NEW AGENTLESS AI Software Development
Vložit
- čas přidán 23. 08. 2024
- A new leader for open-source software development: AGENTLESS. Best in class according to performance metrics on the SWE-bench Lite benchmark, including other open source agents w/ GPT-4 omni and Claude 3.5, like Aider or Devon or SWE-Agent.
AGENTLESS challenges the prevailing notion that complex autonomous agents are necessary for automating software development tasks. It leverages a simplified two-phase process-localization and repair-without the need for agents to decide on future actions or manage complex tools. This simplicity not only reduces the cognitive load involved in understanding and debugging the process but also significantly lowers operational costs. The empirical results demonstrate that AGENTLESS achieves competitive performance metrics on the SWE-bench Lite benchmark, outperforming all other open-source approaches in both effectiveness and efficiency.
The insights from AGENTLESS's performance suggest a paradigm shift in the development of software engineering tools, emphasizing the effectiveness of simpler, more interpretable methods over complex autonomous systems. This approach not only makes it easier to understand and maintain the system but also highlights the potential for significant cost savings and efficiency improvements. The research encourages further exploration into refining these simplistic approaches, suggesting that future advancements could focus on enhancing the accuracy of localization and repair mechanisms and exploring new forms of integration with existing development environments.
All rights w/ authors:
AGENTLESS :
Demystifying LLM-based Software Engineering Agents
arxiv.org/pdf/...
#airesearch
#newtech
#science
In software engineering class at UCLA (1977) we submitted "PL/C" jobs on punch cards. Intended for teaching, PL/C implemented IBM's PL/1 but "fixed" syntax and a few logic errors for you. Any error became many pages of errors created by PL/C's attempted fixes. So we're nearly back to 50 years ago. 😀
Thanks for this insight. It proves, that we are trying everything we know to make the machines a little bit more ..... clever ? For sure it is not intelligent.
Amazing stuff! Keep it coming.
The fact that companies are selling something that only works 20% of the time is crazy to me
State of the art. Of course only in scientific terms, because if you look at some marketing statement: 104% and more .... smile.
If it saves 5% or more effort in a team of 20+ devs, that's already more than 1 FTE so it could be useful.
Praise the Lord ☀️ for these sweet drops 🍭 of youtube manna 😭😭😭
when less is more... the we need more recommendations with less... errr.... passt scho^^
😎🤖
Agentless? To me it seems to operate in agentic mode, with multiple steps and LLM requests
Ah, I understand. If you want to learn about agentic, and why this here is AGENTLESS; I have a dedicated video on what it means to be AGENTIC: czcams.com/video/yJHdxYJgYhg/video.html