Today we released the BETA of our new feature: User Memory. All Pro accounts should see the beta features appear in their accounts.
You can think of this feature as a memory boost for your Studio. User Memory allows Studios to remember things about a user and share these memories across other Pickaxes in the Studio. Builders can set custom instructions for exactly what to remember.
Here is a demo video that shows off the feature and explains it.
@admin_mike@nathanielmhld such a great feature! Thanks guys. I think the chaining is opening up a sea of opportunities.
Can you tell me what the memory limit is (in case of chaining) and how it could potentially affect the LLM context window? Does it work like a knowledge base where the relevant information is extracted as needed?
I can see how this feature could work beautifully for the projects I create. It would allow for an experience that feels more interactive and personalized, like engaging with a therapist who truly understands you and can respond in a meaningful way by asking about something you said last session.
This is a great feature. These are the some feedbacks I have gathered from few business owners I met and I showed them what Pickaxes can do. Based on their honest feedback these are few things pickaxes can improve on, at least based on their views
the form feature is the winning features of all because the chat function is too similar to chatgpt and it has little differentiation. They think if they can do with chatgpt, claude, perplexity then why bothered to sign up to another AI software. They believe the form is what makes everyone unique with subject expert vibes. You enter the info and this “subject expert” gives you the unique output based on their expertise. It can work like a moat. The form features and customization is still under developed. It needs more functionalities and improvement. People don’t like another chatbot responding but an answer from a proprietary system. Not to say chat is not good. It has different purposes. The “dumb down” version is more preferable than complicated chat where user has no ideas what to ask. They have to be guided like many natives app with specific sections to fill up and to generate results based on user input.
In light with this new feature, I actually got feedback the users should be allowed to fill up a form with customized questions and all the pickaxes will generate based on user filled information. This can be added at the user profile section for more personalization for specific use. For example, a marketing ai, the user have the option to fill up all the information in the profile section and all responses would be based on that, User can always modify at the profile section when needed. It makes it easier and straightforward. The idea is users dont like complicated things.
I would share more feedbacks i got from the ground. I hope this helps pickaxes to improve
I completely concur with you. I have been testing pickaxe as an internal tool. The form interface consistently ranks ahead in terms of adoption than chat interface. Forms reduce the cognitive load of prompting compared to chat and as you said, its a huge differentiator. Your observations are invaluable and completely resonate with me!
This user memory is exactly what I have been waiting for! I’m doing testing today, and I will head back here with a response. I just wanted to tell you first and foremost, thank you. This is the feature that will allow me to move my copywriting software over to Pickaxe, and my users are going to love it!
Question about the memory feature. I understand that it needs to be placed in a Studio in order for it to work with the other pickaxes I’ve created. My client wants a different interface, however. He wants me to embed the bots in a certain way on his web page. My question is, if I put a group of Pickaxes in a Studio behind the scenes, will the memory still work if I need to embed the same pickaxes individually on a web page?
Our next project is a major redesign of how pickaxes>embeds>studios all relate to each other. So the answer to that question is currently being debated in our offices presently