Get instant answers WITHOUT reading the Axe-Fx manual

Jason Scott said:
Or... if you've already read said section(s), just ask the AI. Even if a user hasn't read the manual, asking the AI for help still benefits the learning process because they're learning by doing. Further, the AI can also relay contextual information. But it's not as if users who use AI forfeit their right to read the manual.


Skimming and searching generally takes longer, which translates to more time spent looking at the screen.


Maybe, but the same people who'd ask an AI for help without reading the manual are likely the same type of people who'd ask for help on a forum without reading it. The difference is, users who ask an AI can potentially get help immediately.
Click to expand...
The jig is up! I know you are all AI bots eh! sent here to snatch the manual from my cold dead hands - I shall never forfeit direct viewing of my precious blocks guide - Long live the Table of Contents!

err um .. good points .. carry on
 
Last edited:
sprint said:
not sure where this idea comes from that people think they have to memorize manuals - they don't and they can't (well, I sure don't / can't). It's a reference - you refer to specific section(s) as needed. I'm getting old and don't always remember what I did 1/2 an hour ago - which is ok - just book mark it in the manual and go back to it as many times as you need it. As for phone screens - well - no matter what you do whether its reading a pdf directly or reading what AI is telling you about the PDF, it's gonna be a PITA cuz it's too small (I know - I'm scheduled for cataract surgery lol!). Reminds me also, that, I'd bet those that spend a little quality time with the manual are less likely to be at a gig nervously sweating over their Axfx wondering why they have no sound from output1.
Click to expand...
This AI thing is a step in the right direction. Manuals are ok when you have time to kill, but I have multiple devices, software etc each with its associated manual. It is much faster to query google and get an answer.
If this AI bot can digest multiple manuals / sources of information and give me a natural language interface, it just makes things even quicker.
 
BBN said:
No good deed goes unpunished.... kudos @Jason Scott for incorporating some new tech into the Axe adventure.
Click to expand...

Indeed. The criticism/picking apart of this is a bit over the top and I’m a bit dumbfounded by it, really. Everything starts somewhere. Dude mentioned something that could be potentially helpful now and most certainly very helpful when it’s out of it’s infancy on a forum where we congregate to talk about something that’s continuously pushed technology from it’s infancy and everyone wants to pick it apart. Interesting reaction.
 
RevDrucifer said:
Indeed. The criticism/picking apart of this is a bit over the top and I’m a bit dumbfounded by it, really. Everything starts somewhere. Dude mentioned something that could be potentially helpful now and most certainly very helpful when it’s out of it’s infancy on a forum where we congregate to talk about something that’s continuously pushed technology from it’s infancy and everyone wants to pick it apart. Interesting reaction.
Click to expand...
To me the headline seems intended to get the exact response it got in this forum. It didn't say "potentially useful tool to interface with Fractal user manual" or something else hard to argue with :)
 
Kelleys Heroes said:
To me the headline seems intended to get the exact response it got in this forum. It didn't say "potentially useful tool to interface with Fractal user manual" or something else hard to argue with :)
Click to expand...
The title was simply intended to draw attention by emphasizing the word "WITHOUT". Why? Because I think the tech is really cool and potentially useful for a lot of users (like myself), and I wanted as many eyeballs as possible to notice it. With all due respect, a title like "potentially useful tool to interface with Fractal user manual" sounds about as interesting as reading the terms and conditions of a bank loan, in my opinion, but that's me. I simply wanted a title that grabs attention and wasn't clickbait.

Having said that, the fact is, the tech can do exactly what the title states; provide answers without the need to read the manual. Why anyone would find that controversial, I don't know.
 
Last edited:
Jason Scott said:
No, the headline was simply intended to draw attention by emphasizing the word "WITHOUT". With respect, anyone who thinks it was intended to create drama is potentially projecting, in my opinion.
Click to expand...
Drama is not a word I would have used. Reminds me of all the YT videos with sensationalist headlines intended to spike views/likes, promote a product or whatever. If I was an editor, or critiquing an AI response to a query, that's a comment I might make.
 
Jason Scott said:
Information relayed by the AI, at least in this case, is derived directly from the manual.
Click to expand...
Correct. This AI only knows about information in the manual.

Its behavior was trained on a huge corpus of unorganized text, with probably only a smattering of papers from experts. What information it’s drawing from in the case of the FX3 manual is based only on what Admin M@ wrote. In other words, its opportunity for error is extremely small.
 
Kelleys Heroes said:
Drama is not a word I would have used. Reminds me of all the YT videos with sensationalist headlines intended to spike views/likes, promote a product or whatever. If I was an editor, or critiquing an AI response to a query, that's a comment I might make.
Click to expand...
I agree, though the difference is it's not clickbait. You stated that the title seemed "intended to get the exact response it got in this forum". And what kind of response was that? Controversy. That's why your comment struck me as implying that the title was intended to foment drama.
 
Last edited:
toneseeker911 said:
If this AI bot can digest multiple manuals / sources of information and give me a natural language interface, it just makes things even quicker.
Click to expand...
It can currently only digest one PDF; The site says it can do more than one but I could find no way to submit additional documents to this instance.

Admittedly I want to see how good it really is, because the current situation is not very challenging for what AI is supposed to be capable of doing.
 
Greg Ferguson said:
It can currently only digest one PDF; The site says it can do more than one but I could find no way to submit additional documents to this instance.
Click to expand...
The FAQ states that there's no limit to the number of PDF's you can upload; however, the caveat is, each interaction is limited to a single PDF. For multiple document queries, check out Visus.ai. It's similar to AskyourPDF but capable of providing aggregate answers based on multiple documents. It works well, though it's worth noting that the service is ultimately subscription based. If I recall, they provide 25 free queries per day and an additional (one-time) 25 free queries if you join their Discord server.
 
Last edited:
Jason Scott said:
The FAQ states that there's no limit to the number of PDF's you can upload; however, the caveat is, each interaction is limited to a single PDF. For multiple document queries, check out Visus.ai. It's similar to AskyourPDF but capable of providing aggregate answers based on multiple documents. It works well, though it's worth noting that the service is ultimately subscription based. If I recall, they provide 25 free queries per day and an additional (one-time) 25 free queries if you join their Discord server.
Click to expand...
The FAQ says, yes, but did you try? I did and can’t find a way to, nor does the FAQ explain how.

I suspect their goal is to allow it eventually, and the FAQ was written with that goal in mind.
 
Greg Ferguson said:
The FAQ says, yes, but did you try? I did and can’t find a way to, nor does the FAQ explain how.

I suspect their goal is to allow it eventually, and the FAQ was written with that goal in mind.
Click to expand...
I haven't tried, nor have I seen any indications it's possible.

Maybe this was recently added, but this is from the current FAQ:1684297723036.png

1684297567354.png
 
IronSean said:
It will not, ChatGTP learns from it's training data, not from it's interactions with users. Nothing you do with it or tell it persists to new chat sessions.
Click to expand...
This is not completely correct. LLMs can be incrementally trained so replies and good / bad interactions can be fed back into the model to improve its behavior on specific subjects. Happy to go into a LOT more detail if you're interested. Also chatgpt is not the only game in town, there's open source models that can be trained on curated corpuses that have similar performance to cgpt 3.5.

This is a good link to start reading more if you care about the subject: https://github.com/Hannibal046/Awesome-LLM
 
RevDrucifer said:
Indeed. The criticism/picking apart of this is a bit over the top and I’m a bit dumbfounded by it, really. Everything starts somewhere. Dude mentioned something that could be potentially helpful now and most certainly very helpful when it’s out of it’s infancy on a forum where we congregate to talk about something that’s continuously pushed technology from it’s infancy and everyone wants to pick it apart. Interesting reaction.
Click to expand...
It's just garden variety gatekeeping, not that surprising tbh. It's disappointing but not surprising.
 
Jason Scott said:
Using a PDF with an FAQ would also work.
Click to expand...
Someone got to compile it. I'm not just thinking about the manual, also thinking all the great content that has been created over the years on the forum. But sure you do you, if you have a problem with technology creates after 2019 then by all means don't use it.
 
luckymethod said:
This is not completely correct. LLMs can be incrementally trained so replies and good / bad interactions can be fed back into the model to improve its behavior on specific subjects. Happy to go into a LOT more detail if you're interested. Also chatgpt is not the only game in town, there's open source models that can be trained on curated corpuses that have similar performance to cgpt 3.5.

This is a good link to start reading more if you care about the subject: https://github.com/Hannibal046/Awesome-LLM
Click to expand...
Nice link! Thanks!
 
You must log in or register to reply here.
Back
Top Bottom