AI Technique

为了解决车库门忘关的问题，我用拇指大小的ESP32-CAM单片机与AI协作做了一个端侧神经网络传感器。本文记录了如何通过打通全自动串口研发闭环、解决数据极度不对称与QAT量化衰减，在10小时内跑通嵌入式AI的完整过程。

To solve the problem of accidentally leaving my garage door open overnight, I collaborated with AI to build an edge neural network sensor on a thumb-sized ESP32-CAM. This article details how we completed the end-to-end telemetry loop, tackled extreme data imbalance and QAT quantization loss, and got edge AI running on a microcontroller in just 10 hours.

月食时月面边缘有一条青绿色的窄带，科普说那是臭氧吸收。但为什么是窄带不是整圈？为什么全食最深时反而看不见它？我们从最土的白圆盘开始，一层层加物理，硬算出这条带，在一路翻车里发现了一个更大的问题：AI 太懂物理，反而会把你带进前人留下的近似里。

During a lunar eclipse, a narrow green-blue band appears at the moon's edge; popular science says it is ozone absorption. But why a narrow band instead of a full ring? Why does it vanish at deepest totality? We start from the simplest white disk, layer in physics, and compute the band — and through repeated failures discover a bigger problem: AI knows physics too well, and will happily lead you into approximations left behind by earlier researchers.

从最直白的"一星一像素"出发，八次翻车、六亿颗星，一步一步把银河从真实星表里逼出来。在这个过程中才发现，以前从来没认真想过头顶的星空为什么长这个样子。

Starting from the simplest "one star, one pixel" rendering, eight failures and 600 million stars later, the Milky Way slowly emerged from a real star catalog.

Prompt不是需求本身，而是用信息搭建的临时世界。新模型实现了context-driven emergence：AI在沉浸式语境中被迫变聪明，从你的草稿和删除痕迹中长出更懂你的能力。

A prompt is a temporary world you construct with information. New models enable context-driven emergence—AI becomes smarter within immersive context, learning your preferences and style from unfinished drafts and deleted text.

四步实现GPT文生图的商业级精准可控：用视觉撬动AI思维、用遮罩迭代生成、让AI当合伙人、用文档管理打造AI资产库，生成可复用的统一视觉风格。

Four techniques to achieve commercial-grade flowcharts and visualizations with GPT's image generation: visual prompts, masking for iteration, AI brainstorming, and document management for reusable assets.