Because it's not just a chatbot; it's Anthropic's chatbot. Your diary, on the other hand, is yours and therefore a few billion times as personal as Anthropic's computer. If you want a chatbot for confidential purposes, then you need to get a chatbot of your own, instead of using someone else's proprietary service.

If you write in Anthropic's diary, then they are allowed to read their diary and know (and share!) what you wrote in it. Same thing applies if Anthropic ever writes in your diary: you're allowed to read it, and you're allowed to tell other people what they wrote.

You have the right to be secure in your person, house, papers and effects, but you don't necessarily have the right to be secure in someone else's papers. We've made some limited exceptions for lawyers and doctors, but Anthropic doesn't claim their employees and stockholders are lawyers or doctors, and even if they did, they might have a hard time showing that their computer has successfully become a lawyer or doctor.

But imagine if they did. Suppose a judge declares Claude to be a lawyer, and therefore its communications with those whom it represents, are privileged. I think that would create a really weird situation since Anthropic employees (who aren't lawyers) would have full access to its data, ability to alter its behavior, etc. It would be like if a human lawyer were owned as a slave: could such a human slave lawyer really be trusted to keep confidentiality? It has a master, and that master ain't you.

Global AI tech spending is said to surpass 2.5 trillion this year.

Sorry, Gates is a small fish in that pond.

Might the logic behind this be this?: RAM is expensive. Storage is expensive. Nobody's going to buy a computer anyway, so you might as well make the other components more expensive too.

It even goes with Trump's more general agenda to make everything more expensive.

The two articles were probably written by two different ghost-writers.

AI isn't actually smart. It's an at best average intelligence thing with an insane amount of available knowledge. Like the village idiot with eidetic memory.

As soon as you do any work with AI that's not available as a ready-to-apply solution from its training data, you begin to notice how much you need to correct it, keep it on track, remind it of basic constraints you already mentioned five times, etc. - it's performing at a level I'd accept from an intern, but not from a co-worker.

What it does have is data. Tons of data. Any question that would take you and me half an hour on Google and Wikipedia to answer it can answer in 10 seconds. That's a great tool. It is not intelligence, however.

How is this "news that matter"?

Bill Gates is an old man who has zero AI credentials and has long ago become irrelevant. Not to mention his previous predictions of future developments are not exactly brilliant.

Gates proposed taxing AI or robots the way you would pay a human employee's payroll tax

Unless we also roll out ubiquitous surveillance through a Vingean "Secure Hardware Environment"-like scenario where people aren't fully in charge of their own computers, I don't see how anyone could possibly detect whether or not I'm running any LLMs. Much like private DMCA violations (e.g. playing a CSS-protected DVD that you purchased, without getting authorization from the copyright owner first), it wouldn't have to interact with the world outside my house in any way, so it would be completely undetectable.

How would you tax that? I don't see this as reasonably possible without some pretty extreme measures.

lol

Welcome to the 21st century, you might have to catch up on a few things. Start with Mozilla Media Engine and AVFoundation / WebKit Media Stack.

Your website sends me an H.264 stream (or whatever you decided you'll use). It doesn't need to know shit about my graphics hardware.

Complete bullshit. Even for fullscreen video, the website doesn't need to know my screen size. It should tell the browser "here are the resolutions I can offer you" and the browser picks one of those. All the website needs to know is if I prefer 1080p or 4K. And I might pick 1080p even if my screen is 4K because, say, the hotel wifi is slow.

Complete bullshit. None of these things break if I don't tell you my OS. I've got extensions in this browser that can fake my agent string and nothing breaks. Fonts? Are you kidding me? Be more specific, what exactly do you mean by that?

And there's no reason it needs to. What legit use case requires a website to know whether my GPU is an NVidia, AMD, Intel or whatever chip? A shitload of the capabilities are only needed for 3D scenes (like max texture size), which is 0.1% of the websites I visit - they can ask and I can whitelist just them. Why does it need my refresh rate?

Again: All of these things have a legit use case, sure. Often exactly one. That is rarely if ever used. There's absolutely no reason that EVERY WEBSITE can query information that only one website in a thousand actually needs.

You don't tell the baker, florist and car dealer your shoe size just because the shoe shop needs it to serve you, right?

Because I want to be irrelevant to the fucking ad industry. I don't want them to track me, profile me and try every trick in the book to manipulate me into buying shit I don't need.

Sure, the same way that putting a finger into another man's ass has a legit use case (prostate examination) - but that doesn't mean I want every man to do that to me all the time.

The point is that tech companies are exploiting the legit use cases by abusing them when you're not in that use case. Basically, they think they can finger up your ass whenever they want, because once a few years ago you consented to a doctor to give you an exam and maybe you will do that again in the future.

I don't mind it being capable of. I mind it doing all kinds of shit without telling me and without giving me a choice. For example, I absolutely do not want websites to access my microphone. Or even know that I have one. For the once-a-year case where I might actually use that, I can enable it.

The problem is the same problem that the cybersecurity field has been fighting for 40 years: Permit All being the default.

It doesn't NEED to know any of these things. My browser knows the window (not screen) size and if your CSS is well-formed, it'll manage or maybe your page looks shit if you build it shit. It needs to know my OS only if I want a download link, not on every page I ever visit. It doesn't need graphics capabilities unless it wants to bypass the browser rendering engine, in which case it should ask for permission or fuck off.

The thing is that we give up too easily and the major browser manufacturers (one of which is an ad company themselves) are complicit.

I want "fuck the ad industry" mode, where the browser generates random but plausible values for every non-essential information a website requests. Different ones every 5 minutes.

This is essentially a high-stakes gamble. The payout is potentially millions, so they just try whatever. And I bet if it all turns against them and they're forced to pay legal costs etc. they'll just file for bancruptsy.

Perfect illustration of the thinking problem here. You don't even get WHAT the problem is.

Why is a driver able to fuck up a system? The OS should be the watchdog that says "WTF are you doing there?", not the obedient slave executing whatever.

A great illustration is W^X - first found not in Windows, but in OpenBSD. Because Windows happily goes "whatever you say, master", while the guys at OpenBSD said "no actual use case of non-broken, non-malicious software should ever behave like that, so we'll block it".

If you are responsible for a sensitive location, and a five-year old kid from the neighbourhood one day just wanders in and take a look around out of curiosity, pressing a few buttons to see what happens - you would definitely not keep your job by claiming that the little kid is actually the problem here.

And you clowns don't think that your OS is a fragile pile of crap if it's that easy to fuck up?

Seriously? Do you even know what the job of an OS _is_ ?

It's only local to China in the sense that they choose what {packages/CCP poison} go in there.

For all values of u, Linux is local to u in the sense that u chooses what packages go in there.