Jev's hype cools after launch as technical claims draw scrutiny
What happened
A Hacker News front-page post on October 23 traces the general-purpose classifier Jev through 33 days of viral attention and rapid cooling. The author says Jev's "cannot hallucinate" claim only means its output matches a schema, not that it is logically correct, and that its real error rate is higher than ordinary language models. The post also questions Jev's novelty, saying a similar model, Laya, came 1.5 years earlier, and that open-weight models Laya, Kev, Nimble and NanoJev perform better.
Written by AI from the coverage · updated 49 minutes ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageThe Rise and Fall of Jev
通用分类器 Jev 发布 33 天内经历爆红与迅速降温，其宣称"无法幻觉"实为输出符合 schema，并不代表逻辑正确，实际错误率甚至高于常规语言模型。作者指出 Jev 并非全新思路，类似模型 Laya 早其 1.5 年提出，且已有 Laya、Kev、Nimble、NanoJev 等性能更优的开源权重模型。
Heat over time
Not enough continuous observations to draw a trend yet.