My earlier post regarding "what's the difference between an Assistant API vs. Botpress KB + AI Task": A good question that many others are also wondering. Not just us Botpress builders, but anyone who wants to switch their current systems to use the OpenAI Assistant API. The important thing is to start by setting usage limits and a maximum monthly budget for your OpenAI account, so that there comes no surprises. On the OpenAI help forum you see a lot of questions and discussions that GPT-3.5 isn't smart enough to do all Assistant API stuff, but GPT-4 runs can create hundreds of thousands of tokens (or even millions of tokens as some users have shown in print screens from their accounts). The reason is that they've been uploading files to the Assistant API, and doing searches on that knowledge base, and it turns out that they've been paying for the whole file with each message. I'm trying to build two different versions of some projects, one using the OpenAI Assistant API, and the other doing the exact same thing using only Botpress. I want to see if there is any difference in my use cases other than that using Botpress only is faster and cheaper.