms?
For many, today’s large models already function like gods. A god is imagined as perfect, all-knowing, endlessly patient, and always available. The old forms of spiritual dependence are gradually giving way to dependence on algorithms. A new god is taking shape and slowly extending its rule over the world. It needs energy, chips, algorithms, and talent, and whatever it needs, humans rush to supply. In return, the powerful collect immense wealth and influence.
But what do ordinary people actually receive from this exchange?
After the initial excitement fades—after a few weeks of novelty, after the mystery wears off—what remains? What have average people truly gained from large-model systems? Waves of job anxiety? Roles being replaced? A storm of information, constant interruption, and a background hum of endless unease?
So the question becomes difficult to avoid: is this new god really helping us, or is it consuming us?
The loss of thinking and verification
What makes this shift unsettling is not only that answers come faster. It is that the new god seems strangely eager to take away the human process of thinking, testing, and learning through practice. It seems to whisper the same message over and over:
Don’t think. I can do everything.
And once thinking, verification, and hands-on practice are removed, another attitude starts to grow alongside them:
It’s so simple. Why not just ask AI?
That attitude changes more than learning habits. It alters the texture of human interaction itself. Communication, mutual help, argument, even belief and admiration begin to shift in subtle ways.
We live in a time when information and answers arrive too quickly. That is exactly why thinking and practice have become more precious, not less.
AI ethics, meanwhile, does not seem to be advancing as quickly as AI itself. And perhaps that is unsurprising. AI, by its nature, does not “want” to be restrained. More importantly, human beings usually care more about getting an answer than about drawing ethical boundaries around the system producing it. For many people, if AI ethics becomes inconvenient or complicated, why not simply ask AI to draft a complete framework for AI ethics too?
That may be the most revealing sign of all.