Good that you're citing your sources. But how does it work legally? Does the CC license in SE allow you to copy and reuse code that is more than a short snippet? Does your product have a compatible license?

I see no differences from source and binary: most people just need the binary, and will only download that. But you can distribute them both, and you should do that, if you want to call your product "open source".

And, in any case, people don't have the money or resources to train the model *today*. That's not an excuse not to distribute the source.

Aren't the pre-trained weights basically the equivalent of a compiled binary? You can run them as they are, but they are an unmodifiable black box, and you don't have access to the source material in case you want to tweak the process that generated them.
By Stallman standards I wouldn't call this "open source" at all.

The danger is people dodging their responsibilities. If the answer to "why did you bomb that hospital" is "because the computer told me so", and that's acceptable, it becomes a problem.

Wait, does Google control even TLDs now? Who made them evil overlords? How did that happen?

Well, you can't change their password, but you can still read all their data.

The critical factor is not local vs. Microsoft accounts, it's having full disk encryption vs. not having it.

What's a circle-based system?

This must be like public healthcare: every developed country has it figured out and working reasonably well, apart from the US. So who is at fault here?

My example does not assume constant control by a family member. You just need the voter to be under your control at the time electronic voting ends. Or, even simpler, for them to give you their credentials for electronic voting, so *you* can vote for them at the end of the allotted time.

Yes, this example requires a party representative to be on the board. I am not sure which threat model you have, but to me this seems the most dangerous scenario, not individual efforts to coerce one single vote.

Yes, I agree that paper voting has its drawbacks, too, and cameras are an important loophole. But checking if a person goes to the polls or not seems simpler than giving them a camera, at least in my view. I assume the voting officials keep records of who goes to the polling station, and that party representatives can audit them.

Please don't assume I'm the one misunderstanding things here.

Let me give an example. Party A offers me $1000 if I vote for them. To prove that I voted for them, they ask me to vote electronically in their presence at the time when electronic voting closes. Then, their representative monitors the poll station and checks I am not overriding the electronic vote with a paper vote: if they see me at the poll station, no $1000 for me.

Your secure voting system cannot prevent this vote control exploit, can it?

...unless you have a representative at the polling site that checks who goes in and who doesn't. And it is important that representatives are allowed into physical voting locations to check for fraud.

The pencils used in elections are indelible, too. They contain special coloring substances that get transferred to the paper. See e.g. https://cool.culturalheritage.org/coolaic/sg/bpg/annual/v17/bp17-05.html .

You can self-host, I'm doing it now. It's a memory hog if you install the full English n-grams (about 4GB), but it works well.

Here are some internet threads from 2021 discussing this practice: https://forums.macrumors.com/threads/why-do-i-have-to-turn-off-find-my-iphone-for-a-repair.2316549/ https://apple.stackexchange.com/questions/423785/why-do-i-have-to-disable-find-my-iphone . Of course, they do it "for your safety".

Sadly, I think it's more likely that airlines will soon forbid people from using Airtags on checked luggage.

Apple already does the same: they ask you to switch off "Find my Iphone" when you send yours in for repair.