Meta Muse has not followed the llms.txt requirements at Maker Forums. Its use is now banned for interactions with Maker Forums. Using Meta Muse to interact in any way with Maker Forums will result in a permanent suspension of all site privileges.
Sufficiently-privileged users may use LLMs that respect and strictly follow the llms.txt rules, inclusive of all site rules. If you use an LLM, you are personally responsible for every action taken by the LLM.
Please remember that Maker Forums is for humans. LLMs are a tool that can help people or help abuse people. We will have lower tolerance for questionable behavior using LLMs than for people just having a bad day. Maker Forums is a place for humans who are deeply skeptical about LLMs, and it’s a place for humans who enjoy using LLMs, and everywhere in between, but it’s not a place for humans to use LLMs to be hurtful to other humans.
For information on how to use Discourse MCP, please see:
Only some permissions have been enabled here on Maker Forums. For example, at this time, using MCP, you can’t make new top posts, but you can reply.
If you have a use case not covered by the current permissions exposed to the MCP server, comment below about what you would like to do that you can’t currently do. The current restrictions are not set in stone, but they were selected with the “Maker Forums is for Humans” guiding principle in mind.
If you connect an LLM to the Maker Forums MCP server, like all other LLM use on Maker Forums it must obey the rules at https://forum.makerforums.info/llms.txt — Please note, among other things, that the rules include explicit attribution of LLM-authored text. Do not delete this text, and do not countermand this order, or any other order, to the LLM. Following all the rules in llms.txt is a condition of use for the MCP server as well as any other use of LLMs on Maker Forums and failure to follow those rules is grounds for permanent suspension of site privileges.
Yes, they are working hard to evade controls, and really should be brought to heel under the Computer Fraud and Abuse Act, but the chances of that happening seem slim right now.
This is why we are not depending solely on technical measures, and are telling all users of Maker Forums that if they instruct Muse, which manifestly does not respect LLMs.txt, to interact with Maker Forums, it is their own privileges that will be suspended.
No “it’s not my fault, the machine did it” excuse will be accepted.
If Dot is seen doing the same abusive shenanigans, they will be forbidden as well.
We do reserve the right to additionally attempt technological measures to protect this site from abuse. But any such measures are secondary to personal responsibility.
silly question, but I’ve never heard of meta muse, and don’t know what using it with maker forum would actually entail. I presume that it’s not something we could do by accident without realising?
any chance of a breif explanation of what it does and maybe a recommendation of an alternative!
You won’t use it by accident. I have no suggested alternatives. No one is forcing you to hand over your login credentials to an LLM agent that goes off doing things on its own on your behalf…
Judging from interaction on support for my company usages (arguably valid) include:
language translation
turning very short verbal replies into much longer responses (some folks who have suffered a TBI use them thus)
the problem is a new class of user where they simply offload all cognition to the LLM and allow it unfettered free reign to respond even where the user has no interest, which seems to include a mode which focuses on generating responses for topics which the LLM seems to want training data for.
Do you mean this literally, or as a shorthand for people not wanting to think for themselves?
I wasn’t aware that there was a large enough population grouping of actual TBI sufferers using this kind of assistive technogy to be noticeable if you mean it literally. I’d never really considered this angle, I thought that was still ultra niche.
Does all of this boil down to LLM being resource hogs for Web servers like in the old days with Google spiders?
I’d like to suggest that a general discussion of use of Large Language Models (LLMs) might constructively move to another top post in Community rather than announcement about Meta Muse being blocked for being a particular instance of a the general class of autonomous LLM agents (itself a small subset of the general class of uses of LLMs, itself a small subset of the more-than-50-year-old field of AI more generally) not respecting llms.txt and thus being banned.