Why should we wait for LLMs to become good enough to spoon feed the end result when we can do it ourselves?
These are beautiful, collaborative efforts and require members to constantly improve their learning, communication, patience, and technical skills. These are skills that transfer to other parts of life, same as other hobbies like sports and martial arts. It's an incredible manifestation of the hacker spirit. Do we as a community not enjoy tinkering anymore?
This isn't a consumable product. We aren't selling anything here trying to make a quick buck. There's no "rush to market" pressure to finish these projects.
The perspective I see creeping up in many engineering areas of "wait for AI to automate everything in life" seems to value the end product more than the means.
It's a valid question. The short answer to the question is counterproductive in my opinion.
Can you please elaborate on why distributing decompiled source code is violating copyright? And Nintendo is litigious, if it was black and white surely they would have taken down everything by now? How do you know the major LLMs are not trained on these decomp projects and/or leaks already?
Of course any random C file is going to have pointers. Where can anything in the linked code be exploited? It seems like they're testing for bad input data with asserts to catch bugs in some functions, and properly validating bad inputs in others. Just because they're writing C doesn't mean it's vulnerable.
How can you claim nobody cares about 5%? A 5% performance increase is significant. And video decoding is not always for playback, where 5% may not matter as much.
On the main page, it took me a minute after laughing at "std::byte arithmetic" to realize it was April 1st. All of those C++29 "features" are on point, very funny. Though surely there's somewhere SFINAE can be mentioned...
This website https://www.learncpp.com/, while it may look like other half-baked tutorial websites, is excellent is filled with up to date nuanced information. It covers basically everything you will need. It's extremely high quality.
Even if you don't follow it step by step sequentially, it's an excellent reference if learning from another resource.
This + actually working on a small project you want to accomplish is probably the best way to learn.
RISC-V is supported on QEMU. The available devices don't have a ton of peripherals compared to aarch64, but it exists. Even FreeRTOS has a QEMU virt port for RISC-V. And if you have unit tests QEMU could easily run those accurately.
That's like the car you bought in 2022 to run on the fuel used in the year 2500. Whether it's annoying or not, that it does or doesn't run on new technology, doesn't matter. You paid for a product, not upgrades (unless you did pay for upgrades, of course). You shouldn't expect free upgrades because they certainly were not free to create.
Well, it makes them no more money from me. My maximum value has been extracted. Whether that's a solid business plan or not is not my argument. I'm saying that if I paid for something I don't expect any changes at all. If good changes happen, great. If bad changes happen, I should be able to decide if I can retain the product that I paid for. Whether the changes are free or not is not important because changing anything at all would be make the thing a new unique product. I just want the choice to keep it the same, something subscriptions don't offer. Because, well, that's what a subscription service means. But if I bought a newspaper subscription, I wouldn't expect them to fix a typo after they already delivered a newspaper. Typos happen, bugs happen, I don't buy a product expecting perfection. I expect to get to use the product I paid for, for the time agreed upon.
In my opinion you shouldn't expect future updates of any kind from software that you own. I wouldn't expect them to keep updating it because it obviously makes them no money. I would expect them to release a new product to keep making money.
Are you able to perform a good majority of functions without taking your eyes off the road over 95% of the time? In my perfect setup I would have configurable buttons to map critical functions to and then a touch screen for non-important things. I have a very hard time seeing how a touch screen alone would be better than touch screen + a couple buttons.
These are beautiful, collaborative efforts and require members to constantly improve their learning, communication, patience, and technical skills. These are skills that transfer to other parts of life, same as other hobbies like sports and martial arts. It's an incredible manifestation of the hacker spirit. Do we as a community not enjoy tinkering anymore?
This isn't a consumable product. We aren't selling anything here trying to make a quick buck. There's no "rush to market" pressure to finish these projects.
The perspective I see creeping up in many engineering areas of "wait for AI to automate everything in life" seems to value the end product more than the means.
It's a valid question. The short answer to the question is counterproductive in my opinion.