BIBLIOGRAPHY

darjeeling. 2025. “Llm을 활용한 코드 작성시 유의할 14가지.” GeekNews. March 12, 2025. https://news.hada.io/topic?id=19705.

Simon Willison. 2025. “Simon Willison’s Newsletter | Substack.” March 26, 2025. https://simonw.substack.com/.

———. (2023) 2025. “Simonw/Llm.” https://github.com/simonw/llm.

———. n.d.-a. “Here’s How I Use Llms to Help Me Write Code - Darjeeling.” Simon Willison’s Weblog. Accessed March 26, 2025. https://simonwillison.net/2025/Mar/11/using-llms-for-code/.

———. n.d.-b. “Llm: A Cli Utility and Python Library for Interacting with Large Language Models.” Accessed March 26, 2025. https://llm.datasette.io/en/stable/index.html.

Willison, Simon. n.d.-a. “Github Issues for Notes 깃허브 활용한 노트 관리.” Simon Willison’s Weblog. Accessed June 2, 2025. https://simonwillison.net/2025/May/26/notes/.

———. n.d.-b. “Llm, Ttok and Strip-Tags—Cli Tools for Working with Chatgpt and Other Llms.” Simon Willison’s Weblog. Accessed March 27, 2025. https://simonwillison.net/2023/May/18/cli-tools-for-llms/.

xguru. 2025. “Simon Willison의 O3-Mini에 대한 노트.” GeekNews. February 2, 2025. https://news.hada.io/topic?id=19018.

관련메타

History

  • [2025-06-04 Wed 19:41] 좋은 글들이 많아.
  • [2025-03-27 Thu 07:19]

Simon Willison’s Newsletter | Substack

(Simon Willison 2025)

  • Simon Willison
  • AI, LLMs, web engineering, open source, data science, Datasette, SQLite, Python and more. Click to read Simon Willison’s Newsletter, a Substack publication with tens of thousands of subscribers.
  • 2025

아티글

LLM을 활용한 코드 작성시 유의할 14가지

(darjeeling 2025)

  • darjeeling
    1. 합리적인 기대치 설정 (Set reasonable expectations)주요 내용:LLM은 본질적으로 고급 자동완성 도구임을 인지해야 하며, 전적으로 의존할 경우 오류나 부정확한 결과가 발생할 수 있음.활용법:자신의 코딩 능력을 보완하는 보조 도구로 활용하며, 과도한 기대를 버리고 실수를 검증하는 태도가 필요함.2. 학습 데이터의 종료 시점을 고려

Here’s how I use LLMs to help me write code - darjeeling

(Simon Willison n.d.-a)

  • Willison, Simon
  • Online discussions about using Large Language Models to help write code inevitably produce comments from developers who’s experiences have been disappointing. They often ask what they’re doing wrong—how come some …

GitHub issues for notes 깃허브 활용한 노트 관리

(Willison n.d.-a) Willison, Simon

GitHub issues is almost the best notebook in the world. Free and unlimited, for both public and private notes. Comprehensive Markdown support, including syntax highlighting for almost any language. Plus …

Simon Willison의 o3-mini에 대한 노트

(xguru 2025)

  • xguru

  • OpenAI의 o3-mini 모델이 새로 공개되었고, 다른 o 시리즈 모델과 마찬가지로 평가하기가 약간 어려움이제 기존 GPT-4o, o1, o1 Pro 등에 이어 사용 목적에 따라 어떤 모델을 선택할지 결정해야 하는 상황o3-mini System Card (PDF)에 따르면 일부 평가 지표에서 o3-mini가 GPT-4o, o1보다 우수하지만, 모든 항목

  • 2025

LLM: A CLI utility and Python library for interacting with Large Language Models

(Simon Willison n.d.-b)

  • Simon Willison

llm, ttok and strip-tags—CLI tools for working with ChatGPT and other LLMs

(Willison n.d.-b)

  • Willison, Simon

  • I’ve been building out a small suite of command-line tools for working with ChatGPT, GPT-4 and potentially other language models in the future. The three tools I’ve built so far …

리포

[2025-03-27 Thu 10:53]

훌륭하다.

>

simonw/llm

(Simon Willison [2023] 2025)

  • Simon Willison
  • Access large language models from the command-line
  • 2025

LLM 프로젝트 개요

LLM은 Simon Willison이 개발한 커맨드라인 도구이자 Python 라이브러리로, 다양한 대형 언어 모델과 상호작용할 수 있게 해줍니다.

주요 특징
  • 이중 인터페이스: CLI(커맨드라인) + Python API
  • 다중 모델 지원: OpenAI GPT, Anthropic Claude, Google Gemini, Meta Llama 등
  • 로컬 모델 지원: Ollama 플러그인을 통한 로컬 LLM 실행
  • SQLite 로깅: 모든 프롬프트와 응답을 SQLite DB에 자동 저장
  • 임베딩 생성: 텍스트 임베딩 생성 및 저장 기능
  • 멀티모달 지원: 이미지, 오디오, 비디오 처리
  • 플러그인 시스템: 확장 가능한 아키텍처
설치 방법
# pipx를 이용한 설치 (권장)
pipx install llm
 
# Homebrew
brew install llm
 
# pip
pip install llm
기본 사용법
# 간단한 프롬프트 실행
llm "Explain how transformers work"
 
# 특정 모델 지정
llm -m claude-3-5-sonnet "질문"
llm -m gpt-4o "질문"
 
# 대화형 세션
llm chat -m claude-3-5-sonnet
 
# 로컬 모델 사용 (Ollama)
llm install llm-ollama
llm -m llama3.2 "질문"
Python API 사용
import llm
 
# 모델 가져오기
model = llm.get_model("gpt-4o-mini")
response = model.prompt("Your prompt here")
print(response.text())
 
# 대화 세션
conversation = model.conversation()
conversation.prompt("첫 번째 질문")
conversation.prompt("후속 질문")
 
# 비동기 지원
async_model = llm.get_async_model("claude-3-5-sonnet")
 
# 임베딩 생성
embedding = model.embed("텍스트 내용")
활용 사례
  1. 로컬 LLM 통합: Ollama와 연동하여 로컬 GPU에서 모델 실행
  2. 배치 처리: 대량의 텍스트 데이터 자동 처리
  3. RAG 시스템: 임베딩 생성 및 시맨틱 검색
  4. 워크플로우 자동화: n8n 등의 도구와 Python API 통합
  5. 비용 최적화: 로컬 모델과 API 모델을 상황에 따라 선택적 사용
플러그인 개발
import llm
from llm import hookimpl
 
@hookimpl
def register_models(register):
    register(CustomModel())
 
class CustomModel(llm.Model):
    def execute(self, prompt, **kwargs):
        # 커스텀 로직 구현
        return response
참고 링크

로그

|2025-03-27 Thu 07:19|

@user Simon Willison에 대해서 알려줘 Datasette를 만든 사람이며 아래 블로그를 운영하고 있어.

https://simonwillison.net/

@assistant Simon Willison은 Datasette를 만든 개발자로, 데이터베이스를 웹으로 쉽게 공유할 수 있게 해주는 도구입니다. 그는 또한 블로그 https://simonwillison.net/%EB%A5%BC 운영하며, 기술 관련 글과 프로젝트를 공유합니다.

@user 그의 대표적인 글들을 소개해줘.

@assistant Simon Willison의 블로그에서 그의 대표적인 글을 소개하면 다음과 같습니다:

  1. “What to blog about” (2022년 11월 6일): 블로그를 시작할 때 무엇을 작성해야 할지 고민하는 사람들에게 TIL(오늘 배운 것)과 프로젝트 설명을 추천합니다. 이 글은 블로그 작성의 압박을 줄이고, 배우고 만든 것을 공유하는 방법에 대해 이야기합니다[3].

  2. TIL 시리즈: Simon Willison은 2020년 4월부터 TIL을 시작하여 현재 346개 이상의 글을 작성했습니다. 이 시리즈는 그가 배운 것을 간단하게 공유하는 형식으로, 대부분 10분 이내에 작성되었습니다[3].

  3. Datasette와 관련된 글: 그는 Datasette와 관련된 다양한 프로젝트와 기술에 대한 글도 많이 작성하고 있습니다. Datasette는 데이터베이스를 쉽게 웹으로 공유할 수 있게 해주는 도구입니다[4][5].

  4. Citations: [1] https://en.wikipedia.org/wiki/Simon_Willison [2] https://github.com/simonw/simonwillisonblog [3] https://simonwillison.net/2022/Nov/6/what-to-blog-about/ [4] https://github.com/simonw [5] https://substack.com/@simonw

@user