This question is less hypothetical than it seems. In recent years, there has been amazing breakthroughs in low-light photography with machine learning, such as [1]. It's not hard to imagine that it would be applied to future surveillance cameras.
I imagine that in most cases, the "AI" would work just fine at guessing things. But important edge cases exist, including the possibility of adversarial machine learning spoofing attacks to create false images. Imagine faking a crime scene this way, it would something straight from cyberpunk fiction.
> Confidently exclaiming that downscaling, gaussian blur, and taking a picture of screen showing that from far away can easily be restored by deconvolution...
I don't think HN (or Reddit) posters are arguing that deconvolution was actually being used, or that it would be practical to do so. Instead, they're just nitpicking about the technical correctness of this particular statement - "the (ir)reversibility of a perfect digital gaussian blur" - even if it doesn't affect the conclusion of the experiment.
It's just the standard nerd-sniping as seen in all tech communities... If the debate on whether you can "reverse a photo of a gaussian blur taken by a camera from the monitor" became hot enough, eventually someone may even spend a weekend to code a prototype to show how it actually works better than most people's imagination, with the solely purpose to win an argument on the Internet.
This problem is not new. In 2019, Huawei introduced a special image processing feature in its smartphone camera app, the "Moon Mode" (opt-in). Missing details are added to the moon photos via machine learning inference from a pre-trained model. Huawei then started marketing these processed images as a showcase of its new smartphone's photography performance. In China, it was widely criticized by tech reviewers [1][2] as misleading, and "Moon Mode" became a running gag among tech enthusiasts for a while.
It seems that Samsung simply adopted the same tactic to compete...
On the Huawei "Moon Mode" controversy, one can even find a research paper [3] published in a peer-reviewed social studies (!) journal, Media, Culture & Society:
> This is where the controversy began: Chinese tech critic Wang’s (2019) posting on Weibo, the Chinese equivalent of Twitter, made quite a splash. In his post, Wang put forward a shocking argument: he said that Huawei’s Moon Mode actually photoshops moon images. He contended that, based on his self-conducted experiments, the system ‘paints in pre-existing imagery’ onto photographed takes, re-constructing details that are not captured in the original shots. Huawei immediately refuted these claims, stressing that the Moon Mode system ‘operates on the same principle as other Master AI modes that recognize and optimize details within an image to help individuals take better photo'
The author showed it at the end of the article. It's the "effective speed" reported by some benchmark programs (including calling subroutines, running for loops, iterating on a string, etc). These are simple and trivial programs and can be highly optimized in a simulator on modern x86_64. Real-world programs, like games, is slower, as acknowledged in the article.
Better title: Clocking a 6502 Simulator to 15 GHz. There are multiple efforts to recreate the physical 6502 CPU on modern hardware, this is not one of them and should not be confused with that.
Wow, they are using conventional discrete log cryptography with 256-bit key? What the actual fxxk?! It's literally the equivalent of using 256-bit Diffie-Hellman, or RSA-256, DSA-256, dammit, even NSA's bad DH_EXPORT ciphers use 512-bit. Totally unbelievable.
Yes. I guess the fact C doesn't have advanced features like generics and templates probably have made the problem worse, making it difficult to create a "drop-in" API.
Creating a lightweight library would make C programming much more productive, and I'm sure many have tried, but I wonder why none of them see widespread applications, instead of writing some half-baked, ad-hoc helper functions.
One reason, I guess, is the diverse range of applications of C, another reason is the lack of advanced features like generics and templates.
But it would still be useful to create a simple library, intended for Unix-like platforms (beyond the BSD extension of stdlib). So it's more of a cultural problem?
> Pathfinding is actually a murderously complex programming challenge—one that, depending on the path, might be NP-complete.
> The "eureka!" moment came when Castle and team realized they didn't have to build perfect—or even really great—pathfinding. Rather, they had to build pathfinding that didn't look stupid to the player.
> The most interesting thing, at least to me, is that pathfinding is still a difficult problem even for modern games, and simply throwing more CPU power at the problem doesn't really help that much—such is the curse of NP-complete problems.
It reminds me of my experience of some pathfinding bugs in OpenRA (free and open source Command & Conquer clone): one day I was playing OpenRA on Linux, and I find the game was running extremely slowly like a snail, the more I played the more slow it was. So I decided to check GitHub, it turned out that AI has built a huge number of soldiers, but unfortunately, due to geography features of the map and the battles, a lot of them have trapped inside "islands" with no available path connected to the rest of the world, or they want to attack targets inside "islands". One example is isolated waterways, another is land with destroyed bridges.
But edge cases like these were still not handled well in the OpenRA engine, the AI would keep ordering those units to move and attack. OpenRA uses a variant of the A* algorithm, and those AI orders will force the engine to run a huge number of A* pathfinding attempts on every second and exhaust most of your CPU time.
The solution should make the AI to detect these endless attempts and "bail out". The developers have already fixed some of those edge cases, but there are still a few to hunt down.