If you are looking around, there are no many products like Pickaxe. If you were here 12 months ago, I’m sure you appreciate how much the product has improved. This wouldn’t be possible without a clear vision and a commitment to excellence.
Obviously, to have the best possible product in the fast-changing world we are in means you have to make decisions that make the product more robust, scalable, and adaptable.
Yes, there are some features that are slightly different from when you bought the LTD, but you also have new features that were not available then. Would you rather have Pickaxe V1 or have a product that offers the latest features and functions?
By the way, if the 50 documents per Studio are a limitation, I can think of at least three workarounds:
- Change the model to the new GPT 4.1 Mini which has a larger context window (1M tokens - equivalent to approx 1500 pages)
- Use a custom action that connects to an external DB like Pinecone that gives you unlimited storage
(How scalable is the knowledge base? - #3 by ab2308) - Merge some of the documents and compress them to add more text in each file. Or summarise / restructure some of the content.
I haven’t seen many chatbots using 50GB of data (50 documents x 100MB each) but if you have one, let me know as I’m curious. Happy to offer some ideas on how to handle the knowledge base.