What do people think of Sweep AI?
The community submitted 14 reviews to tell
us what they like about Sweep AI, what Sweep AI can do better, and
more.
What do you think about Sweep AI?
Leave a rating or review for the community
2.4/5All time (14 reviews)
4/5
Recently (1 review)14 Reviews
useless, worse than copilot
William Zeng@william_zeng1
Hey Livi, what bugs did you try to fix with Sweep? I can look into those for you.
Share
I use this daily. It's 100x better than Github Copilot X and Code Interpreter.
Thanks Sagar! Appreciate the user feedback.
Anyone who doesn't care about the correctness or quality of their code would love this tool!
I was curious at the reviews so I tried the product for a few days. The audacity to claim you are "replacing junior devs" with a primitive GPT wrapper is laughable. It is just as incompetent as ChatGPT.
This tool falls short of its promises. Just like ChatGPT, it is a mindless bot that lacks any semblance of independent thinking. The code it generates for any non-trivial task is either incorrect or introduces more problems than it solves, so its basically a glorified ChatGPT wrapper. There's a reason why AI generated content is banned on stackoverflow, and developers would be wise to avoid this tool if they seek reliable and efficient coding solutions.
Hey Andrej, issues were you having? I can look into those for you.
don’t buy this product. the creator is a university dropout who is delusional about the capabilities of AI (and himself).
this looks like the state of the art in replicating the junior dev using AI at the moment. the team has the right approach and are only limited by the inherent limitations of generative AI right now. would give 5 stars if the actual experience right now would be great (right now it is just the best there is, but not great). as the foundational models (including open source ones) improve, companies like Sweep AI will be prepared to take advantage of them. a big risk for all startups in the space is the release of similar features and services by OpenAI or Microsoft Co-Pilot themselves.
Like any machine learning algorithm, it has its strengths and weaknesses. Right now there are too many issues of text-based AI to be considered robust for any developer's use (most notably, its answers are confidently incorrect about 99% of the time). It's genuinely fascinating and sad how much people overhype generative AI this year.