I mean, if people here don't like how Reddit took advantage of user comment data, why should we archive the same without consent from the people who wrote them? Legally speaking Reddit holds the copyright also.
I admit C++ ain't safe, but wonder if there's an alternative to going Rust. Don't get me wrong, I love the language. But Rust is a beast on its own. I read here that game devs generally can't adapt Rust because the language forces frequent refactoring, which doesn't fit the business speed of game development.
This is just labeling. You can label everything as bad at will. I'm fine with that, it's called "you're entitled to your opinion". That's not objective though.
I can say it's both on point and not. For the not, you can ban the gun in the UK and it will be very difficult to bring one from the continent. Peace. But the same is not true for AI. If the UK government bans AI, Russia can still bring it through the internet.
And then I can still counter-argue that one, and then counter-argue this one also. See what a mess a metaphoric arguments bring.
I already write one reply to tell my main point. But whatever argument you come up with, I don't think that'll match the reality as viewed by AI researchers. If you give me specific short questions I'd be happy to engage in a discussion, with conditions on time.
In any case, I won't listen to metaphoric arguments like yours with guns because metaphoric arguments are very difficult to do scientifically. Every situation is different. I mean that anybody can always end the discussion saying "that's oranges vs apples", and everything time this happens you'd not have an objective way to counter that.
The reality is, passing this huge amount of data was the only way these crazy current AI models work as powerful as ChatGPT. With a restriction like you fantasize, the AI programs would have been dominated by bad actors and the west would not have a counter technology for a decade if not longer.
Regulating the outputs of AIs would be a separate story. But it's still overwhelmingly difficult. OpenAI is actually advanced in this region in the sense that they have in pocket the single best technology to politically balance the replies by a chatbot.
Generally speaking, all runtimes have been traditionally called programs. (On Unix systems runtimes are often synonymous to executables. I guess the term runtime is used more often by devs on the Windows and Java platform, and I think it is specifically an antonym of library, but not sure because I don't develop on those often) Applications traditionally referred to programs that were exposed to the user through a mouse interaction by intention. On macOS an app has the .app extension and is thus a special type of a program.
Although, depending on the context, an "application" might just mean programs because even official tech manuals aren't perfectly rigorous.
On Linux and Windows it is similar. They don't have a specific extension (some .exe binaries on Windows are meant to be run through the commandline.)
Software is the antonym of hardware, as I wrote in another comment.
Honestly I'm surprised that people here don't share this. The terminology was rather cleanly separated before iPhone. Unfortunately, due to smartphones the word "app" entered the mass population and it lost meaning as usual.
We aren't naive. We all knew this will happen. But, as it happened, it was better than banning AI in the free world and giving dictators advantages in AI tech.
I think one has to gather more proof before concluding that the gap is due to LLMs. It can also be that the engagement was lost due to third party app drop. We don't have stats to distinguish them.
I mean, if people here don't like how Reddit took advantage of user comment data, why should we archive the same without consent from the people who wrote them? Legally speaking Reddit holds the copyright also.