𝙲𝚑𝚊𝚒𝚛𝚖𝚊𝚗 𝙼𝚎𝚘𝚠

Nothing lasts forever. But for now, it’s decent enough.

I’m surprised to hear GIMP crashed on you, I don’t think I’ve ever had it crash on me.

LinEUx

Our company has directly profited from a competitor that leaked sensitive data, because some of their large corporate customers decided to switch to us.

Business don’t like being on the receiving end of a data leak either you know.

I think you’re being too pessimistic about IT security, particularly in the Financial sector. A lot of the security rules and audits aren’t even government-run, it’s the sector regulating itself. And trust me, they are pretty thorough and quite nitpicky about stuff.

The cost of failing an audit also often isn’t even a fine, it’s direct exclusion from a payment scheme. Basically, do it right or don’t do it at all. Given that that is a strict requirement for staying in business, most of these companies will have sufficiently invested in IT security.

Of course it’s not airtight, no system really is. But particularly in the financial sector most companies really do have their IT security in order.

That’s not entirely true. In order to be allowed to keep processing transactions you have to adhere to strict rules which do get regularly audited. And then there’s the whole “customers will switch to another more reliable party in case of outages or security problems”. And trust me, I’ve seen first-hand that they do.

Not Tesla though, it relies on cameras only.

Except the part where it said downloading videos is against their terms of service? Which was my only point?

Did you completely fail to read the part “except where authorized”? That bit of legalese is a blanket “you can’t use this software in a way we don’t want to”.

You physically cannot download files to a browser. A browser is a piece of software. It does not allow you to download anything

Ah, you just have zero clue what you’re talking about, but you think you do. I can point out exactly where you are on the Dunning-Kruger curve.

This is such a wild conversation and ridiculous mental gymnastics. I think we’re done here.

Hilarious coming from you, who has ignored every bit of information people have thrown at you to get you to understand. But agreed, this is not going anywhere.

Yes, by allowing you to download the video file to the browser. This snippet of legal terms didn’t really reinforce any of your points.

But it actually is helpful for mine. In legalese, downloading and storing a file actually falls under reproduction, as this essentially creates an unauthorized copy of the data if not expressly allowed. It’s legally separate from downloading, which is just the act of moving data from one computer to another. Downloading also kind of pedantically necessitates reproduction to the temporary memory of the computer (eg RAM), but this temporary reproduction is most cases allowed (except when it comes to copyrighted material from an illegal source, for example).

In legalese here, the “downloading” specifically refers to retrieving server data in an unauthorized manner (eg a bot farm downloading videos, or trying to watch a video that’s not supposed to be out yet). Storing this data to file falls under the legal definition of reproduction instead.

except: (a) as expressly authorized by the Service

Can you read?

No, that’s “Download to file” or “Download and save”. Just because some people like to refer to downloading and saving as just “downloading”, doesn’t mean that that magically now means that. You out of all people, who likes to rail against people using wrong definitions, should realise this.

The CS definition has never directly implied that downloading must also store the received data.

Would they? The XZ utils backdoor was only discovered by what can only be described as an insanely attentive developer who happened to be testing something unrelated and who happened to notice a small increase in the startup time of the library, and was curious enough to go and figure out why.

Open does not mean “can’t be backdoored”.

For example your second source says “downloaded over the internet” and since YouTube doesn’t allow you to download videos, YT videos would be omitted from that definition.

Everything on the internet is “downloaded” to your device, otherwise you can’t view it. It just means receiving data from a remote server.

In general, you should pay for content that you’re going to use commercially

Sure, but merely linking to a page isn’t reusing the content. If said content was being embedded, rehashed or otherwise shown then a compensation would be fair. But merely linking to a page should absolutely be free. That’s a massively important cornerstone of the internet that shouldn’t be compromised on.

Linking directs traffic which can be monetized by the website itself, it shouldn’t require additional fees on top.

Yes, but at least there they still use “Earth time”, just slowed down. For the moon it gets a little bit more complicated I guess.

Time moves at a different speed due to the moon’s reduced gravity. It’s not just the length of a day.

Eh, I have a few things from Kickstarter that were successful. Exploding Kittens is probably the most successful one of all the ones I own.

Isn’t Umbraco the one that struggled loading a page that didn’t exist, taking several seconds to load the PageNotFound page and causing very high CPU load in the meantime? Like, an issue they had for years?

Somehow I don’t have great faith in that solution, but perhaps it’s improved in recent years.

What they didn’t prove, at least by my reading of this paper, is that achieving general intelligence itself is an NP-hard problem. It’s just that this particular method of inferential training, what they call “AI-by-Learning,” is an NP-hard computational problem.

This is exactly what they’ve proven. They found that if you can solve AI-by-Learning in polynomial time, you can also solve random-vs-chance (or whatever it was called) in a tractable time, which is a known NP-Hard problem. Ergo, the current learning techniques which are tractable will never result in AGI, and any technique that could must necessarily be considerably slower (otherwise you can use the exact same proof presented in the paper again).

They merely mentioned these methods to show that it doesn’t matter which method you pick. The explicit point is to show that it doesn’t matter if you use LLMs or RNNs or whatever; it will never be able to turn into a true AGI. It could be a good AI of course, but that G is pretty important here.

But it’s easy to just define general intelligence as something approximating what humans already do.

No, General Intelligence has a set definition that the paper’s authors stick with. It’s not as simple as “it’s a human-like intelligence” or something that merely approximates it.

Yes, hence we’re not “right around the corner”, it’s a figure of speech that uses spatial distance to metaphorically show we’re very far away from something.