"What good is an AI that can't answer questions about Tiananmen Square?!" You're getting so close to understanding... Censorship aside, LLMs are not question-answering machines. They can't answer questions about anything.
My hope is that the fear of "China spying on our AI queries" and "China making AI spit out propaganda answers" would be enough to convince people that maybe treating LLMs as question-answering machines is a terrible idea all around.
@davidnjoku every LLM is inherently biased based on both the views of the team behind it and the materials it’s trained on.
How we’re this far down the “AI” conversation with so many people choosing to disregard that is somewhat alarming.
Of course next up is some LLM aggregator that uses its own model to query all the others then patch together a response taking into account all replies.
That said, I share your struggles editing #Wikipedia and have only made minor typo-level edits, and added exactly one image to an article. AFAIK they're trying to improve the editing experience.
Sidenote: I am reassured that there are efforts to improve representation. The WikiEdu non-profit works with universities to turn Wikipedia editing into student projects! Such as improving #LGBTQ inclusion:
I applied the KISS principle (Keep It Simple) because mainstream media does such a terrible job covering the Fediverse. I didn't want to add extra complexity.
Before Google decided to let extremists rename bodies of water, it used common sense.
From 2008: "...if a ruler announced that henceforth the Pacific Ocean would be named after her mother, we would not add that placemark unless and until the name came into common usage."
Some #Republican senators have campaigned against #PEPFAR’s reauthorization for 5 yrs, alleging that the program promoted abortions. In March, the program was renewed for 1 yr.
Without treatment, virus levels in people w/ #HIV will quickly spike, hobbling the immune systems of the infected people & increasing the odds that they will spread the virus to others.
About 1 in 3 untreated pregnant women may pass the virus onto their babies.
"The AI community is abuzz over DeepSeek R1, a new open-source reasoning model. The model was developed by the Chinese AI startup #DeepSeek which claims that R1 matches or even surpasses OpenAI’s ChatGPT o1 on multiple key benchmarks but operates at a fraction of the cost."
It warms my little shriveled heart to see a software company that invested in optimizing their software and got a huge benefit from it, even if it is an AI company. It can at least write wrong things while consuming 100x fewer resources 🥰
I love optimizing software and it's annoying that current software company business theory won't let programmers do that to a first order approximation
@overholt "A previous version of this article stated that the Harvard Slavery Remembrance Program had been disbanded. To clarify, according to a University spokesperson, the program has not been formally discontinued, though its employees and leaders have been laid off."
"Trump, I don’t really like traveling to the U.S. much; it’s a bit boring, but I admit there are some commendable things. I enjoy visiting the Black neighborhoods in Washington. There, I witnessed an entire battle in the U.S. capital between Black and Latino communities, with barricades and all, which I thought was nonsense because they should unite instead. . . ."