Open louiseliu opened 1 year ago
Thank you for your suggestion. AutoGPT indeed introduces a valuable concept, but in most cases, it may not provide practical solutions. We aim for Jarvis to solve users' problems effectively and be genuinely useful in daily life. We are keeping an eye on AutoGPT's development. If it reaches a production-ready state, we would be delighted to consider its implementation or even integrate it directly. However, a significant drawback is that AutoGPT requires the use of GPT4, which is prohibitively expensive. Conversely, Jarvis achieves satisfactory results with GPT3.5, which is more cost-efficient. In conclusion, our goal is to ensure that Jarvis is truly practical and usable.
After using it, I feel that Jarvis is an assistant that is ready to serve you at any time. At least in terms of user experience, it is convenient to adjust the strategy with AI through natural language interaction. Both Jarvis and OpenDAN are actually solving user problems, but OpenDAN hasn't fully maximized its capabilities yet. I'm looking forward to future version updates.
During my use of Jarvis, I found that it only executes one task at a time when planning tasks. However, we know that AutoGPT has the ability to break down complex issues into a group of simpler, smaller ones, and it can automatically orchestrate the execution of tasks. Therefore, I'm wondering, why not directly use AutoGPT as the task planning module for Jarvis, to enhance its ability to handle multiple tasks?