Hi Product Hunters! 👋
I'm happy to show you our vision for the future of app development - GPT Pilot beta.
🤔 Why we decided to build GPT Pilot?
After talking to over 200 users of our previous project Pythagora, there was one feedback that kept popping up. That many developers are more prone to trusting the code that’s generated from tests than the test that’s generated from code.
When you look at it closely, an automated test is nothing more than a description of what needs to happen and since LLMs process natural language so good, we decided to create a tool that would create features from a developer’s prompt.
So, we sketched how could that work and the idea behind GPT Pilot is that it serves as a blueprint of how an app development process looks like. For example, when you start working on a feature, you need to first break down the feature into smaller steps, then write initial code, run it, look at the logs, debug the code, etc.
LLMs (eg. GPT-4) can write code amazingly well. However, writing code is only 1 part of creating an application. To fully build an app, LLMs need to be able to
- write code
- run it
- read the output
- add logs around the code to better understand an issue
- debug the code
- etc.
GPT pilot provides a way so that LLMs can interact with the machine resources like the ones listed above so that AI can, not only generate code, but do real work of app development.
🏗️ How GPT Pilot works?
Here are the steps GPT Pilot takes to create an app:
1. You enter the app name and the description
2. Product Owner agent asks a couple of questions to understand the requirements better
3. Product Owner agent writes user stories and asks you if they are all correct (this helps it create code later on)
4. Architect agent writes up technologies that will be used for the app
5. DevOps agent checks if all technologies are installed on the machine and installs them if they are not
6. Tech Lead agent writes up development tasks that Developer will need to implement. This is an important part because, for each step, Tech Lead needs to specify how the user (real world developer) can review if the task is done (e.g. open localhost:3000 and do something)
7. Developer agent takes each task and writes up what needs to be done to implement it. The description is in human-readable form.
8. Finally, Code Monkey agent takes the Developer's description and the existing file and implements the changes into it. We realized this works much better than giving it to Developer right away to implement changes.
🔮 Vision behind GPT Pilot
We see how LLMs can code amazingly well. However, they are not 100% perfect so developers will be needed until we get to AGI.
We think that now, developers can act as tech leads who are reviewing the code that LLMs generate instead of coding the entire day. With this, we're imagining a future in which developers can create full production-ready apps in days rather than in months.
📖 Additional resources
👨💻 [Github repo](https://github.com/Pythagora-io/...)
📝 [Technical blog post explaining how GPT Pilot works](https://blog.pythagora.ai/2023/0...) [part 1/3]
📝 [Technical blog post explaining how GPT Pilot coding works](https://blog.pythagora.ai/2023/0...) [part 2/3]
### Finally, since GPT Pilot is open source, we’d be so grateful if you could star GPT Pilot repo on Github - https://github.com/Pythagora-io/....
After a year of iterating on 4 different test automation products, I believe we finally hit the market need so I’m super excited to present to you Pythagora -
an open-source dev tool that creates automated integration tests by analyzing server activity without you having to write a single line of code!
1. You enter the app name and the description 2. Product Owner agent asks a couple of questions to understand the requirements better 3. Product Owner agent writes user stories and asks you if they are all correct (this helps it create code later on) 4. Architect agent writes up technologies that will be used for the app 5. DevOps agent checks if all technologies are installed on the machine and installs them if they are not 6. Tech Lead agent writes up development tasks that Developer will need to implement. This is an important part because, for each step, Tech Lead needs to specify how the user (real world developer) can review if the task is done (e.g. open localhost:3000 and do something) 7. Developer agent takes each task and writes up what needs to be done to implement it. The description is in human-readable form. 8. Finally, Code Monkey agent takes the Developer's description and the existing file and implements the changes into it. We realized this works much better than giving it to Developer right away to implement changes. 🔮 Vision behind GPT Pilot We see how LLMs can code amazingly well. However, they are not 100% perfect so developers will be needed until we get to AGI. We think that now, developers can act as tech leads who are reviewing the code that LLMs generate instead of coding the entire day. With this, we're imagining a future in which developers can create full production-ready apps in days rather than in months. 📖 Additional resources 👨💻 [Github repo](https://github.com/Pythagora-io/...) 📝 [Technical blog post explaining how GPT Pilot works](https://blog.pythagora.ai/2023/0...) [part 1/3] 📝 [Technical blog post explaining how GPT Pilot coding works](https://blog.pythagora.ai/2023/0...) [part 2/3] ### Finally, since GPT Pilot is open source, we’d be so grateful if you could star GPT Pilot repo on Github - https://github.com/Pythagora-io/.... After a year of iterating on 4 different test automation products, I believe we finally hit the market need so I’m super excited to present to you Pythagora - an open-source dev tool that creates automated integration tests by analyzing server activity without you having to write a single line of code!