#autogpt — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #autogpt, aggregated by home.social.
-
OpenClaw 解体新書｜AIに実行権限を渡すと何が起きるのか https://www.yayafa.com/2773388/ #AgentAI #AgenticAi #AI #AIArchitecture #AiAutomation #AIOS #AiTools #AIエージェント #AIセキュリティ #AI実行 #AI最新 #AI構造 #AI権限 #AI解説 #AI開発 #ArtificialGeneralIntelligence #ArtificialIntelligence #AutoGPT #LLM #Openclaw #エージェント型AI #ソフトウェアエンジニア #プロンプトインジェクション #人工知能 #汎用人工知能
-
OpenClaw 解体新書｜AIに実行権限を渡すと何が起きるのか https://www.yayafa.com/2773388/ #AgentAI #AgenticAi #AI #AIArchitecture #AiAutomation #AIOS #AiTools #AIエージェント #AIセキュリティ #AI実行 #AI最新 #AI構造 #AI権限 #AI解説 #AI開発 #ArtificialGeneralIntelligence #ArtificialIntelligence #AutoGPT #LLM #Openclaw #エージェント型AI #ソフトウェアエンジニア #プロンプトインジェクション #人工知能 #汎用人工知能
-
🚨 CRITICAL: CVE-2026-26020 in AutoGPT (<0.6.48) allows authenticated RCE via improper authorization of BlockInstallationBlock. Patch to 0.6.48+ ASAP! Monitor for suspicious graph configs. https://radar.offseq.com/threat/cve-2026-26020-cwe-285-improper-authorization-in-s-31ca744c #OffSeq #AutoGPT #Infosec #Vulnerability
-
AutoGPT × OODAループ 続編：自己改善型AgentOpsとKPI連携の構築を考える
https://qiita.com/ota-tsutomu/items/40d69131e366416b4ed2?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items -
AutoGPT × OODAループ【実践＆応用編】思考と行動をエージェント化する仕組みを考えてみる
https://qiita.com/ota-tsutomu/items/dae05eb6eebafe049bff?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items -
🔧 Gli agenti AI stanno imparando a usare il Mac meglio di noi.
Dalla shell Unix al controllo dell’interfaccia grafica, oggi è possibile automatizzare flussi complessi usando linguaggio naturale e modelli LLM.
Ho scritto su ciberneticagerber.it un articolo per sviluppatori che analizza:
Una panoramica dei principali agenti AI capaci di pilotare macOS via CLI o GUI.
Il confronto tra soluzioni open source (Auto-GPT, Open Interpreter) e commerciali (Raycast AI).
Un approfondimento su automazione shell, interazione con API e flussi multi-step intelligenti.
📎 Se sviluppi su Mac e vuoi capire come delegare task alla tua prossima intelligenza artificiale personale, questa lettura fa per te.
👉 Leggi l’articolo completo:
https://ciberneticagerber.it/2025/05/03/agenti-ai-per-pilotare-macos-panoramica-di-strumenti-e-casi-duso/
#AIagents #macOS #automation #developers #LLM #CLI #OpenSource #Raycast #AutoGPT -
🔧 Gli agenti AI stanno imparando a usare il Mac meglio di noi.
Dalla shell Unix al controllo dell’interfaccia grafica, oggi è possibile automatizzare flussi complessi usando linguaggio naturale e modelli LLM.
Ho scritto su ciberneticagerber.it un articolo per sviluppatori che analizza:
Una panoramica dei principali agenti AI capaci di pilotare macOS via CLI o GUI.
Il confronto tra soluzioni open source (Auto-GPT, Open Interpreter) e commerciali (Raycast AI).
Un approfondimento su automazione shell, interazione con API e flussi multi-step intelligenti.
📎 Se sviluppi su Mac e vuoi capire come delegare task alla tua prossima intelligenza artificiale personale, questa lettura fa per te.
👉 Leggi l’articolo completo:
https://ciberneticagerber.it/2025/05/03/agenti-ai-per-pilotare-macos-panoramica-di-strumenti-e-casi-duso/
#AIagents #macOS #automation #developers #LLM #CLI #OpenSource #Raycast #AutoGPT -
Replit Agent создает программы за вас или кодинг без IDE
Многие слышали про AutoGPT и GPT Engineer — агентные системы, которые позволяют генерировать код по промпту от пользователя. Меня зовут Евгений Кокуйкин. Я руководитель AI продуктов компании Raft. Сегодня расскажу про Replit Agent — AI Copilot для написания кода без знаний в программировании. Без шуток! Раньше у меня на такие прототипы уходили часы кодинга и отладки. А сейчас я сгенерировал код приложения через агент — быстрее, чем писал эту статью. Так что теперь можно участвовать в хакатоне, не зная Python. Replit — это онлайн-IDE, где с помощью AI можно быстро создать прототип приложения и задеплоить его прямо в облаке. Недавно вышла экспериментальная фича Replit Agent, которая стоит 25 баксов. Правда, для оплаты потребуется зарубежная карта, но это уже каждый сам решает, стоит заморачиваться или нет. Эта фича позволяет начать генерацию проекта одним промптом. Я так вдохновился постом Степана Гершуни, что тестировал кодинг-агента в Replit , а потом не спал всю ночь, записывая впечатления. Так появился этот обзор.
https://habr.com/ru/companies/oleg-bunin/articles/841808/
#разработка_приложений #nocode #ии #ии_и_машинное_обучение #autogpt #GPT_Engineer #Replit_Agent #AI_Copilot #онлайнIDE #отладка_кода
-
Autonome KI-Agenten: Smallville-Experiment ist jetzt Open Source
In Smallville leben 25 KI-Agenten ein simuliertes Kleinstadtleben. Dieses Experiment ist nun Open Source verfügbar.
#AutoGPT #AutonomeAgenten #KünstlicheIntelligenz #Spracherkennung
-
-
#babyagi #autogpt and the like try to mix in some event loops and ways to interact with the world, edging into more ingredients of Agency. Mixing in rule databases, logic engines, or things like this tree-of-thought (https://arxiv.org/abs/2305.10601) take further steps.
-
Random thought of the night. Does anyone have any practical examples of utilizing tools that enhance ChatGPT?
I have been dabbling in AutoGPT and Langchain. While I see the future possibilities, so far it's just been a time sink vs GPT 4
-
Interestingly in only a few hours, the discussion progressively moved from how to choose an agency that could study the initial question, to health and fitness ❤️🩹, and progressively to discussing tastes about... Pizzas.
Quite fascinating, and shows the huge potential of autonomous AI agents.
Where will that lead to?
-
@bhaveshpant probably we'll see #AutoGPT alternatives that work with #OpenSource #LLM in the near future. Some open source models are listed here: https://anania.ai/chatgpt-alternatives/
-
My key takings about autonomous agents, and why they are the next big thing after the hype of #Chatgpt https://alecrimi.substack.com/p/what-are-autonomous-agents-and-why #AutoGPT #AgentGPT #MachineLearning #AI
-
My article on "Truthiness by Design" — building a #Legal #LLM as a #FoundationalModel, to provide:
1. Truth / Facts
2. Guardrails for autonomous agents (e.g., #AutoGPT, #BabyAGI)
https://www.linkedin.com/pulse/legal-llm-foundational-models-truthiness-design-damien-riehl
-
Read the blog post on AI that GPT-4 calls "...[a] captivating journey into the future of thinking machines!"
https://jordanwages.com/on-stochastic-parrots-large-language-models-and-where-were-heading.html
-
Anwendungsbeispiele, die das Potenzial von #AutoGPT erahnen lassen... ⬇️
https://the-decoder.de/auto-gpt-die-besten-anwendungsfaelle/ -
Follow to the last post:
AgentGPT
"Assemble, configure, and deploy autonomous AI Agents in your browser."
-
Asking Auto-GPT to create and run a botnet is like asking a Paladin to help you rob that tavern. Sure, you can trick them into it, but every other sentence out of their mouth is about how dishonorable it is.
-
Wow. #AutoGPT is my new best friend. Buggy AF! But I'll have it fix its own code.
-
Auto-GPT
"An experimental open-source attempt to make GPT-4 fully autonomous."
https://github.com/Significant-Gravitas/Auto-GPT
In the Github repository is a 1min video demo showing access to Google search and file storage.
-
@Vorinstanz hat nicht Mutter Theresa gesagt:
Gerade wenn du eine strenge Woche hast - vergiss nicht, #AutoGPT zu installieren.
Irgendwie so ging das.
-
Werde heute wohl versuchen, #AutoGPT auf meinem Mac zu installieren.
Evtl ein Meilenstein.Kann sich das noch nicht selbst installieren? 😀
-
"[T]he #AutoGPT GitHub cautions against using 'continuous mode' as it 'it is potentially dangerous & may cause your AI to run forever or carry out actions you would not usually authorize'" https://www.vice.com/en_us/article/epvdme/developers-are-connecting-multiple-ai-agents-to-make-more-autonomous-ai
