I think it's a fascist eugenics program. They think they are the strong ones and everyone else is weak, and the weak should die off so people like Elon Musk can repopulate the country with their superior genes.

Without evidence, the idea that there's a hidden camera in the exam room is just paranoid speculation. By contrast, this was an actual visible camera in the exam room, and a camera that was obviously the doctor's own personal device. It makes sense to worry more about the camera you know is real.

Where I am in Canada we do have problems with wait times, and many people can't find a family doctor. But it wasn't always like this. Successive Conservative provincial governments have cut healthcare funding again and again, trying to force people into private for-profit healthcare. It's the same story in the UK: public healthcare has been undermined by every government since Thatcher's in the 1980s, and the NHS is not what it was.

Both are still better than the US system though, even after all the cuts. I usually have to wait about 2 weeks to see my doctor unless it's an emergency. Emergencies can mean an hours-long wait at the hospital. I was recently admitted after an emergency and needed surgery and follow-up examinations including MRI, EKG, ECG, and a whole bunch of other tests with specialists. The care has been very thorough and the doctors excellent, and I have not had to worry once about costs.

Many are convinced you can't have freedom and an affordable doctor at the same time. Americans are subjected to extraordinary levels of propaganda.

Companies always say they are forced to pass their increased costs on to consumers. Oddly, they don't feel any such obligation when it comes to their reduced costs.

As I understand it, you're right that you need a good length of output to be able to detect the watermark, because only then can you see the statistical effect with confidence. And in code there are usually various options for how you get something done, and they could watermark generated code by adding a distinctive pattern to its preferences for certain constructs over others. But again, it would have to be subtle, so you'd need a large enough sample of its output before you could see the effect.

You can't enforce that with text though, because people will just delete the words. They use a subtle statistical signal because people can't just spot it and delete it. Unfortunately it raises a problem: either they have to make the signal a secret, in which case only Anthropic and those contracted to keep the secret could detect it, or it's not, in which case people can use a detector and some other AI to discover how to eliminate the signal. I expect Anthropic would try to keep control over the detection secrets and provide a detection API for third-party software to use.

It's not that obvious. They have ways of subtly statistically manipulating the output of their models, so that a computer can analyze it and recognize patterns that deviate in distinctive ways from what humans normally do. Just typing it out again won't change the statistics. Rewording might. But if you're going to rewrite the output, you lose the time savings you gained by getting AI to write your stuff, whether it's code or an essay.

Gotta pump those numbers up. Those are rookie numbers.

I'm on GrapheneOS too, but I think hardware backdoors are a genuine concern. There's not much an OS can do to fight a hardware backdoor. So I don't entirely trust it but it's better than not trying for privacy. In any case, the risk of hardware backdoors is present on pretty much every device we compute on. There's not a lot of fully open hardware out there, especially for phones.

The Trump regime has a habit of not monitoring things that would make them look bad. I wonder how much of this difference is real, and how much is due to one country trying to track cases while the other tries not to.