Sauron

It would be really weird for them to go with a downright different architecture for only one of their SKUs... it would be really hard to enforce full compatibility. If they're going to switch I'd expect it would be across all devices.

My point isn't that games couldn't be smaller while still looking good - my point is it has nothing to do with programming. Let me show you. Take, for example, DOOM 2016, a 70gb game: Almost all of it is accounted for by textures, models, video and sound. The executables and DLLs are only a few hundred kb each. Take any modern game and the same will apply. No matter how big and poorly made your program is, producing an executable that is more than a gig is a challenge in its own right. -edit- unless of course you package the assets inside the executable, but you get my point

Yes it is. Well, textures, models, audio etc. Not always the case, and dlss upscales the whole output image, not individual textures and models. It's possible that sometimes assets are needlessly high resolution or not well compressed and end up being mostly wasted space, I think there was controversy a while back because a COD game did not bother compressing the audio and ended up with 50gb of FLAC files... but that's not a given. And again it has nothing to do with code optimization. Pretty irrelevant for size. Prebaked lighting would occupy more space on disk, not less. It would make the game run faster but definitely not less "bloated". I agree with the sentiment but it kind of depends on what "AI" is even understood to mean. Just a few years back you would use "AI" to refer to simple game NPC behavior.

I'm not sure that's true and even if it were, which "new technology" would have been responsible for that? BIG [citation needed] on that one. Comparing neural networks to the human brain has always been a stretch and CNNs have lost even the superficial appearance of human neurons. Are they though? Recent "breakthroughs" have mostly been the result of iteration on ideas introduced several years ago. The difference in performance between GPT generations are in large part due to simply adding more data in the training sets. Similarly, afaik hardware has gotten faster mainly by adding more stuff rather than radically changing the way processing is done, at least in the last few years. This has absolutely nothing to do with programming or optimization. The vast majority of data in a modern videogame is assets like textures and character models. The more detailed, the "larger" - it's just the inescapable reality of things. Better compression could help a little, but don't expect an orders of magnitude difference since (almost) everything is already compressed. Poor code optimization can lead to bad performance, not to large size on disk.

Quantum computing is not meant to be a replacement for traditional computing, it's an application specific technology that is multiple orders of magnitude better than traditional computing at those specific tasks. Currently there's no reason to believe this will happen in the near future, nor has Altman provided any.

You can find all factors of a number and filter out results that would go above the size limits. Increase the total by one at a time and use the first number that gives you a result.

I wonder if this means they're dropping EA for star wars games...

Maybe they're trying to bullshit you into believing that, but the premise is absurd. Like, actual lunacy. We already have solutions to climate change, we just don't want to actually implement them because it would reduce powerful people's profits. Trying to "tech the climate change away" by funneling money into nebulous hypothetical solutions that delay any action taken by a factor of years or decades is part of the problem. Russia and China have nukes. The US already has the military capability to easily beat both except entering a direct war with them means getting nuked. AI doesn't, wouldn't and can't ever change that situation. I also posit that if at all possible armed conflict should be avoided rather than looking for ways to win one before it's even on the table.

The manhattan project was seen as vital for winning a potentially civilization ending war (not saying that was necessarily true, but it was the perception of the US government at the time). This would do what, make the same stuff nvidia does but starting from scratch? I'm for having some competition in the space, but that money would be better invested in getting existing competitors up to speed, especially since right now the main hurdle is software and not hardware; CUDA is too widely used for others to make a dent in nvidia's market, even if the hardware were good enough.

I don't even really care whether the price could theoretically be justified in terms of hardware costs, because it just seems to be a bad product with barely any software support and few use cases. Maybe future iterations will be good, possibly even worth the exorbitant price, but right now I'd rather light my money on fire for warmth.

We went over this in your other thread - you don't know what this compiles to and, further, this is going to be scheduled by the operating system, meaning it will not always complete in the same time. I'm pretty sure the compiler will just delete your for loop if it doesn't do anything, as is the case here. The explicit goto instruction will probably not be deleted, because the variable could be used elsewhere in the code and therefore it must be incremented; it's possible the compiler will just substitute the useless goto with a simple addition, but I'm not certain.

You're looking for a profiler

I'd say the problem is more with the misleading naming that would imply these are new products when, in fact, they're just renamed old chips

Quoting from Apple: These very much look like API level restrictions, not just a pinky promise the developer makes when signing the app store's terms of service. These can still be enforced on sideloads by the operating system. Also don't be tricked into believing that Apple can somehow prevent facebook from knowing and using your email address if you use it to create your account... all this means is that iOS won't make that information available to the app. Maybe they should fix their shitty API then, rather than use it as an excuse to maintain a monopoly and block open source developers. If they don't know they will just use the app store. It's already there. Ideally Apple would bury the sideload enabling option three menus deep like android and leave it at that; if you don't know you'll just get a safe default. Heck, make it an obscure ritual like old school console cheats... I don't care. As long as the option is there for people who want it and the only barrier is knowledge, it's fine.

There have always been ways to circumvent app store restrictions with little effort. Also just because the app is sideloaded doesn't mean it suddenly gets privileged access to the operating system; things that are blocked by the API, like tracking location without consent, will still not be possible. This is the case on Android as well. Speaking of which, on Android sideloading has always been possible and yet there are almost no mainstream apps requiring their own store or a sideload - because in most cases it would just cost them a huge portion of their install base. Fortnite is one of very few apps that can afford to attempt this and it's guaranteed to still lose them a lot of users (although considering they're playing this against being forced out of the platform it's probably not going to bug them). If they're worried about that, they can just keep using the app store. Nobody cares about the epic store. They care about being able to install what they want on their device. Incidentally Apple's bullshit sideload license means it will be almost impossible to distribute apps outside of something like the epic store, so if that is what's going to happen then the fault lies solely at Apple's feet.

Further, knowing or suspecting that your messages could theoretically be accessed by snapchat is not the same as knowing for a fact that they will be, or that there are automatic filters that will alert international police based on keywords like "taliban". We're not even sure that's what happened considering it could also have been reported by someone in the chat (in which case, I'd say the person to be charged with causing a false alarm would be the one who reported it, knowing perfectly well it was a joke) - this may even be more likely considering they knew not just the name of the person but also which plane they would have been boarding, although I'm sure snapchat also collects location data. I also question the legitimacy of just mass screening people's private conversation without a warrant or prior suspicion, if that's what actually happened, and using it as evidence in a trial. I'm almost certain this is illegal where I live (we even had a big case about the police not being allowed to use intercepted phone calls as evidence, and that was with significant evidence of wrongdoing). You can't even be certain that the person using a given account or phone number is who they claim to be.

It says and/or... https://www.reuters.com/technology/apple-allow-downloads-outside-app-store-eu-with-new-fees-2024-01-25/ I don't think this complies with the spirit of the law. You can't allow sideload but only if they pay you for it. This is already stupidly expensive but if this logic holds what's to stop them from charging you a billion for each sideload, effectively making it impossible? It doesn't work for open source apps that can't afford to pay these fees. Also a huge reason for sideloading is specifically to avoid Apple gatekeeping what you can or can't install on your device... it doesn't work if your app still has to be approved by them before you can install it. With this they can still decide which apps or app stores are even allowed on the device, regardless of the fee. I don't give Tim Sweeney too much credit but at least he has some legal experience on the matter...:

I agree, I'm saying that somehow I doubt this would have caused anything to happen if he had written "gonna blow up the train lmao" at a train station.

I don't think you can in C. Once the value has overflowed or underflowed its impossible to tell the current value is not correct. The only way to prevent overflows is to check for them before running the operation or assignment. Instead of this: unsigned int a; //some code here a -= 1; do this: unsigned int a; //some code here if (a > 0) { a -= 1; } C will not prevent you from writing bad code. Some languages, like Rust, have built-in utilities to avoid this problem: https://doc.rust-lang.org/std/primitive.i32.html#method.checked_add but fundamentally the logic is the same; check that the operation would not cause an overflow before running it.

Do you not see the difference between a bomb threat being directly sent to a school or a bag being left unattended, and someone writing a joke in a text chat they assume to be private? Who's intercepting your snapchat while you board a train? At the very least, if you're going to prosecute people for causing false alarms you need to let them know what the triggers are. Personally I think intercepting everyone's private conversations, without any evidence at all of them being dangerous, is a privacy violation and is bound to cause expensive false alarms, on top of not once (at least to my knowledge) actually preventing an attack. This is literally a case where they got it wrong, 100k was spent for no reason and some guy is currently under trial for shitposting on a private messaging group. The fighter jet pilots had to subjectively judge the situation and decided there was no real threat; imagine if the pilot had made a mistake and the fighters mistook it for an attempt at dive bombing something, then shot down the passenger plane? The idea that there's no subjective judgment at play here is ridiculous. Do we send a paramedics team out every time some teenager goes "haha I can't I'm literally gonna kms dude lmao" in their group chat? Some followup is certainly warranted in this case but immediately assuming it's a fully serious bomb threat is not.

You can read the ELF header of the executable at runtime https://stackoverflow.com/a/34960487 Honestly this whole thing looks like more work than it's worth. Why do you even need to know which optimization level was used at runtime?

Or you could carry a bomb into a crowded train station and do the same if not more damage. There's no security to intercept you at any point there. I wonder why we're ok with that risk but even a single intercepted text message is enough to scramble everyone and their dog when a plane is involved. Regardless, I don't think it's at all reasonable to expect this guy would know his message could cause this reaction, so even if it is deemed necessary then it's just the cost of doing business this way.

It's a long chain of unlikely hypotheticals that I don't think the teenager in question can be held responsible for. I must stress that no terror attack has ever been prevented this way.

Again not something fighter jets can stop... As I explained this is just excessive precaution. Terrorist acts can be performed outside of airports and yet in daily life no such precautions are taken. The one thing that makes planes special is that they can be hijacked and used to, say, ram skyscrapers; this is now impossible because the pilot's cabin cannot be opened from the outside on modern planes. Airports already have, by far, the highest security level of any form of transport and the chances of you making it to a plane with explosives is extremely low; it's much easier and equally (or more) effective to stage the attack in the airport itself at that point. So yes, they can just take the risk and they do in virtually every other situation. Terror attacks are often prevented through normal investigative methods, not by randomly intercepting a single snapchat message where a teenager writes "gonna blow up the plane lmao". And that one was a scandal because it's not supposed to do that. A random terrorist wouldn't know that when not even the pilots did. If a pilot is involved you don't need an explosive to hijack the plane...

If they scrambled jet interceptors every time anyone says anything that might be considered threatening, despite no prior history of doing anything wrong and no other signs of it being a serious threat, then you'd have them in the air 24/7. The attack that caused airport security to be the way it is was carried out by people who were already known to law enforcement and could (and should) have been stopped upon checking their documents; cases of people getting through modern airport security and carrying out any kind of attack currently sit squarely at 0 as far as I know. Also I'm curious what fighter jets are supposed to do about a passenger trying to blow up the plane. If anything they're more likely to cause someone with a bomb to panic and decide to just do it. You can't even enter pilot cabins from the outside anymore so something like the WTC attack is just not possible; at worst you have a hostage situation, which again is not helped by jet fighters. Meanwhile anyone can just walk on a train and blow it up with nothing stopping them. You tell me if this makes any sense to you. Yeah, if you shout it out at the airport, of course, you're causing public distress and panic. Not so with a text chat. If it's an in depth investigation then yes, chat logs can be used as indicators and possibly evidence in a trial, but it's never a single message immediately triggering a full force response... at most I could have seen him be briefly detained on arrival.