manicdave

and now a foreign football team has had to ban their own supporters from travelling here because our government is so sycophantic they were committed to causing a race riot in the country's second city which would probably have forced their own collapse

For example, Britain's national mapping organisation's brand is associated in our national consciousness with going to a small shop in a quaint village to get a map showing how to walk up a mountain. It's called Ordnance Survey. If that sounds like Artillery Research to you, that's because the project started because the king wanted to know how to accurately bomb Scotland.

So I was just reading this thread about deepseek refusing to answer questions about Tianenmen square.

It seems obvious from screenshots of people trying to jailbreak the webapp that there's some middleware that just drops the connection when the incident is mentioned. However I've already asked the self hosted model multiple controversial China questions and it's answered them all.

The poster of the thread was also running the model locally, the 14b model to be specific, so what's happening? I decide to check for myself and lo and behold, I get the same "I am sorry, I cannot answer that question. I am an AI assistant designed to provide helpful and harmless responses."

Is it just that specific model being censored? Is it because it's the qwen model it's distilled from that's censored? But isn't the 7b model also distilled from qwen?

So I check the 7b model again, and this time round that's also censored. I panic for a few seconds. Have the Chinese somehow broken into my local model to cover it up after I downloaded it.

I check the screenshot I have of it answering the first time I asked and ask the exact same question again, and not only does it work, it acknowledges the previous question.

So wtf is going on? It seems that "Tianenmen square" will clumsily shut down any kind of response, but Tiananmen square is completely fine to discuss.

So the local model actually is censored, but the filter is so shit, you might not even notice it.

It'll be interesting to see what happens with the next release. Will the censorship be less thorough, stay the same, or will china again piss away a massive amount of soft power and goodwill over something that everybody knows about anyway?

Just listened to him ramble on about AI for an eternity on the radio.

We're all going to die.

Are his advisors all so lacking in knowing anything as to have never spotted a mistake in Google's AI interventions.

We're just on the brink of bankruptcy and he's diving headfirst into what the rest of the world is beginning to realise is a boondoggle.

When people say there's been an "𝑥 fold increase in such and such." They mean such and such is 𝑥 times as big.

If you get something that actually folds like a sheet of paper, the amount of layers doubles each time. One fold = twice as many layers. Two folds = four times as many layers...

This warning is on the Indian release only

As far as I know this isn't a feature that exists, but I know the protocol should make it fairly easy.

What I'm thinking is for Lemmy to basically have the option to inherit the comment thread when posting a URL to another fediverse service.

E.g. If crossposting a Lemmy thread, you get a tickbox saying "inherit comment section" or something and it makes the new thread effectively a symlink to the original.

This could also be used to bootstrap other fediverse services like pixelfed and peertube by enabling people to comment directly from their Lemmy instance.