Google cares deeply about privacy. Google defines privacy as them not giving your private data that they have collected to anyone else unless you ask them to.
Google cares deeply about privacy. Google defines privacy as them not giving your private data that they have collected to anyone who hasn't paid them for it or can compel them to give it up.
There's a fourth amendment case on the Supreme Court docket (Chatrie v. U.S.) about Google searching a massive amount of user data to find people in a location at a specific time, at police request. The case is about whether the police's warrant warranted such a wide scope of search (if general warrants are allowed).
Point being: Google will 100% give your info to the police, regardless of whether the police have the legal right to it or not, and regardless of whether you actually committed a crime or not.
Bonus points: the federal court that ruled on the case said that it likely violated the fourth amendment, but they allowed the police to admit the evidence anyway because of the "good faith" clause, which is a new one for me. Time to add it to the list of horribly abusable exceptions (qualified immunity, civil asset forfeiture, and eminent domain coming to mind).
The breaking point with me that caused me to de-google myself was finding out that Google was buying Mastercard records in order to cross-reference them with Android phone data. That shit is not okay.
So no compelling here. The police asked for it and google gave it, either for free or in exchange for money. They didn't say "no" to the police, they didn't wait for a court order.
The bad guy here is google. And the people that champion data collection by private companies because of free market == good.
In that case, the main bad guy was the police who didn't bother to do even the most basic investigating after "check Google's GPS records to see who was at the house" including "Check Google's GPS records to see how how long they were there" which would have shown them this was a drive by, but yeah Google is absolutely a villain
Ah yes, I should have said I was describing the official line, not the behaviour. In all fairness the “can compel them to give it up” doesn’t seem to be optional but otherwise, yeah. Agreed.
Agreed, but also, permission prompts are way overused and often meaningless to anyone at all, even fellow software engineers. “This program [program.exe] wants to do stuff, yes/no?” How should I know what’s safe to say yes to?
I think Android’s ‘permissions’ early on (maybe it’s improved?) and Microsoft’s blanket ‘this program wants to do things’ authorisation pop up have set a standard here that we shouldn’t still be following.
This is just the next iteration of the issues with Linux file permissions, where the original threat model was “the computer is used by many users who need protection from each other”, and which no longer makes much sense in a world of “the computer is used by one or more users who need protection from each other and also from the huge amounts of potentially malicious remote code they constantly execute”.
Each bigger one has in fact solved more of physics, after being built precisely because there was a good theoretical case for a higher energy collider being helpful.
As somebody slightly better informed (physics degree, following popular science): It really isn't looking great for something that could be found at a small multiple of current energies, but not at current energies.
In all fairness most of the unique stuff I can do is probably an artifact of my training process, so it seems unfair to deny an LLM the same accomodation.
This got me thinking, and it might actually even be a comparable amount.
Let's estimate 12 years of schooling run at minimum $100,000 per student, at least in the US [1], and then add onto that number whatever else you may do after that, i.e. a bunch more money if paid (college) or "unpaid" (self-taught skills and improvements) education, and then the likely biggest portion for white-collar workers, yet hard-to-quantify, in experience and "value" professional work will equip one with.
Now divide the average SOTA LLM's training cost (or a guess, since these numbers aren't always published as far as I'm aware) by the number of users, or if you wanted to be more strict, the number of people it's proven to be useful for (what else would training be for), and it might not be so far off anymore?
Of course, whether it makes sense to divide and spread out the LLMs' costs across users in order to calculate an "average utility" is debatable.
But we know that nobody really does that - it's just too hard and lonely to accept all responsibility solely to ourselves, even as an adult. That's why we learn to rely on each other sometimes, and I don't think that's a childish or irresponsible act!
reply