"review the recent changes, looking for: gaps, flaws, inconsistency, or overtuning."

"oh shit! update the tests and docs too."

Based on what? What makes something wrong or right in production providing the toolkit works and exists to allow it to be widely adopted and mass produced, and the end goal of the device is achieved?

Exactly. About the only thing that matters for "in production" is "meets requirements", whatever those may be. It's pretty rare that one of those is "squeeze the maximum possible performance out of a single-cycle multi-hundred MHz MCU". Sure for mass, low cost consumer products, it's worth pending a lot to save 10 cents on the MCU but it isn't for most of us.

I've seen similar complaints before: "Arduino has no place in production,

I've seen that too. I now point them to Finder PLC. It' an industrial unit that is basically an arduino with relays etc in a nice box.

Dude.

Back in the 80s I was running basic on a machine with a tenth of the RAM and less than a thousandth of the computing power. It did not "confound the learning process". These days I also often use the Arduino environment because the performance is adequate for me needs and so it's a waste of time to slog through the datasheet and do it all by hand.

Who cares if they are a bit different from python? I don't write C++ on an MCU in the same way I'm writing it on the server because I'm writing very different things.

This really sounds like nerd snobbery to me. Modern microcontrollers are so ludicrously powerful now many people dipping into them won't need to scratch the surface of the performance on offer. And that's ok.

The very rich intend on owning serfs. People keep telling me how California is full of far left liberalism, but that is not what I've seen living in Silicon Valley for almost 30 years.

You can use non-standard extensions with C++ today to cover some of what Rust does, like -Wlifetime in C++ for something roughly equivalent to Rust's compile-time borrow checker. UBSan/ASan lets you enforce address and undefined behavior with runtime checks, it's useful and easy to use, but it has a big performance impact so you normally turn it off in production code.

Google Carbon seems promising, but it deviates enough from C++ that it feels like yet another language. But theoretically you will be able to have a project that has a mix of C++ and Carbon and it should Just Work(tm).

Ada / SPARK is an alternative to must of what Rust offers, it's a wildly different approach and not a 1:1 equivalent. But you can make reliable software and even do some formal verification and static analysis on it. It's easier to deal with violation messages from the Ada compiler than with Rust. But it's also a lot more work to put all the contracts and declarations in place (you need multiple files just to make a module, so it's a very different workflow compared to C/C++/Rust).

Coming from Pascal where write/writeln is a special language construct that doesn't work like any other function/procedure, it's refreshing to see that Rust made printing a macro, and that you can roll your own easily. That every macro ends in ! makes a lot of sense in terms of transparency, but it was a confusing little doodad in the syntax to see on my first day.

And I prefer C to C++ or Rust. And I prefer shell script, awk, or Lua to Python. So I'm a bit of an odd duck.

But Rust is pretty good. It's tedious at hell to write anything safely, but you can do it and you can have high confidence in the results. (obviously you still need a full test suite)

Thanks.

What will happen? The same that happened to the Soviet Union when it wasn't able to raise the money to compete, a collapse.

Except this time around it will be the orange shitgibbon's territory that'll fail, because:

https://www.goodreads.com/book...

Data center satellites will not be happening, same as Mars colonies. Or at least not in the next 50 years or so and by then the human race will be in a fight for survival against climate change and have other priorities.

These two things are just hallucinations of the Space Nazi on drugs, who has no clue how Physics works despite supposedly having a Physics BA. I guess he bought that one and never actually studied the Science.

Do not write off that Backyard Bomb just yet. You find the same kind of utter crazy among some Christian groups as well.

Nice example of telling people you are a blithering idiot without telling people you are a blithering idiot.

You should restrict your propaganda-lies to your cult members. Nobody else believes you.

Indeed. I wonder how the human race even made it up to here.

Yes. Unfortunately for them, they cannot actually get the open models banned.

I also highly doubt that IPO will work after the complete crap that Musk pulled.

Definitely PsyIOps.

OpenAI and Anthropic will cease to exist very fast when the stupid money dries up. All others that are burning tons of cash, for example Google and Microsoft will be at least in very serious trouble and Nvidia and Oracle and some others will probably be dead as well. and hence they try to inflate what their toy can do to gloss over the worse and worse news as to what it cannot do and what massive risks it comes with. For example, there are still no substantial revenue gains from LLM use anywhere and profitability of the LLM offers can probably not even be reached anymore for the frontier model makers.

Now, AI could actually, by accident, wipe out the human race in some scenarios. Think nuclear launch decision relying on it, for example. But that would still be on the humans that did the abysmally stupid thing, not the AI.