Neutral mention · rule
This has been proposed/exlored in 2023 already: ChatDev: Communicative Agents for Software Development - https://arxiv.org/abs/2307.07924
Back to the collection
ChatDev: Communicative Agents for Software Development
Chen Qian, Wei Liu, Hongzhang Liu
Show all 14 authors
Nuo Chen, Yufan Dang, Jiahao Li, Cheng Yang, Weize Chen, Yusheng Su, Xin Cong, Juyuan Xu, Dahai Li, Zhiyuan Liu, Maosong Sun
2023 · arXiv
- HN score
- 0
- Mentions
- 8
- Readers
- 4
- Months
- 4
From the abstract
Software development is a complex task that necessitates cooperation among multiple members with diverse skills. Numerous studies used deep learning to improve specific phases in a waterfall model, such as design, coding, and testing. However, the deep learning model in each phase requires unique designs, leading to technical inconsistencies across various phases, which results in a fragmented and ineffective development process. In this paper, we introduce ChatDev, a chat-powered software development framework in which specialized agents driven by large language models (LLMs) are guided in what to communicate (via chat chain) and how to communicate (via communicative dehallucination). These agents actively contribute to the design, coding, and testing phases through unified language-based communication, with solutions derived from their multi-turn dialogues. We found their utilization of natural language is advantageous for system design, and communicating in programming language proves helpful in debugging. This paradigm demonstrates how linguistic communication facilitates multi-agent collaboration, establishing language as a unifying bridge for autonomous task-solving among LLM agents. The code and data are available at https://github.com/OpenBMB/ChatDev.
Metadata: arXiv. Abstracts retain their original authors’ rights.
In the conversation
Original HN comments, with labels from simple text rules. A label describes the whole comment and may miss context or irony. Read the discussion to make up your own mind. About the labels
8 selected excerpts · 8 total mentions
Neutral mention · rule
>The ChatGPT "AI" isn't an autonomous thinker performing its own actions based on reasoning of what's fed to it. Yes it is. Or it very well could be. Agency is trivial to implement in LLMs. https://github.com/microsoft/autogen https://arxiv.org/abs/2307.07924 The intelligence and tool access of the LLM in question is the only thing stopping things from being particularly dangerous (to humanity).
Neutral mention · rule
Not to discount your experience, I would just like to flag that your conclusion does not quite follow from the premises. I don't believe that there's anything inherent about LLMs that would make one incapable of recognizing the falsity of an argument made by another LLM. Rather it's just the they're not all powerful, and struggle with complex tasks when they don't have any way of directly checking their code's outputs (or error messages), same as humans. There is a lot of recent progress in collaboration between LLMs when they're given the ability to run their code and assess their progress [0]. [0] E.g. the ChatDev project - https://arxiv.org/abs/2307.07924 , described in this recent video by "Two Minute Papers"[1] [1] https://youtu.be/Zlgkzjndpak
Neutral mention · rule
I mean, it's just a clickbait video. The research paper [0] is cool and all, I just wish we didn't have these hyperbolic clickbait titles like "OpenAI’s ChatGPT Makes A Game For $1!" - we really are in a bubble. [0] https://arxiv.org/abs/2307.07924
Neutral mention · rule
https://arxiv.org/abs/2307.07924 at the very least, you can get things done that would be extremely difficult or practically impossible to do with a single instance.
Neutral mention · rule
AI taking to AI to do things is not necessarily an exercise in futility. This paper used 3.5, not even 4 and is basically a couple AI agents collaborating to create software. https://arxiv.org/abs/2307.07924 This is another example at multi AI agent collaboration. https://arxiv.org/abs/2307.02485 Getting relevant intelligent text to arbitrary previous text means generating text capable of making decisions. That is the end game here. Your machine can think for you now.
Neutral mention · rule
It's more than just a react loop but i think there's promising research in multi gent cooperation. Notice the first doesn't even use GPT-4. https://arxiv.org/abs/2307.07924 https://arxiv.org/abs/2307.02485
Neutral mention · rule
>The lack of effectiveness and usefulness of these agents has been shown time and over. Has it ? https://arxiv.org/abs/2307.07924 https://arxiv.org/abs/2307.02485 The first one doesn't even use GPT-4
Keep following the thread.
More in Machine learningMore from the same topics.
2023arXiv
Sparks of Artificial General Intelligence: Early experiments with GPT-4
Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan +11
Artificial intelligence (AI) researchers have been developing and refining large language models (LLMs) that exhibit remarkable capabilities across a variety of domains and tasks, challenging our understanding of learning and cognition.
60 mentions14 months
3.5HN score
2020arXiv
Language Models are Few-Shot Learners
Tom B. Brown, Benjamin Mann, Nick Ryder +28
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task.
83 mentions35 months
3HN score
2016arXiv
Understanding deep learning requires rethinking generalization
Chiyuan Zhang, Samy Bengio, Moritz Hardt +2
Despite their massive size, successful deep artificial neural networks can exhibit a remarkably small difference between training and test performance.
16 mentions14 months
2.5HN score
2016arXiv
End to End Learning for Self-Driving Cars
Mariusz Bojarski, Davide Del Testa, Daniel Dworakowski +10
We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands.
8 mentions7 months
2.5HN score