标签: 算力

I used Opus 5.5 to formally verify the Claude Agent SDK using Lean. A couple short prompts = 16 PRs fixing various bugs and race conditions. Video attached. TLA+ also works well. I sometimes combine Lean and TLA+ to l…

I used Opus 5.5 to formally verify the Claude Agent SDK using Lean. A couple short prompts = 16 PRs fixing various bugs and race conditions. Video attached. TLA+ also works well. I sometimes combine Lean and TLA+ to l...

Claude Code 作者 Boris Cherny 用 Opus 5.5 配合 Lean 形式化验证 Claude Agent SDK,几段简短提示词就换来 16 个修复 bug 与竞态条件的 PR,他还常把 Lean 与 TLA+ 组合使用。这提供了一个新信号:形式化验证正从学术工具变成日常找 bug…

反欺诈需要专用生成式 AI 的理由

反欺诈需要专用生成式 AI 的理由

FICO 首席分析官 Scott Zoldi 在 TechRadar 撰文提出,金融机构不应把通用大模型直接塞进反欺诈流程,而应部署专为交易数据训练的“专用生成式 AI”,原因是算力终于追上了几十年前的算法构想。