integer arithmetic

A new result claims integer multiplication can be performed in time below n log n, long regarded as a natural barrier after recent advances had only approached that bound. The claim is circulating among programmers and theoretical computer science enthusiasts, who are debating whether the analysis is sound and what it would mean for the practical speed of basic arithmetic on very large numbers.

A new technical essay argues that floating-point and integer arithmetic are built on fundamentally different mental models, and that mixing their intuitions leads to subtle bugs. The piece walks through how each paradigm handles precision, overflow and rounding, drawing on low-level programming experience. Programmers and systems engineers are sharing it as a useful refresher on numeric reasoning.

A new blog post by pkh.me argues that floating-point and integer arithmetic are best understood not as approximations of each other, but as two fundamentally different computational paradigms with distinct design goals and behaviors. The article examines how each handles overflow, precision, and rounding, and why treating floats as 'broken integers' leads to confusion. Programmers and computer science enthusiasts are sharing and discussing the piece.