Despite rapid generation of functional code, LLMs are introducing critical, compounding security flaws, posing serious risks for developers.

Baron Discovery Fund highlights a new position in JFrog Ltd. as a leader in binary management. Read the Q4 2025 report for full investment insights.

Baron Discovery Fund reports Q4 2025 performance and highlights key holdings like Exact Sciences and JFrog. Read the full ...

基准测试本身也引发了讨论。Opus 4.5在去年11月达到63.3%，12月却跌到43.8%，波动之大让一些人质疑其可靠性。但支持者认为这正体现了真实场景的不确定性，每月使用全新问题正是避免数据污染的设计初衷。也有人指出，模型在Python上被过度训练，希望看到持续更新的多语言基准测试，比如Elixir或Rust。