It's all correct and also pointless and irrelevant. It all boils down to "Rust doesn't fix every possible problem so why not stick with something that still has every possible problem?"

In other words, they're upset that a new thing isn't popular so they're trying to think of any argument against it, but none of their arguments are relevant. Yes, you can still write bugs in Rust; of course you can. What you can't do is write memory safety bugs in Rust, which are a huge proportion of security bugs that occur. Rust gives you a massive decrease in attack surface automatically.

This is ignoring the ecosystem, which is full of better tooling, better library packaging, better testing, and just in general an overall better language, but instead of trying to argue the language on its merits they're trying to argue that it's not perfect so why bother.

I've also heard the same arguments about C++; 'anything you can do in C++ you can do in C!', which is technically true but ignores the fact that if I want to do something C++ does it usually makes more sense to use C++ to do it rather than e.g. trying to hack the concept of objects, private methods, templates, etc. into C myself.

1. You don't know what you're doing - everyone makes mistakes.

2. You can still write bugs in Rust but the point is you are far less likely to.

Yeah but 1000x less mistake ????? I mean these people behind android project is atleast one of the better engineer but jesus christ if they can improve so much then I dont know how much average joe can benefit from that

It's not 1000x fewer mistakes overall, it's 1000x fewer mistakes of this one specific family that Rust is designed to eliminate.

They've also seen improvements in developer confidence and onboarding time, but not to the same degree.

>You don't know what you're doing - everyone makes mistakes.

I mean if you don't know what you are doing you are going to make mistakes that go beyond memory safety. Look at Log4shell for example.

The important thing is the likelihood of mistakes getting past the compiler. According to Google's numbers the likelihood for memory safety reduces by several orders of magnitude, and the likelihood for other kinds of mistakes reduces by a factor of ~4 (depending on how you interpret their numbers).

Just saying "but you can still make mistakes" is dumb and irrelevant and it's kind of disappointing that it's such a commonly bandied non-argument that Google still had to address it in this post.

On the sense that you can write a Rust compiler in C and use it to program your software in a better language, yes, all of that is correct.

It isn’t wrong, it’s misguided. You can write the same code in a Turing machine, too.

Sure, but the discussion against Rust is basically saying that you can have the same features without the downsides of using Rust, which is objectively true. For example, memory analyzers like valgrind work great.

If you assume that the only benefit of using rust is built-in valgrind, sure, but there's more to the language than preventing off-by-one errors.

If people don't want to use the language then that's fine, no problem, but a lot of people do want to use the language because it's just a great language to use as well as having memory safety.