TLDR: A solid 3.5 from me, will be higher once the points below are addressed.

Full disclosure - my AI prompting and leveraging skills are early stage, therefore I can only relate my my experience of Dfirst as a AI wrapper and what value it adds to my usage of AI within the remit of Dfirst.

The frustration of using individual models and cobbling together the data into actions is the overlong LLM repetative steps, long outputs and very scattered data and research across queries, sessions and AI models.
Dfirst is a wrapper which aims to add value to other ai tools. It's an attractive prospect to have all the best models available, workflows, persistent assets and a whiteboard to work on that appears to address the issues, however the devil is in the detail, only through real world use can I see if it works better.
And thats the key for me at the stage is the interface delivering sufficient efficiency and quality to be worthwhile? I am not wholly convinced yet, the whiteboard for my project has become rather difficult to work with, its very large - so large I have to continually zoom in and out to move around it or know where I am in it. Its a bit of a mess but with plenty of quality within it but its become a rather disorganised mess. On the one hand If it hadn't been so much work to get to this stage I would start again but on the other hand I'm really no clearer how to use it properly so there is not much point to me doing that.
My attempt to produce a video failed, it seemed to want me to make a separate prompt and video for every scene then stitch it all together after (too much work and not confident it would be usable)
Another irritation is UI, large whiteboards are difficult to work with, (a very large dark warehouse with a tiny torch) that requires continual zooming in/out, - exporting the good bits of content out of Dfirst didnt work well, with multiple copy/pastes for each paragraph, I gave up on it because that wasnt working.
Other issues I had were errors that I could not correct etc just a bit frustrating and with the amount of confusion I did start to wonder if it was worthwhile using Dfirst at all.

However, the objective of Dfirst is extremely valuable, it is needed.
I am hopeful that we (Dfirst and myself!) will get there;

Heres what I think
- whiteboard interface : the UI is not yet working well for me, it soon became confusing and awkward to use, possibly I went too wild with agentic ai on a startup project, im not sure.

Training materials:
- its not intuitive enough: I think this should be addressed with a multiprong user education strategy: within the app by embedding tool tips etc and user guides, help articles etc.

- also some video workshops we can refer too, the rather dry "user guide" type of information is always a priority to give a clear "bottom line" on how things work. what it can/cannot do, but i feel with a tool like this and "over the shoulder" workshop of an expert using the tool and talking through what he is thinking/deciding/doing would be excellent - Appsumo (not Jay) already did something like this which was great, I really rate that guys videos BUT he was a novice with the software exploring it and not an expert demonstrating it.

- I 'd like mutiple over the shoulder demos by an expert with Dfirst/prompting/AI/marketing demonstrating all Dfirst's capabilities with real world examples multiples of them actually, they can just be raw not overedited videos for speed - so users can get to extract value from the app as quickly as possible and also become proficient users of it so that they can provide valuable feedback that will actually help the DFirst team rather than burden them with support tickets, or provoke negative reviews.

To be clear these demos would ideally be a mix of :
1) Cradle to grave Demos : fully featured new startup projects, EVERYTHING that Dfirst can do for you, how to do in Dfirst (how to organise it into 1 whiteboard or multiple?)
2) Discrete Task Demos: eg. "How to create a logo for your new start up" , "how to create a promo video" "how to create a high converting sales page from scratch"

So overall I think the potential of Dfirst is massive, the UI and UX is not delivering yet but I am hopeful it can do - but I think it needs some work. As a user I need more support materials to be able to use this effectively at all and also to make it worthwhile from an efficiency perspective.
3) I would suggest recording user webinars too and adding them to your members materials or youtube chanel. Powerful software like this lends itself to webinars especially in the format of : Part 1 demo/tutorial of specific tasks/features
Part 2 user questions answered by the expert with an online demostration.
I think this would also be a perfect way for the team to fully test the UX and UI and refine it too because sometimes it feels like either I'm doing something wrong or the team cannot have used this tool for what I want.