即使在2026年，把AI从demo做成产品也不是一件容易的事。比如中翻英，大家都觉得早就被LLM解决了，不就是调个API的事情嘛。但我们最近因为要把Superlinear Academy社区加入一个中翻英自动同步的功能，才发现开发体验这么差。
这个问 …
Articles in the Computing category
From Process Certainty to Outcome Certainty: A Different Kind of Confidence in the Age of AI
Even in 2026, turning an AI demo into a production-ready product is surprisingly hard. Take Chinese-to-English translation. Everyone assumes LLMs solved this ages ago—just call an API, right? But when we recently tried to add an automatic translation sync feature to the Superlinear Academy community, we discovered just how …
使用Nano Banana Pro生成整套PPT：疯狂，挑战和工作流
最近Nano Banana Pro（NBP）非常火。我们看到了各种应用：给漫画上色、生成海报、甚至直接生成以假乱真的演示文稿PPT。看起来像一张PPT …
Generating Entire Slide Decks with Nano Banana Pro: The Madness, The Challenges, and The Workflow
Recently, Nano Banana Pro (NBP) has been on fire. We’ve seen all sorts of applications: coloring comics, generating posters, and even creating photorealistic presentation slides. They look like slides, but in reality, they are just images.
However, in actual presentation workflows, most people still treat it merely as an …
一次性软件与被压缩的现实：AI Native 的本质是策略重构
一次性软件与决策解析度
一切从一个十分钟的需求开始。在机器学习研发工作中我们用Labelbox来让人类标注数据。有一次我发现标注的数据质 …