No one here claimed the AI models made something truly new.
The commenter flessner asserted it couldn't, despite having no way to demonstrate it. They are passing off faith as fact.
Assuming someone did want to show AI models can make new stuff...
> How could you possibly know that? Could you prove that an image wasn't copying from images in its training data?
This isn't even that hard. You just need to know what images are your training data when you train your model. A researcher with a small grant could do this.
Redbubble was sued by the Pokemon Company two months prior to the launch of Pokemon Go, so you picked the exact wrong company and moment to try this with
It's an Emscripten-compiled version of Basilisk II [1], which is closely related to SheepShaver. (The two emulators target different generations of Apple hardware - Basilisk does 68k CPUs, SheepShaver does PowerPC.)
> Speaking of game loading, it is surprisingly slow. You plug in a cartridge and turn on the 7800+, which takes a moment, and then there is a “Loading game” screen that takes 15 to 20 seconds to actually load the game.
This would have been nearly instantaneous on the original console. You could hit reset and be back at the title screen basically before you'd leaned back from pushing the button.
It is very funny that as computers have become more powerful, we've lost the incentive (and probably the skills, outside of the demoscene) to make things run fast and close to the bone.
Also, I think "disappointing joypad" is faithful to the Atari experience, having used those spongy suckers for many hours in the early 90s.
It is not funny but completely insulting to the customer.
The people selling those consoles should feel ashamed about this. It shows complete lack of care for the customer.
Any emulator can load those cartridges near instantaneous. This does not show lack of optimization but lack of complete care. It probably boots up the whole operating system or something similarly insane.
This is plastic rubbish preying on nostalgia. They can't even get the basic function of playing the games right.
You will have a better time just playing the games on a cheap laptop. It is one of the easiest systems to emulate anyway.
and yet they sell them like hot cakes. so either the customers don't care, or the customers don't even notice. younger players that are used to waiting for loading probably don't notice at all. only older people might even remember, but the younger kids won't believe them because the olds memory is probably not right.
I wish games still came with instruction manuals; then you would have something to read while you wait 15 minutes for the base game to install, followed by however long it takes to download and install 60GB worth of day 1 patches and updates.
I'm a casual gamer at most lenient definitions of my gaming time. I got tired of having an hour to kill, and half of that time being spent on updates. It pretty much was the final nail in the coffin for my interest in gaming.
Of course I'm ignoring any needed system or firmware updates. For a while Sony's updates seemed magically synchronized to my gaming schedule.
Or try being an online gamer, buying a new game or expansion, installing the game and all of its update patches, and finding that there is a multi-hour queue to log on and actually play the game.
Fortunately remote play allows you to boot up your console remotely from your smartphone, install any updates, and get into the login queue. So when you get home you can log on and play for 5 minutes before the servers shut down for maintenance.
It's a fair bit of engineering to design an interface to the original ROMs that can map them into the memory space of whatever modern hardware they're using. It's much easier to design an interface that allows you to bit-bang the data out of the ROM and snarf it into RAM. The loading game screen might've been the only way to keep costs below a certain threshold.
And I have a 2600+. Games usually take ~2-3s to load on it.
If you were to write a program to make the Atari 2600 copy its own ROM--not that A) it had the RAM and B) it ever had a need, it would very roughly take about 15-20 cycles per byte, and 15-20 more cycles per 256 bytes for looping. Heavily depends on the code used to copy it, probably less if optimized.
The 6507 is a 1Mhz cpu, and again very roughly about 64,000 microseconds to copy 4096 bytes, or twice as many to copy 8192 bytes. Under a second on the original hardware.
So even 2-3 seconds is a long time. Is the embedded OS on the 2600+ bit-banging the ROM via GPIO pins? Is it a Python script doing that from userspace (lol maybe something recycled from a ROM dumper's toolkit posted on some random forum in the early 00's)? Not bad for a script to do that in 2-3 seconds to be honest if that's what it is.
Interesting. I wonder why TFA said 15-20s? An optimized small Linux boot takes ~1s, so what is the delay from?
Still likely a lot faster than loading a PS1/2/3 game from optical disc. And miles ahead of installing a game on PS4/5 (and downloading gigabytes of patches.)
It's not clear if we're talking consistently about the sizes. If 15-20s is referring to a 7800 cartridge that's up to 48kb, that makes sense. This subthread is talking about 2600 games up to 4kb.
Also it's not clear how the console reads cartridges with bankswitching hardware. It can't really know if a cart has that. So it may just have to try all the common schemes on every cart, and read through the entire ROM space multiple times in case anything changed. That would account for the timing.
Hmm, this is disappointing and something that they should really fix.
As I understand it, the whole advantage of ROM cartridges on Atari-like systems is that no loading is required - the games run directly off the cartridge ROM. Enthusiast emulation systems can boot in a couple of seconds.
The Nintendo Switch cartridges are more like a read-only file system, but games still load quickly without a lengthy installation process and can usually run without an internet connection.
In the classic cartridge systems, you plug in the cartridge, and the data is directly accessible on the bus of the system. Some milliseconds to initialize hardware, and the game can start.
What they're doing here is dumping the cartridge, and loading it into an emulator. That, and probably the boot of whatever OS the device is running, is the "load time" for the game.
> Prior to the Internet/Web any sort of pause in software was anathema
We should really go back to the blazing speed of cassette tapes, floppy disks, and optical media.
> Now AppleOS spins for 20 or 30 seconds on file security checks.
Trying to recreate the experience of loading a program from floppy or optical drives.
I feel your pain though: syspolicyd is my sworn enemy. Not only does it cause apps to hang on startup (ugh) while it tries to phone home (no thanks), it also periodically decides it needs to rescan all huge files it can find (like game updates that I've downloaded over the years) in case they might have suddenly turned into known malware. Whenever my laptop heats up for no obvious reason, it's usually some infernal combination of evil macOS demons (not a typo), (syspolicyd, mdworker, photoanalysisd).
It's a shame that in the current malware environment syspolicyd seems to serve enough of a useful purpose (for some subset of Mac users) to be turned on by default.
In 1992, unless you were rich you had a $2,000 386SX-33 sourced from parts out of the back of Computer Shopper magazine.
Here is a 386SX-33 running Windows 3.1 (1992) launching WordPerfect for Windows 5.2 (1992) a perfect representation of what a normal, typical, average person in 1992 would have as a PC: https://www.youtube.com/watch?v=pYjrX2P4iY4
WordPerfect takes 45 seconds just to start. You dropped $350 on an ATI Wonder that has hardware acceleration of Windows 3.1 video calls just to get smooth window moving and resizing.
What about a 486? Lol you didn't have a 486, those were for CAD workstations and servers. You had a 386.
Three years later, lusting after Windows 95 and frustrated with slow performance, you shell out ANOTHER $2,000 on a Pentium 75 with 8MB of ram and a 528MB hard drive.
Sure, with the lower FSB that means you're less than half the performance of a Pentium 133 despite having 56% the clock, but you're a regular person with regular person money.
Windows 95 takes over a minute to boot, Space Cadet Pinball takes 15 seconds to launch, and you buy City Streets for Windows which takes 3-5 seconds to redraw its viewport every single time you reposition the map.
From the day you could first afford a PC you chased upgrades: more ram, a hard drive with a higher rotational speed and bigger cache, holy crap they can do 3d in hardware now, can I afford a 256kB COAST module?
Then one day, in about 2011-2012 that all stops. Sandy Bridge. SSDs. Now you only upgrade every 4-6 years: when performance doubles or windows stops supporting your CPU or you drop your laptop and break it.
And even that 386 was hot stuff in 1992. I was in high school then and stuck with a 286 with 16-color graphics and like a 40mb hard drive or something. We got a 486 right around Christmas 1993 I think, which was a mind blowing upgrade at the time. 256 colors in DOS video games and Windows could display 16k colors. CD-ROMs that had little postage stamp videos and we got one of those CD-ROM encyclopedias.
We had like 286 class machines in most of the lab my senior year of High school in 1994 with amber displays and we learned Turbo Pascal on them. The school had a handful of 486s but they were setup to run exactly the same software as the 286s, no Windows for us. They must have waited till they replaced everything to start using Windows.
In spring 1995 I spent like $250 (or was it $500?) of my summer job earnings on 4MB RAM for our 486 to get Doom to run better. The upgrade struggle was real. That was a ton of money for me and my parents thought I was absolutely nuts but eventually gave in.
When I went to college I spent essentially my entire summer's earnings on my first laptop, it was close to $4k back then and it was not exactly great!
I seem to remember my old Commodore 64 being fully booted and READY at its BASIC REPL in about a second. Shorter than it took for the CRT to warm up. A computer from 40 years ago. How we have fallen!
Cartridges were probably pretty fast to load, but I don't expect that Commodore's floppy drives were particularly fast, even if they were much faster than cassette tape.
Not sure about the commodore but atari's floppy drive was connected via a 19200 bps serial connection so even for those days it was really slow.
Both atari 8 bit and commodore had a disk drive with a trimmed down 6502 in it so you basically bought a second computer. This is why the drives cost more than the computer itself.
I was an Apple ][ (well, Franklin Ace, technically) guy back in the day, but my high school used mostly C64s. The slow speed of Commodore floppies infuriated me. And they even had their own CPU inside the drive! By all rights, they should have been faster than Apple drives which were run by the computer's CPU.
um, what? how many times would a game get to a point that would require you to remove the disk from the drive to replace with a different disk because all of the data didn't fit on one disk?
You mean the Silk Road which exposed the real IP of the web server due to misconfiguration? Tor can be compromised (run a bunch of exit nodes and do traffic correlation) but Silk Road made pretty basic mistakes.
