Dlan
@dlan22.bsky.social
He/him
Enjoyer of TTRPGs, nature, and computer stuff
Fuck generative “AI”
My player 2 🩷 @nellie.starts.party
Enjoyer of TTRPGs, nature, and computer stuff
Fuck generative “AI”
My player 2 🩷 @nellie.starts.party
Reposted by Dlan
liberal walking into the car dealership: i want the second shittiest car you have. i will buy it for literally any price
February 16, 2026 at 2:39 AM
liberal walking into the car dealership: i want the second shittiest car you have. i will buy it for literally any price
Even if people swap out terminology (which they don’t, I’ve seen unalive for years), they could understand it within days.
The sort of language models used for modern social media moderation are way ahead of the days when you could censor a word to get past an explicitly defined bad words list.
The sort of language models used for modern social media moderation are way ahead of the days when you could censor a word to get past an explicitly defined bad words list.
January 27, 2026 at 2:08 AM
Even if people swap out terminology (which they don’t, I’ve seen unalive for years), they could understand it within days.
The sort of language models used for modern social media moderation are way ahead of the days when you could censor a word to get past an explicitly defined bad words list.
The sort of language models used for modern social media moderation are way ahead of the days when you could censor a word to get past an explicitly defined bad words list.
Even then, that’s not how it works. Semantic similarity has been researched since the 2000s and it’s a relatively easy probably for modern language models. Meaning extraction relies far more on context than on any single individual word; drop in replacements like “unalive” are trivial to detect.
January 27, 2026 at 2:08 AM
Even then, that’s not how it works. Semantic similarity has been researched since the 2000s and it’s a relatively easy probably for modern language models. Meaning extraction relies far more on context than on any single individual word; drop in replacements like “unalive” are trivial to detect.
I once saw an unmarked gray Toyota rav4 (or some similar model) turn on police lights and pull someone over. Had no idea there were non-ford/chevy unmarked police vehicles in the us
January 27, 2026 at 1:53 AM
I once saw an unmarked gray Toyota rav4 (or some similar model) turn on police lights and pull someone over. Had no idea there were non-ford/chevy unmarked police vehicles in the us
Reposted by Dlan
I mean the idea that these words are a) shadow banned and b) none of the educated professionals running the shadow ban lists can figure out "unalive" is actually a completely user driven problem on par with Boomers posting "I DONT give MarkZUCKERBERG rights to my CONTENT" on Facebook tbh.
January 26, 2026 at 10:26 PM
I mean the idea that these words are a) shadow banned and b) none of the educated professionals running the shadow ban lists can figure out "unalive" is actually a completely user driven problem on par with Boomers posting "I DONT give MarkZUCKERBERG rights to my CONTENT" on Facebook tbh.
They also think it universally applies to every platform, and the list of words (if it existed) just can’t also include euphemisms.
It’s a ridiculous hoax through and through
It’s a ridiculous hoax through and through
January 27, 2026 at 12:23 AM
They also think it universally applies to every platform, and the list of words (if it existed) just can’t also include euphemisms.
It’s a ridiculous hoax through and through
It’s a ridiculous hoax through and through