The real deal y0

  • 2 Posts
  • 231 Comments
Joined 1 year ago
cake
Cake day: July 16th, 2023

help-circle











  • He isnt wrong. This comes from somebody who technically uses ai daily to help develop ( github copilot in visual studio to assist in code prediction based on the code base of the solution ), but AI is marketed even worse than blockchain back in 2017. Its everywhere, in every product, even if it doesnt have ai or has nothing to do with it. Monitor ai shit? Mouse with ai? Hell, ive seen a sketch of a fucking toaster with ‘ai’.
    There is shit like microsoft recall, apple intelligence, bing co pilot, office co pilot, …
    All of those are just… Nothing special or useful. There are also chatbots which bring nothing new to the table either.
    Everyone and everything wants to market there stuff with ai and its disgusting.
    Does that mean that current ai tech cant bring anything to the table? No, it totally can, but 90% of ai stuff out there is, just like linus says, marketing bullshit.



  • As a software developer i know what iterative development means, its in our blood and brains ( or at least it should be ). Simulations can indeed only get you so far, and i agree sometimes you have to make things and take a plunge. However, and i would like to be really wrong here so correct me if im wrong, but other companies like nasa, do not just shoot shit up in space and hope for the best. They arent allowed to do so for a reason. They test and calculate everything very rigoursly to make sure itll hold up as expected. From thruster power, resistance to continues extreme heat from reentry, …
    All of that they do here, on earth, before shooting anything up into space. Otherwise things like the rover on mars would have needed like 20 tries instead of 2.

    These are things that looks like spacex is just throwing out the window.
    To take it back to software development, they are doing an iterative development ( which is very good for what they are doing! ) but their testing before production/release of software is so basic theyll just see how it responds out there. Thats a huge nono to me if youre going to end up crashing all those rockets in the sea killing a shit ton of nature in the process. Sometimes the means dont justify the costs to me, and this is one of them. Yes, the booster catching was nice to see ( eventhough it nearly ended badly ) and its idea is very good and needed, but the way to get there is…messy.


  • Specially this. How space x handles failures is a very hard nono in my book. “But we test in the field” is what space x says, and as a software developer its like saying “we test in production”.
    Yes youll get something use able faster, but its way way more costly in the long run and is nasty in between.
    My arse they cant test this stuff on earth. We have simulations, models, calculations, test, everything. Yes, things can and will sometimes still fail when going in production ( in flight ) but you want to lower the risk of it failing cause its costly as fuck.

    They dont seem to care though.

    Also, im not saying what they are building towards is bad, it really really isnt, but their methods is… Bad






  • I think you have things wrong. Any other languages can have libraries be distributed as some format that would allow applications to use it, be it linux/gcc and .a files ( which are actually archives with elf/object files of the code ), or a full on library like .so/.dll.
    Rust can only do .o/.dll and only have it expose like a c library afaik. Even .net has improved on the .dll and includes all its language features in it. Rust has none of that. Its not true that libraries not rebuilding are only for closed source. Its also ease of use/access and less problem prone. What if i build my library using a different version of the compiler than you and your application? I could have no problems building my library, while you cant build your application because the library i made gets rebuild and errors.
    These errors happen and are all because there is no stable interface/abi and all other languages have overcome this.

    Also, by default, nothing in c is rebuild unless it needs to. Thats why the intermediate .o ( elf object ) files exist, so it only has to do the relinking and not recompile and thats why .a archive/libraries in c work, because it doesnt recompile. Unless you meant the fact rust can rebuild part of a file, without recompiling it completely?

    I think you dont fully understand how c compilers ( gcc specifically ) work when using multi file projects ( and not just doing gcc input.c -o output.exe ) just how i dont fully know how the rust compiler works. Also, anything using IL will always have an abi, because how else will it jump from code to IL code, so its obvious that rust to wasm will have to abide by that haha. Be it c wasm, c# wasm or rust wasm calling one another. Wasm is wasm, and you only need an exposed interface to call or include the other wasm ( c#/blazor having NativeFileReference in the csproj )

    Again, i like the idea of rust, but it has a long way to go to be viable atm. And it has many pitfalls to avoid so it doesnt become the hot mess that is any framework based on node.js