I think you make some valid points. I like to imagine most of us have other interests and projects we are engaged with and my be less motivated in some areas when we engage with other things. This is almost always the cause if my headline hot take behavior or unwillingness to read a text wall. I’m primarily here for the inadequate dopamine hit of social media; not as much for the personal growth potential.

I think the primary issue is an education system that makes reading and learning a nuisance and chore. This is a problem that can be solved in the coming decade with the use of technology, but it will take a serious overhaul of the entire system. There is no room for proprietary software and exploitation in education. The entire system should be standardised on open source software, funding should be allocated to run a small independent and offline AI server and the teacher’s role should be divided between the AI system and a traditional group oriented role. This will allow individualized education without exploitation. An AI agent that is specifically designed for this task and paired with the teacher’s supervision makes it possible for each child to follow the path that best suits them. They can read any book they want that meets certain requirements. They can progress at their own pace. Issues can be identified long before any current teacher is capable of spotting. Most importantly, this is not about AI as a product or replacing the teacher in any way. This is making use of a tool, and doing so ethically. This kind of thing can not be done for profit or by contractors. The privacy of such a system should be of paramount importance that is not possible long term with any company focused on profitability. The only people with access to the AI should be the students, parents, and teachers. Even IT staff at the school should not have access to the AI logs and data, and there should be no persistent storage long term. It has to be a tool that is used by the teacher only.

To be clear, I am a hobbyist working on such a tool for my own self education with the computer science curriculum. This is about AI agents. This is not about a raw AI LLM. An agent is a collection of LLMs connected through a code base, and connected to databases. This does not rely on the model training alone for answers. This is a system where the final answer is checked and reviewed multiple times and verified against accurate sources before a final reply is made. Most people here are likely unfamiliar with this and what it is capable of doing.

This is the inevitable future, it is only a question of how long it takes people to adapt to the new potential. This level of individualized education has only been available to the ultra rich, but it is now possible for everyone at scale.