AI概览

# REF
[阿里云百炼大模型 API](https://bailian.console.aliyun.com/cn-beijing?tab=model#/model-market/detail/qwen3.6-flash?serviceSite=asia-pacific-china)

# AI分层
- model
  - prompt
- agent
- skill
[软件构架设计/状态机设计.rst](https://github.com/Kenneth-Lee/MySummary/commit/5a56bca951b746f3054d0a03362281974c3ccbc5#diff-646956da0617cedf9eed59f6312937f0946004902ce91b5989ae3fc6faf9118d)
- harness

## code
cc-switch 配置API、base_url

- gemini cli
- claude code
- codex
- opencode

### aider
aider是一个结对辅助编程助手，
/help question

- Repository map

用tree-sitter生成AST
This repository map is now built automatically using tree-sitter to extract symbol definitions from source files. Tree-sitter is used by many IDEs, editors and LSP servers to help humans search and navigate large codebases. Aider now uses it to help GPT better comprehend, navigate and edit code in larger repos.

生成代码地图，只给大模型相关的代码
Of course, for large repositories even just the repo map might be too large for GPT’s context window. Aider solves this problem by sending just the most relevant portions of the repo map. It does this by analyzing the full repo map using a graph ranking algorithm, computed on a graph where each source file is a node and edges connect files which have dependencies. Aider optimizes the repo map by selecting the most important parts of the codebase which will fit into the token budget assigned by the user (via the --map-tokens switch, which defaults to 1k tokens).

可以预填充，达到模型上下文限制时，从之前的状态接着输出
When aider is collecting code edits from a model and it hits the output token limit, aider simply initiates another LLM request with the partial response prefilled. This prompts the model to continue where it left off, generating more of the desired response. This prefilling of the partially completed response can be repeated, allowing for very long outputs. Joining the text across these output limit boundaries requires some heuristics, but is typically fairly reliable.

# local model
- lm studio
- ollma
- llma.cpp
## local model remote use
$ irm https://lmstudio.ai/install.ps1 | iex

![](/media/202608/2026-08-20_144117_1826760.6166835535713534.png)

![](/media/202608/2026-08-20_144335_4287660.5897323326195637.png)

$ pi install npm:pi-lmstudio

# 大模型使用技巧
AI的能力边界在哪？

## skills
目的：结论客观性
展示信息：
- 信源
- 思考过程/思维模型
- 复盘/事实修正

目的：AI思考透明化
思维链
(思维链条可能有事后编造的嫌疑)

## 原理层面
- context length