If you've worked with ChatGPT for a while, certain things start to irk you: even when you pay for access, it will too often remark that as an AI large language model it has its limitations, it will make numerous mistakes, and it tends to exhibit woke behavior, wanting to be politically correct all the time.
Enter Open Assistant: an open source project that is aimed at training an open source counterpart to OpenAI's ChatGPT since that has not been open source despite its earlier roots when Elon Musk invested 100 million dollars in it. Open Assistant is a chat model that has a few extra functions aimed at training the model: you can provide it with prompts, but it is also possible to help train the model by reviewing prompts and their replies.
This interview discusses the case of Open Assistant vs. ChatGPT, and how a spontaneous group of enthusiasts came op with a system that can learn while it operates.