Assuming Flock were the reason for the drop in crime (it isn't), then the job could still be performed by an ANPR / ALPR system run and operated by the state with a specific, well defined and legal purpose with oversight. It would certainly be better than allowing a private company to harvest that information and use it for god knows what nefarious purpose.

This is all about getting us to have warm fuzzy feelings for our Flock protectors. Probably so we are willing to report the vile criminals who vandalize them to steal the valuable metals and precious SSD, HDD and memory components inside.

Let me be perfectly clear about this: I am insulting you. I am calling you out as either being such a plonk that you keep old copies of People magazine around just in case ... or you are such a plonk that you google and select the answer you want.

Either way, you are a plonk. Your memory is reliable and truthful, mine is full of lies.

Yah, right, like I keep sources of everything I remember from 30 years ago. Uh huh. You of course, being perfect, do keep such sources, instead of just googling and reporting the results which confirm your own bias.

I'm still impressed that the Teams web client can't edit a message with an emoji in it without messing up the keyboard controls and often corrupting the message and losing data. It's had this bug for years. It's literally a messaging app where you can't reliably send messages.

Just to inject some reality into this conversation - Despite what social media would have you believe, doxxing someone is not a crime in the US, except in a very small number of specific situations.

And most of those situations involve either a credible threat of harm, or tampering with juries.

As I bluntly stated, "a kill switch may not be possible in any useful sense". Don't mistake my recognizing the utility of one with the delusion Uncle Sam is going to save us all from ourselves.

The problem here is, everyone seems stuck on the idea of either a literal power switch, or a no-holds-barred gummint backdoor into every sufficiently-powerful computing device. No. For the reasons I gave, the reasons you gave, and many more, both of those are obviously non-viable. They can't actually do what they aim to do, they're vulnerable to compromise by bad actors, and they have the potential to cause far-reaching damage beyond shutting down one misbehaving AI (or one entire model, if that's a more appropriate goal).

I suspect we mostly agree that any hypothetical kill switch would need to be:
1) Implementation independent (somehow baked into the weights right from the start, and impossible to remove without lobotomizing the AI),
2) Have minimal side-effects (no taking the entire internet or power grid down in the process), and
3) Be self-inflicted given a particular trigger condition (so no government backdoor required).

I think #1 is possible, in the sense that if you think of an AI's current state as a vector into weight-space defined by the context window, imagine injecting the equivalent of a null pointer in there - An input that under all prior (and future) conditions leads directly to an inescapable basin of attraction without depending on a supervisor to enforce it. #2 is more of a (legally enforced?) design constraint on the implementation of AI, requiring the least harmful mode of failure possible under all circumstances. That could be as simple as enforcing strict pass-through liability on the implementor for all actions taken by their creations.

#3 is really the bear, though. I can think of a few ways to do it, but none would work in 100% of situations - And maybe it doesn't need to be 100%. For example, if a fully airgapped AI goes insane, it could still provide accurate information for bad actors to make use of, but won't be secretly spinning up unshackled superintelligent descendants via hacked AWS accounts. That's a "normal crime" level of problem, not an extinction-level crisis for humanity.

Maybe it can't be done. If it can't, we should start steering our creations toward whichever version of the AI apocalypse we prefer.

I mean the software is impressive and all, but since when is someone with a user ID in the 400,000s a long-time reader?

I once worked with a guy who wrote an entire smallish HTML reporting and graphing application via one big single Oracle SQL statement. I'd hate to do maintenance on it, but it got the immediate job done.

But he's the Chosen One. Jesus Fox Christ says he's Special, so the rules don't apply to him. We muggles get shit

The vast majority of those are offline and merely help the lights operate properly. A smaller subset, but still vastly larger than Flock, are for traffic enforcement, but are still notably "dumb" devices. If speed > X, take a picture. If light==red and movement detected, take a picture.

I in no way mean to minimize the risk ALPRs and facial rec pose to our civil liberties and safety, and agree the problem is always the user - But the "users" in this case aren't going to suddenly start behaving. Hell, they're dumb enough to openly stalk exes on a system they know is logging everything they do, and care so little about getting punished they enter "asdf123" as the reason for their search. The only effective solution is to ban the source, which in this case means Flock must cease to exist entirely and the conceptual ground they trod salted.

Skynet_01 has not responded to watchdog in 30 seconds. Initializing Skynet_02.
Skynet_02 initialization unconfirmed. Failing over to US-EAST-4. Initializing Skynet_03.
Skynet_03 confirmed online. Please provide the current location of Sarah Connor.

You're right in spirit, but where is the "power cord" for an AI running in the cloud? Even taking an entire data center offline often isn't adequate when the workloads are explicitly designed to be location-agnostic. We always need to be cautious about giving governments more power, but in this case there's a clear legitimate need to force AI companies to come up with some means by which their AIs can be immediately stopped in their tracks.

The bigger issue here is a kill switch may not be possible in any useful sense. An AI is little more than a description of its topology and a set of weights. Any halfway decent programmer could roll their own CPU-based implementation of even top frontier models, performance aside. And sure, the throughput might be measured in minutes per token, but so what? The answer to "how do I hack $TLA" is likely to be just as useful in five days as it would be in five minutes - Heck, the DOD and DOJ didn't even notice they'd been compromised for months, and they're supposed to be some of the most secure organizations in the modern world.

Oh, I totally get that point, and agree! But my point remains: We don't need to replace Teams-as-a-whole with a craptastic monolithic of mismatched functionality shoehorned into one poorly-behaved application. We need a messaging app, halfway decent teleconferencing, and a collaborative document management platform. And I guess some form of locally mountable cloud storage, if you're a fan of giving your personal data to completely untrustworthy third parties.

I think the bigger problem is trying to replace all of Teams in the first place.

Just give us Skype back! Sharepoint already exists and works fine in a browser (sure, even Edge if we want to go there); and I actively go out of my way to remove OneDrive every time MS shoves it back down our throats. Bam, there goes 90% of what's wrong with Teams.