Romain Huet (@romainhuet)

트윗은 @steipete가 Codex를 활용해 공개적으로 개발한 사례를 칭찬하며, moltbot을 통해 많은 사람에게 도달한 점을 감탄합니다. 작성자는 Codex를 만든 이유와 오픈 생태계 지원의 중요성을 강조하며, 개발자 중심의 개방형 AI 도구 및 사례 확산을 지지하는 메시지를 전달합니다.

https://x.com/romainhuet/status/2016314931566805182

#codex #ai #developer #openecosystem #moltbot

Oleg Akbarov (@olegakbarov)

OpenAIDevs의 Codex 관련 피드백: 서브에이전트(subagents)들의 진행 상태(프로그램 진행 인디케이터)가 부족하다고 지적하며, 토큰 한도에 걸려 에이전트가 멈춘 듯해 사용자(작성자)가 직접 'how are we doing buddy'라며 에이전트를 깨워야 했다는 경험을 공유. 개발자용 에이전트 UX·모니터링 문제 제기.

https://x.com/olegakbarov/status/2016364422290407751

#openai #codex #agents #devtools

Mukund Jha (@mukundjha)

샘 알트먼(@sama)의 타운홀에서 더 빠른 Codex가 곧 출시된다고 발표했습니다. 코드 생성 모델 Codex의 속도 개선(퍼포먼스 향상)이 예고된 소식으로, 개발자용 코드 생성·자동화 워크플로우의 체감 성능과 생산성에 직접적 영향을 줄 가능성이 있습니다.

https://x.com/mukundjha/status/2015961243153264977

#openai #codex #codegeneration #developertools

Runko Shownotes-Editor Vibeslop Alpha #sweetsolution #pwa #codex #quip

"Codex CLI is our cross-platform local software agent, designed to produce high-quality, reliable software changes while operating safely and efficiently on your machine. We’ve learned a tremendous amount about how to build a world-class software agent since we first launched the CLI in April⁠. To unpack those insights, this is the first post in an ongoing series where we’ll explore various aspects of how Codex works, as well as hard-earned lessons. (For an even more granular view on how the Codex CLI is built, check out our open source repository at https://github.com/openai/codex. Many of the finer details of our design decisions are memorialized in GitHub issues and pull requests if you’d like to learn more.)

To kick off, we’ll focus on the agent loop, which is the core logic in Codex CLI that is responsible for orchestrating the interaction between the user, the model, and the tools the model invokes to perform meaningful software work. We hope this post gives you a good view into the role our agent (or “harness”) plays in making use of an LLM.

Before we dive in, a quick note on terminology: at OpenAI, “Codex” encompasses a suite of software agent offerings, including Codex CLI, Codex Cloud, and the Codex VS Code extension. This post focuses on the Codex harness, which provides the core agent loop and execution logic that underlies all Codex experiences and is surfaced through the Codex CLI. For ease here, we’ll use the terms “Codex” and “Codex CLI” interchangeably."

https://openai.com/index/unrolling-the-codex-agent-loop/

#AI #GenerativeAI #LLMs #OpenAI #CodexCLI #Codex #AIAgents #AgenticAI #Programming #SoftwareDevelopment

How hard can it be to build a browser from scratch for three platforms anyways?

Apparently 20K lines of code and ~70 hours from first commit to last.

https://emsh.cat/one-human-one-agent-one-browser/

#llm #llms #ai #codex #openai

↳ continue
↳ continue
↳ continue
↳ continue

#codex #CodexCLI

🤖 OpenAI apre le porte al futuro, svelando i dettagli tecnici dell'agente AI Codex. Un passo avanti verso l'#IntelligenzaArtificiale. #OpenAI #Codex 💡

🔗 https://www.tomshw.it/hardware/openai-svela-come-funziona-il-suo-agente-ai-2026-01-27

Two interesting points from #OpenAI’s AI agent explainer:

⁃ #Codex’s compaction uses data from the #LLM’s internal latent representation of the conversation rather than a text summary. Considering the computational expense of reprocessing tokens, this must a big efficiency win.
⁃ The client gets this data encrypted, with the keys held by OpenAI. This satisfies #ZDR compliance, but I suspect also blocks model internals from reversing.

https://openai.com/index/unrolling-the-codex-agent-loop/

OpenAI spills technical details about how its AI coding agent works https://arstechni.ca/p5Qg #largelanguagemodels #AIdevelopmenttools #machinelearning #DeveloperTools #promptcaching #Programming #codeagents #agenticAI #AIagents #AIcoding #Biz&IT #openai #Codex #API #AI

Trong bối cảnh AI như Copilot/Cursor tăng sản lượng code, liệu các Leader Engineering cần công cụ AI giám sát quy trình review?
- Có nên đầu tư công cụ đánh giá chất lượng review, phát hiện lỗi tiềm ẩn hay chỉ là "ảo tưởng"?
- Làm sao cân bằng giữa tốc độ phát triển và an toàn mã nguồn?
Các bạn nghĩ đây là vấn đề thực tế hay chỉ là "đau riêng"?
#AI #Codex #Engineering #ĐảmBảoCode #CôngNghiệpPhầnMềm

**Tag tiếng Việt**: #KhaiThácAI #QuảnLýCode #HiệuSuấtLàmViệc #PhầnMềm #CNPM
**Tag tiế

Which one is better to start with, Codex IDE or Claude Code? #codex

Docker Sandboxes: Run Claude Code and Other Coding Agents Unsupervised (but Safely)
#Docker #Engineering #Products #AIML #Claude #Codex #DockerSandboxes #Gemini

https://www.docker.com/blog/docker-sandboxes-run-claude-code-and-other-coding-agents-unsupervised-but-safely/

TestingCatalog News (@testingcatalog)

Anthropic가 Claude Code용 'Security Center'(구 AutoPatch)를 출시 준비 중입니다. 사용자는 과거 스캔 기록과 감지된 이슈를 조회하고 수동으로 새 스캔을 실행할 수 있게 됩니다. 이 발표는 다가오는 Codex 업그레이드와 관련한 사이버보안 대응의 직접적 대응책이 될지 주목됩니다.

https://x.com/testingcatalog/status/2015454911224582611

#anthropic #claudecode #securitycenter #autopatch #codex

Ollama Launch 출시, AI 코딩 도구와 로컬 모델 연결을 한 줄로

https://aisparkup.com/posts/8664

Alexander Embiricos (@embirico)

Responses API(코덱스가 사용하는)에서 높은 오류율이 발견되어 수정되었으며 현재는 안정화되었을 것으로 보인다는 공지입니다. 문제가 계속 발생하면 회신해 달라고 요청하고 있어 개발자·서비스 운영자에게 중요한 가용성(안정성) 업데이트입니다.

https://x.com/embirico/status/2015206665264726242

#responsesapi #codex #api #reliability #openai

Alexander Embiricos (@embirico)

Responses API(또는 Codex가 사용하는 API)에서 높은 오류율이 발견되어 이를 수정했다고 공지함. 현재는 안정화되었으니 여전히 문제가 발생하면 회신해 달라는 안내입니다.

https://x.com/embirico/status/2015206665264726242

#codex #responsesapi #api #stability

[OpenAI Codex CLI 내부 동작 분석: 에이전트 루프와 프롬프트 캐싱 전략

OpenAI의 Codex CLI 내부 동작 분석 아티클이 공개되었습니다. 에이전트 루프(Agent Loop)의 구조, 프롬프트 구성 및 Responses API의 데이터 흐름, 성능 최적화 전략(프롬프트 캐싱, 대화 압축, 무상태 설계) 등을 심층적으로 다룹니다.

https://news.hada.io/topic?id=26089

#openai #codex #agentloop #promptcaching #zdr

@mjd “Was the alphabetic index invented before the scroll was superseded by the codex?” — Not really. Plinius' “Natural History“ has an index and he tells in his preface about another work with an index, but that seem to be the only cases. The problem is that with a scroll, you cannot actually use the index, because “leafing” through the pages (or better, “scrolling”?) would be too much effort.
It is also said that the introduction of Christianity was the reason why in later times, codices were much more popular than scrolls. The problem was the Bible: To work as a theologian, you have jump backward and forward between gospel and prophet books, between different gospels, and so on — much too much work if you have scrolls. The Bible was the first hypertext!

For more about indices, see “Index, A History of the” by David Duncan (https://en.wikipedia.org/wiki/Index,_A_History_of_the).

#Index #Scroll #Codex #Hypertext

OpenAI stuft die kommenden Codex-Modelle im Preparedness Framework als "Cybersecurity High" ein. Damit besitzen die Systeme Fähigkeiten, die signifikante Vorteile bei der Entwicklung von Exploits bieten. Vor der Veröffentlichung im Februar sind eine explizite Freigabe durch das Board und Red Teaming notwendig. Die Abgrenzung zum reinen Hacking-Werkzeug schwindet. #OpenAI #Codex #Cybersecurity
https://www.all-ai.de/news/news26/codex-sicherheit-high