Dalet Labs 获得由 Lightspeed 领投的 700 万美元 pre-seed 轮融资，用于构建 AI 生成的购物场景。其核心理念是通过 AI 在用户提出明确需求前展示选项，利用大语言模型根据用户喜欢的图片属性构建扩散模型提示词以减少幻觉。同时，该 Feed 流可根据用户互动自主生成和混音内容，无需实景拍摄即可扩大图片池。
AI is becoming the new front door for brand discovery.
The shopping mall is a surprisingly good mental model for consumer AI.
Dalet Labs raised a $7M pre-seed led by Lightspeed to build AI-generated shopping scenes.
The core idea is that AI should show options before you know what to ask, unlike chatbots and shopping agents that need a clear request before they can help.
For grounding, a language model builds diffusion prompts only from attributes found in images people liked, which the company says limits hallucinated details in scenes meant to show real products.
The feed generates its own content: collective user engagement picks which images to remix next, and fresh outputs can become new references, so the image pool grows without photoshoots.
来源：Rohan Paul · x.com