Where many AI models, not only ChatGpt. Many of them is open source. But users will not use ChatGPT always. They will like to switch from one model to another. And we need to implement memory, context, for switching AI. For example, vector-based memory, allowing for only relevant memories to be loaded for the agent at any given time. So, if you have memory model, such model will connect to ChatGPT, and explain current context. So, even if we switch ChatGPT account, such model will make auto communicate on init, and user always can continue dialogue. You can copy such implementation from AutoGpt, take only memory from here.
Wait… What??? 
Vector-based memory is a technique used in machine learning and natural language processing to store and retrieve information in a way that is similar to how the human brain works. In vector-based memory, memories are represented as vectors, which are mathematical objects that have both magnitude and direction. These vectors can be manipulated and compared using mathematical operations, allowing the machine to retrieve relevant memories based on their similarity to the current context.
I don’t know about you, but for me the less of this pseudo artificial ‘intelligence’ and generational AI stuff in my HA, the better.
4 Likes
Okay, let’s use artificial intelligence just to write code.
Artificial intelligence has the potential to revolutionize the way we write and develop code, and it can bring significant benefits to Home Assistant and other smart home systems. By using AI to generate code, software development teams can accelerate the development process while still maintaining a high standard of quality and security.
One of the main benefits of using AI for coding is that it can help to generate code quickly and efficiently. For example, machine learning algorithms can be trained on existing code bases to identify common patterns and structures, and then use this knowledge to generate new code that follows the same conventions. This can help to save developers a lot of time and effort, allowing them to focus on higher-level tasks such as designing and architecting complex systems.
Additionally, AI can help to identify and fix bugs and errors in code. By analyzing code and identifying patterns in the data, AI can help to catch errors that human developers might miss. This can help to improve the overall quality and reliability of the software, and reduce the likelihood of bugs and vulnerabilities slipping through the cracks.
Furthermore, it’s important to note that the code generated by AI will go through the same rigorous review processes and automated tests as code written by human developers. Code review is a critical part of the software development process, as it helps to ensure that code is well-structured, efficient, and secure. Automated tests such as unit tests and integration tests can also help to catch errors and vulnerabilities in the code.
However, it’s worth noting that there are some potential risks and drawbacks to using AI for coding. For example, AI-generated code may not always be well-structured or easy to understand, and it may not be suitable for all applications or scenarios. As such, it’s important to evaluate the potential benefits and risks of using AI for coding on a case-by-case basis, and to ensure that the code generated by AI is carefully reviewed and tested before being deployed.
Overall, the use of AI in coding has the potential to greatly improve the efficiency and effectiveness of software development teams, allowing them to create better and more reliable software in less time. As such, it’s worth considering the benefits that AI can bring to Home Assistant and other smart home systems, and to carefully evaluate the potential risks and benefits of incorporating AI into the development process.
I’m a software engineer by trade. I know exactly how AI generated code looks like. No thanks.
Oh and stop posting stuff generated by ChatGPT. Generational AI self-promoting the use of generational AI is a bit ridiculous.
2 Likes
Code quality is determined by code reviews, unit and UI tests, and automated checks. Accordingly, depending on the experience of the developer, a certain number of reviews, tests and other automatic checks are needed. The development lead can evaluate the level of the developer and individually set up automation for him. With properly configured processes, it doesn’t really matter who writes the code.
And that perfectly demonstrates why I don’t want this stuff in my HA.
At least do us the favor of replying yourself rather than copy and pasting a completely out of context ChatGPT reply.
well i answered it myself
And I’m out. Not in the mood of talking to a glorified autocomplete algorithm. Good luck with your FR.
1 Like
@nick2525 This is bordering on using AI generated responses as answers to questions, a bannable offense. You’re doing it in all your posts and it’s very obvious. Please stop, this is your only warning.
2 Likes
Wait… Isn’t there a helper for that? 
We need local helper, not using cloud.