eric zakariasson@ericzakariasson
19AI 编辑部评分,满分 100
2026-07-30 20:35· 2天前
跳到正文
AI 摘要

每当新模型发布时,许多人会根据自己实际做的事情,用自己的方式来评测它。 对我来说就是编程相关的事情。给现有项目添加功能、从头构建一个FPS游戏、克隆一个现有应用等等。 你用什么来测试新模型?

whenever a new model drops, many people has their own way to benchmark it based on what they actually do

for me its coding stuff. adding a feature to an existing project, building an fps from scratch, cloning an existing app etc

what do you use to test new models?

eric zakariasson · @ericzakariasson · X·2026-07-30 20:35·2天前
在 X 看原推· x.com
AI 摘要

每当新模型发布时,许多人会根据自己实际做的事情,用自己的方式来评测它。 对我来说就是编程相关的事情。给现有项目添加功能、从头构建一个FPS游戏、克隆一个现有应用等等。 你用什么来测试新模型?

whenever a new model drops, many people has their own way to benchmark it based on what they actually do

for me its coding stuff. adding a feature to an existing project, building an fps from scratch, cloning an existing app etc

what do you use to test new models?

在 X 查看原推x.com