TLA+ goes viral, but modeling is only the start
What happened
Boris Cherny 一条爆款推文让 TLA+ 这个三十多年的老工具重新进入大家视野。简单说，TLA+ 是一种描述系统能做什么、以及什么绝对不能发生的语言，比如“绝不能同时有两个 leader”。它自带的模型检查器 TLC 只能穷举有限实例，而且检查的是模型，不是真实代码，所以模型和实现之间天然有距离。Reasonable 团队把这件事往前推了一步...
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pagePickTLA+ goes viral, but modeling is only the start
Boris Cherny's viral tweet put TLA+ in the spotlight. This Reasonable post explains TLA+ as a language for describing system behaviors and temporal properties like 'never two leaders at once.' TLC model checking only explores finite instances, and the model isn't the implementation. The team turned 16,000+ TLA+ spec/property pairs into 3,000+ machine-checked Verus proofs, aiming to connect specification, proof, and Rust code. The post doesn't disclose accuracy or latency numbers for this agentic pipeline.
Heat over time
Not enough continuous observations to draw a trend yet.