When you teach kids, their early responses are generally not even as good as mediocre. You wouldn't get a six-year-old to write your quarterly report.
In fact, all learning works that way. When you learn a foreign language, you start by being wildly incompetent, but only by persisting do you attain competency.
So how when will kids gain the ability to read and understand complex texts and concepts if they are allowed to use LLMs from the start?
As an example, I could perhaps communicate in Chinese if I used an automatic translator all the time, but that wouldn't help me learn Chinese at all, in the same way that driving from A to B doesn't make me a better runner.
Learning how to reason, to read texts critically, to express one's own ideas: these are not like building a keyboard, these are skills that all adults should master. How will this happen?
The argument that people don't change over the generations is not correct. In English-speaking countries, kids are fatter, and venture much less far from their house, than they did generations ago, due to changes in technology: Talking to teachers, at least in the Anglosphere, is pretty frightening now, and it didn't used to be: when the Internet first appeared, teachers would tell me how much better their kids could write than ever before, and how informed they were, thanks to the Internet. Now they tell me that their kids are passive, and hand in LLM results as assignments when they obviously haven't read a word of the outputs.
And again, I bring up this argument: what will these kids do as adults if they rely on LLMs to do their reasoning?
Automation, and other forces, have not been kind to the west: when I was young, a single parent working in a factory could support a stay-at-home wife, send their kids to school, and retire comfortably. Now two working parents can't do that.
What happens if we automated away everything? Do you think the people who took all human knowledge, freely given, and wrapped it in privately owned LLMs will share the wealth? There's no rational basis to believe that.
Thanks for an interesting conversation!