Radiation is not Charismatic
mediocregranular.bsky.social
Radiation is not Charismatic
@mediocregranular.bsky.social
Consistently a bitter human since the late 80's, recertified by recurring Captchas.
We should buy the domain
December 28, 2025 at 12:21 AM
Would at least this force them to be liable for what their own AI's put out? Given how stochastic the replies are and that the training data is scraped a pretty woke source, the internet and literature, it would be be surprising if they don't violate it.
December 19, 2025 at 1:35 PM
The chicken is 3D printed into shurikens
December 17, 2025 at 5:24 PM
I am sure this is nothing like 1979 Argentina where Anaconda Copper, an American company, and the CIA got a democratically elected leader ousted and a nationalization of (their) resources stopped. Followed by a dictator and regime who led the country down a spiral of failed IMF loans.
December 16, 2025 at 2:43 PM
Its almost like it loops back around to his cult members.
December 16, 2025 at 1:25 PM
It is not like they are not scraping the entire internet, in dark or light, to train their Gemini. But hey, this might benefit a human and harm them in the long run.
December 15, 2025 at 7:37 PM
I feel like the days of correlating high electric bills with "grow rooms" are long gone. Now, it might be a bitcoin miner, a kid with a Minecraft server, some bot farmer for an MMO, a dude with a home lab, or the average person with a llama/deepseek open models.
December 12, 2025 at 11:44 PM
I feel like "catastrophic flooding" was one of the headliner repercussions from Day After Tomorrow. So that much.
December 12, 2025 at 11:30 PM
Gratz on having a political leader on a tier reserved for Sid Meier's Great Leaders.
December 12, 2025 at 8:15 PM
For real-real, will this mean that platforms are also liable for shitty stuff their AI Chatbots say?
December 12, 2025 at 8:14 PM
Ah yes, the source of the problem is telling us about the problem so they can sell us a new version of the problem that is the solution.
December 11, 2025 at 12:51 PM
You are assuming they will have API tokens left over for work related questions. They will probably blow their token limit on NSFW image generation.
I also worry that this "workflow" could later be used to excuse horrific warcrimes. Just blame the AI for an illegal decision. Can't NJP a LLM model.
December 10, 2025 at 6:05 PM
Great, wake me when they realize they are leaking a ton of sensitive data via haphazardly generated AI Agents developed by vibe coders hired by DoD contractors who didn't realize their training model had a C2 implant in it because it wasn't part of the minimum security training.
December 10, 2025 at 5:54 PM