Email or username:

Password:

Forgot your password?
Top-level
Mx. Luna Corbden

@accretionist @paul @wikipedia I’m almost certain some group or other is planning ways to destroy Wikipedia as we speak. One way that I’ve heard rumors might be starting already is the use of LLMs to spam edits in edit wars on controversial pages. I don’t take Wikipedia for granted AT ALL.

13 comments
Richard Knipel

@corbden @accretionist @paul @wikipedia Groups of every conceivable brand of bad politics have been attempting this for 20+ years. It has survived the ill-will of various governments, corporations, and other vandals thanks to the resilience of a global volunteer community that indeed does not take it for granted. Wikipedia abides.

DELETED

@Pharos @corbden @paul @wikipedia

absolutely. i admire wikipedia's scrappy resilience

Mx. Luna Corbden

@accretionist @Pharos @paul @wikipedia

I agree with that, and was editing Wikipedia in 2005 so I get it, however:

1. I’ve been online a long time and seen things go away that I never thought possible to destroy or decay.

2. There are tools now available to vandals that have never been available before. I read one thing a few weeks ago by a seasoned wikipedian about seeing floods of automated LLM edits to talk pages on controversial articles that was beginning to tax their ability to manage.

mstdn.social/@kissane/11032842

@accretionist @Pharos @paul @wikipedia

I agree with that, and was editing Wikipedia in 2005 so I get it, however:

1. I’ve been online a long time and seen things go away that I never thought possible to destroy or decay.

2. There are tools now available to vandals that have never been available before. I read one thing a few weeks ago by a seasoned wikipedian about seeing floods of automated LLM edits to talk pages on controversial articles that was beginning to tax their ability to manage.

DELETED

@corbden @Pharos @paul @wikipedia

LLM edits seem to be a new form of spam. "cognitive spam." it's a matter of adopting to the new spam with countermeasures. a constant arms race

Lotus

@corbden While not to destroy it, I only seem to hear skepticism about wikipedia from socially conservative people (Im not american so Im not sure if that applies to american conservatives.) Not sure why specifically.

Stuff like doubts about anyone being able to edit (which I then have to follow with everything needing citations and such.)

DELETED

@LotusHopper @corbden

why change your views to fit the data when you can change the data to fit your worldview. and if a resource like wikipedia is immune to that effort, and stands there unmoved by your denial and handwaving, it must be gaslit and have its credibility attacked. the attacks don't have to make sense, they just have to be constant

this is standard operating procedure for some minds: cognitive dissonance is painful, and changing one's mind is apparently an impossibility, so attack attack attack

@LotusHopper @corbden

why change your views to fit the data when you can change the data to fit your worldview. and if a resource like wikipedia is immune to that effort, and stands there unmoved by your denial and handwaving, it must be gaslit and have its credibility attacked. the attacks don't have to make sense, they just have to be constant

Mx. Luna Corbden

@accretionist @LotusHopper I have defended Wikipedia’s model many-a-time. Lotus, if you’re not in the states, please don’t underestimate the power of insular fundamentalist evangelical groups here, who have been planning a coordinated Christofascist takeover for decades. Highly recommend the docuseries Shiny Happy People which reveals the inner workings of just one such group. I’m a former member of another. They’ve already taken over vast parts of our government. I’ve stopped underestimating the lengths they will go to. They have the will and will do anything within their power to get more power. I know Wikipedia is resilient. I once thought the US government, SCOTUS, the presidency, congress, was resilient, too.

@accretionist @LotusHopper I have defended Wikipedia’s model many-a-time. Lotus, if you’re not in the states, please don’t underestimate the power of insular fundamentalist evangelical groups here, who have been planning a coordinated Christofascist takeover for decades. Highly recommend the docuseries Shiny Happy People which reveals the inner workings of just one such group. I’m a former member of another. They’ve already taken over vast parts of our government. I’ve stopped underestimating the...

DELETED

@corbden @LotusHopper oh yeah, certain trends are deeply disturbing. but this fight is all there ever was, and ever will be. constant erosion and repair. this is all there is in life. that's not a point for despair, that's a point to smile back at the smile of the malicious, and move on with good work

not out of validation from the world. that may come, but it's a side effect. it's for self-validation: i don't litter. not because i fear punishment. but because i fear what i would think of myself

and as long as people like you and i exist, and we always will, there is progress in this world. and don't ever think the malice and despair out there is greater than that

@corbden @LotusHopper oh yeah, certain trends are deeply disturbing. but this fight is all there ever was, and ever will be. constant erosion and repair. this is all there is in life. that's not a point for despair, that's a point to smile back at the smile of the malicious, and move on with good work

Mx. Luna Corbden

@accretionist @LotusHopper I hope so. I try to keep a good attitude and fight. I just think complacency is our enemy. We can’t just rely on our institutions to keep on doing their thing without effort to make sure they do.

Paul Sutton

@corbden @accretionist @paul @wikipedia

Sorry what is LLM?, people keep talking about them but no one has actually explained what it means?

DELETED

@zleap LLM = large language model. think chatgpt. you assemble this huge trove of conversational data and pluck out new artificial conversations in a statistical manner

Mx. Luna Corbden

@zleap @accretionist @paul @wikipedia Large Language Models. ChatGPT and their ilk. A type of “AI” that analyses large sets of language and learns ways to regurgitate it such that it’s hard to tell it’s artificial.

en.wikipedia.org/wiki/Large_la

Go Up