BabyAGI
In plain wordsA tiny 2023 program that kept an AI working through an ever-changing to-do list — doing a task, adding new ones, reprioritizing — and helped start the craze for autonomous AI agents.
A minimal open-source 'task-driven autonomous agent', initially a single short Python script, that ran a large language modellarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code. in a loop over a task list: an execution step completed the top task toward a stated objective, a task-creation step generated new tasks from the result, and a prioritization step reordered the queue, with results stored in a vector database (initially Pinecone) as memory. Its simplicity made the core of an agent loopagent loop, n.: The repeating cycle at the heart of an AI agent: the model decides what to do next, software carries it out, the result is shown to the model, and the cycle repeats until the task is finished. easy to read and copy, and with AutoGPTAutoGPT, prop. n.: A 2023 open-source program that gave GPT-4 a goal and let it run on its own — planning, searching the web and writing files in a loop — becoming the first viral 'autonomous AI agent'. it became an emblem of the spring 2023 agent boom; like AutoGPT, it tended to drift, loop and spawn endless subtasks. The name plays on AGIartificial general intelligence, n.: AI that could learn and do almost any mental task a capable adult can, rather than excelling at one narrow job. Experts disagree on the exact bar and on when, or whether, it will arrive.; later versions experimented with agents that write and store their own functions.
BabyAGI's three-prompt loop is still the clearest teaching example of a task-driven agent.
Origin
Created by venture capitalist Yohei Nakajima, who described the design as a 'task-driven autonomous agent' on social media in late March 2023 and open-sourced it as BabyAGI in early April 2023, days after AutoGPT.
Related
- AutoGPTAutoGPT, prop. n.: A 2023 open-source program that gave GPT-4 a goal and let it run on its own — planning, searching the web and writing files in a loop — becoming the first viral 'autonomous AI agent'.
- AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done.
- agent loopagent loop, n.: The repeating cycle at the heart of an AI agent: the model decides what to do next, software carries it out, the result is shown to the model, and the cycle repeats until the task is finished.
- task decompositiontask decomposition, n.: Breaking a big, vague job into smaller, concrete steps that can each be done and checked — the way a recipe splits a meal into stages.
- agent memoryagent memory, n.: The ways an AI assistant or agent keeps information — your preferences, past conversations, lessons from earlier tasks — and brings it back later, beyond what fits in a single conversation.
Sources
- Nakajima, Y. (2023). BabyAGI. GitHub repository yoheinakajima/babyagi.
- Nakajima, Y. (2023). Birth of BabyAGI. Blog post, 1 April 2023. yoheinakajima.com/birth-of-babyagi
Mentioned in
AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done., AutoGPTAutoGPT, prop. n.: A 2023 open-source program that gave GPT-4 a goal and let it run on its own — planning, searching the web and writing files in a loop — becoming the first viral 'autonomous AI agent'.