I've owned more than one inexpensive Windows laptop that I only discovered to be a touchscreen by accident - like most sane people I see only ergonomic downsides in a touchscreen computer. In fact one of the reasons I prefer to use a computer with a keyboard and mouse is precisely to avoid sliding my finger around a touchscreen. With that said, it appears to add little to the cost, so I guess I am happy to continue completely ignoring it on Macs as well as PCs.

I remember the Rainbow Warrior thing well - that was the French, preventing the RW from trying to stop nuclear bomb testing. The 80s is the decade I remember best, I think :)

I am "implying" that when your ship gets sunk by an enemy mine, you don't really stop to ask where it was assembled or which flag was flying on the vehicle that planted it.

Vault-Tec is a corporate angel compared to some of the real corporate entities in today's world...

Darwin was bombed by the Japanese in WW2, so that takes Australia off your list. Canada was also hit by Japanese balloon attacks, so Canada comes off too. NZ didn't get bombed, but it did have an Axis-laid mine sink an Allied ship in its coastal waters, so borderline.

All that means (and my cynicism continues to run strong) is that the legislation is either competitively advantageous to Anthropic and OpenAI over Google and Apple, or it simply doesn't have big enough loopholes to suit Google and Apple. Big corporations - and Anthropic and OpenAI definitely qualify - don't generally have a masochistic streak; if they are eager to don handcuffs it is because those handcuffs are fur-lined and/or the big corporation has a handcuff key under its tongue.

Both the extremes you mention, and the middle ground between, are merely different flavors of chaos. I've written a bunch of substack articles on related topics (example https://larwe.substack.com/p/w... ), and after all the neuron activity I've performed on this, I'm pretty convinced that "AI" cannot be controlled in the manner that legislators would like. It will always be vulnerable to attack chain type behavior. Look no further than Multivac (also https://larwe.substack.com/p/m... :) ) and "All the Troubles of the World" Asimov short story.

The thing about analysis by CNN is that you can't easily predict what noise is going to interfere with recognition. The CNN maintainers likely have more resources than people trying to shield themselves, and they also are the only people who are in a position to run a bunch of experiments. If this technique truly does allow surveillance of some meaningful kind, it would be invisible to the end-user - so how could the end-user tell if their countermeasures are working?

... which is basically 97.625% of my body by weight, says that the reason the big companies support this legislation is because they have already figured out how to comply with it without affecting their business model or roadmap. In other words, this legislation is a paper tiger - either it has been worded in such a way as to leave important loopholes, or (even more likely, and not mutually exclusive) the big AI players know that their _products_ cannot be analyzed, at least not in finite time, and not by any auditor the government will be likely to employ. Policy documents can be reviewed, and the rubber stamps on them can be checked for color and placement, but that says nothing about what the product does; it's nondeterministic. And if there are no major disasters in Illinois for a year or so, the big companies will point to this legislation as being a "model" for rollout nationwide. At which point they propagate their convenient loopholes across the country, if not the world.

As someone who works with metaphor most of the day - I couldn't have constructed a better analogy. Hat tip to you sir/ma'am.

I file your entire reply under the category of "not even wrong".

You miss the point. Desktop PCs have a physical location and a network location. The network location is used to infer the physical location (inter alia, for the anti-user purposes I mentioned), and more often than not it is sufficiently accurate for those purposes.

We are talking about anti-privacy behavior - and you should look at the Digital Credentials API, which does more or less exactly this.

You didn't, and I do. My point was that prompting y/n for every potentially violating behavior is not a solution - it is an immense noise source that very few users will even understand, and even those who understand it won't have enough information to decide y/n for any given popup. Prompting is NOT a solution to systemic designed-broken antiprivacy architecture.

If I ever had worked for them, I would have been let go at the first RIF for being too rowdy.

Home automation systems don't have any need to go through a cloud to connect a local sensor to a local device (e.g. "motion sensor turns on lamp"). Yet... in practical terms, all the major ecosystems enforce this requirement. So, just because it isn't technically necessary to force people to use a paid cloud service doesn't mean it won't be done :(

Couple things:

• re "location information" on a desktop PC, sure it's not a mobile device, but there is still a reason to geolocate it - partly to determine regulatory domain (do I need to age-verify you because you're calling from a state that mandates it?) and partly for the usual enshittification reasons (can I show you ads for businesses that are close to you and will pay a premium to get eyeballs inside their operational area? can I observe you here frequently during the business day and infer that you work here?).
• "why not just force them to ask permission?" - look at the experience you get if you turn on "prompt every time a cookie is set". Look at the experience you get when manually editing adblock filters. There are hundreds of requests and no way to know which are site-breakingly necessary and which aren't. Yes, it's yet another enshittification outcome, but it's not feasible to ask a user to answer all these prompts. The vast majority of users aren't sufficiently sophisticated to understand what they are, no user is likely able to determine which ones must be accepted in order to accomplish a given outcome, and all users will have the hell annoyed out of them by the constant popups until they turn the prompting off.

Speaking as someone in product management, I cannot realistically believe any product manager at Facebook has the hubris (or perhaps chutzpah) to believe that the mass-market will pay the equivalent of a streaming service subscription just in order to access "AI features". I may be completely underestimating the number of people over there microdosing on shrooms, but I don't think so. Even if I'm wrong about this, it's absolutely bubblethink for an outside observer to believe that the masses will pay for this.

Anyone who wants to generate perfect randomness only needs two ingredients: 1. A cat (Not an Australian cat, for genetic reasons). 2. A bag of catnip.