Why is it all in the kernel?

The blog post discusses the recent controversy surrounding a purported proof of the Collatz conjecture that was ultimately found to be flawed due to a bug in the Lean proof assistant. The author reflects on the implications of soundness bugs in proof systems and critiques the insistence on storing proof objects, suggesting that such practices do not enhance soundness as claimed. Instead, the author promotes the idea of deriving mathematical truths from minimal axioms through rigorous, principled approaches. The piece highlights the complexities of proof theory, recursion, and the pitfalls encountered in contemporary proof assistants.