Currently Online 
A warning from the field of mathematics. What can be wrong about AI solving problems?
AI is solving mathematical problems that have resisted some of the world's greatest mathematicians for decades. Why are mathematicians concerned instead of celebrating? The answer reveals a fundamental misunderstanding of what makes mathematics valuable. From Euler's seven bridges of Königsberg to the race to the Moon, I explain why solving a problem is not necessarily the same as advancing science, and why the real value of mathematical research often lies in the discoveries made along the way. More importantly, I explore a potentially devastating consequence of AI's growing capabilities: what happens when the next generation of talented students no longer sees a reason to become mathematicians? The implications could extend far beyond mathematics and affect the future of scientific progress itself.
Does this apply to coding as well? Sure the current crop of great coders are using AI to work faster, yet kids nowadays are learning just to let AI do all the work. Should AI be used to write entire essays, write entire programs or should it be restricted to help speed up mundane tasks.
The 1:1000 ratio (or whatever it is) might only get worse as where is the motivation to become a good coder and really understand how the code works. (And thus apply that knowledge to new projects and improve on it) Outsourcing critical and creative thinking is a bad idea.
As for ethics, who gets held accountable when a self-driving car kills their passenger or someone else. Same for AI developed treatments. Probably why Anthropic wants AI personhood so the AI can be blamed, just like blaming corporations rather than holding individuals accountable. As of now a case can be made against the AI developers when an AI developed treatment turns out to be harmful.
