A study of 26,811 secondary students in central China found that homework scores rose 18% after students adopted generative AI tools, while closed-book exam scores fell 20%. The OECD’s latest PISA results found that daily AI users scored lower in science than nonusers, while moderate users performed about the same.
It's almost as if using AI only as a tool for execution rather than actual architectural decisions works out well. In HN universe this is the difference between an experienced architect using AI to write code vs a TikTok hipster to tell codex/claude to create an instagram clone for them so they can make a video about it "We dOn'T nEeD eNgiNeErs AnyMoRe"
spwa4•15m ago
As someone who occasionally reads psychometric papers: WOW. That's an impressive result. That is equivalent to a very, very good teacher.
contemporary343•9m ago
uptrails•8m ago
More generally, the problem is almost always motivation and discipline, not the availability of instruction. If you have a motivated kid, you can simply buy them some books. The problem in public schooling is that most kids are not motivated and the emphasis is on getting everyone to the finish line.
And if you want to use AI in education, I guess the big question is whether it can help the kids who have parents who work two shifts, who get pulled into youth gangs, or play games all day and don't even show up to school. Helping talented kids from well-off homes is really not that much of a challenge.
benmanns•6m ago
cyanydeez•4m ago