The funny thing is that not C or Rust as languages “close to the hardware” have more specific bitwise operations - but Common Lisp has:
https://www.lispworks.com/documentation/HyperSpec/Body/f_logand.htm
https://www.lispworks.com/documentation/HyperSpec/Body/f_logcou.htm
https://www.lispworks.com/documentation/HyperSpec/Body/f_boole.htm#boole
(Though Rust has at least popcnt() and count_ones(), which are immensely useful e.g. when processing small sets.)
And that’s why GenAI has chances to leave kind of a double blast crater in tech: Deceptive advertising and completely unsustainable financing, followed by equally unsustainable technical decisions and development practices.
A full-stack developer based in India, who identified himself to The Register but asked not to be named, explained that the financial software company where he’s worked for the past few months has made a concerted effort to force developers to use AI coding tools while downsizing development staff.
[…]
He also said the AI-generated code is often full of bugs. He cited one issue that occurred before his arrival that meant there was no session handling in his employer’s application, so anybody could see the data of any organization using his company’s software.
This kind of things are exactly what I see with a mid-level dev who enthusiastically tries to use GenAI in embedded development: He produces code that seems to work, but misses essential correctness features, like using correct locking in multi-threaded code. With the effect that his code is full of subtle races conditions, unexpected crashes, things that can’t work but would take months to debug because the errors are non-deterministic. He has not fully understood why locks are necessary or what Undefined Behaviour in C++ really means. For example, he does not see a problem with a function with a declared return value to not return a value (inconceivably, gcc accepts such code by default, but using the value is undefined behaviour). He resists to eliminate compiler warnings or instrument his code with -Werror -Wall.
Unfortunately, I am not in the position to fire him. He was the top developer for two years. Also, the company was quite successful in the past and has, over these successful years, developed an unhealthy high level of tolerance for technucal debt.
And more unfortunately, the company’s balance sheet is already underwater, because of extreme short-term thinking in upper management and large shifts in markets, and is unlikely to survive the resulting mess.
It is essentially an attitude and value problem. See Torvalds’ email titled “WE DON’T BREAK USER SPACE” and Rich Hickeys talk “Spec-ulation” on YouTube.
Consequently, the fix is to move to another vendor.
I do not really see that.
The article is short, and myself I like to write longer, more detailed texts. But few people nowadays have the patience to read ten, five, or three pages of text.
Also, I am becoming wary about the trolling / disinformation tactic to qualify something as AI that you do not like. If a piece of text is wrong, it will have logic failures that you can address and point to.
And said that, burn-out is a real problem, I can confirm that. Not only in FOSS software but in other fields of software development, too - but the article also cites real factors which make it worse for open source development. And it is not only a threath for the mental health of individuals, but also for the community.
And the aspect of entitlement of some users is true, too.
I swear by mulinux with Emacs as textual WM.
I know gimp, git, borg, and mutt, but not the others :)
That’s good point.
Another thing that is dangerous are CLAs or “contributor license agreements”, like Google uses. Technically, it is GPL, but Google might demand to hold all the copyright, so as the copyright holder it can change the license at a whim.
It works well for me.
Actually, I am a long-term Debian user (for 15 years) and use it in parallel with Arch, since about ten years, and I had less trouble with Arch: When upgrading from Debian 10 to 12, GNOME broke for me so that I could not log in any more. I spent a day or so to search for the cause - it is related to the user configuration but I could not figure out what it was and I had to time-box the effort, and switched to StumpWM (a tiling window manager, which I had been using before). I had no such problem with Arch, and on top of that I could just install GNOME’s PaperWM extension just to give it a try.
You could argue that my failure to upgrade was GNOME’s fault, not Debians, and in a way this is true. Especially, GNOME should not hide configuration in inscrutinable unreadable files, and of course it should parse for errors coming from backwards-compatible breaking changes.
But the thing is, for software making many small changes is very often much easier than a few big changes. For example because it is far easier to narrow down the source of a problem. So, it is likely that GNOME on Arch had the same problem between minor upgrades, and fixed it without much fuss.
But you also need to see that Arch is primarily a Desktop/end user system, while Debian is, for example, also a server system. Debian is designed for a far larger range of applications and purposes, and having many small breaking upgrades would likely not work well for these.
Oh, it’s not that bad. I use UK international with AltGr to type English, program code, and German. I don’t use the Compose key so often. Just for some extra symbols like Greek alpha or ° C.
I use “PrintScreen” for that …
I use it as the prefix key for my tiling window manager (stumpwm), and have mapped it to the “Super” X11 modifier for Emacs.
(Also, I have mapped CapsLock to the Hyper modifier, which I mostly use for user-defined commands. Not as powerful as the original space cadet keyboard, but not bad!).
BTW, one thing that is great about StumpWM is that you can define commands to script actions on GUI applications. For Example, if you are in a Firefox window, you can script Ctrl-t-B (or perhaps Hyper-B) to go to the adress bar, copy the URL, then call xsel to append the content of the buffer to a file which is called ~/bookmarks.txt, and finally open your preferred editor to add a comment.
Just an example: Of course you can use a private email service. You don’t need to give a copy of all your communications to Google Mail or outlook. Or medical data.
But what helps that, if 97% of the people you communicate with (including your doctor) use outlook or gmail, and all messages you write them are kindly stored there “for them”?
For the moment, that would not be enforceable in respect to people with technical knowledge. Enforcing it would require authoritarian control and even China’s Great Firewall has way to circumvent it.
On the other hand, this is already far more difficult than you might think. You could not install such an app from a server authenticated with TLS because the TLS keys might be subverted - the certification chain has national institutions as the top certificate authorities. You would also not be able to install such an app on an Android phone because Google has decided it needs developer attestation to install apps in a way accesible to end users. You can run Linux now but if all that is taken seriously, your options to run Linux might become limited. E.g. you already can’t run many banking apps on phones with user-controlled OS software. Railway apps like the German one already don’t work. In future, you might not even be able to use a municipial library’s or bookstore’s website this way.
But more to the point, the real application case for this kind of civil rights is not some nerd kids which want to play DnD or minecraft on their own server or test their self-written IRC service. The real application case is what we see in the US, people being dragged out of their house and disappearing just because of their ancestry, how they look, being poor or the area they live in. They don’t have time to compile software or configure port-knocking protocols.
Somebody has called these systems of “democratic” mass surveillance uncovered by Snowden “Turnkey Dictatorship” . I for sure wish they would have been wrong.
Why is this specifically relevant to Linux users?
Well,
To sum up, this is a massive transfer of control.
And that’s why things like PaperWM or niri might be a good compromise on the spectrum between “powerful but complex” and “simple but limited”.
And how much complexity is good for one depends also on the area of application. I use Rust for programming which is complex for sure, but when I have to scan a document, I use “simple-scan” which does exactly one thing, and very well.
They key is repetition, and this means it can be easier to go “all in” and learn, say, only six or eight keyboard chords from stumpwm than to use Xfce with mouse and i3 and more stuff, because the latter is ultimately more complex and requires more things that need to be memorized.
There is a learning program called Anki which is great for repeating learned stuff, it was made for language learning but I’ve used it also for a job where I had to learn like one hundred three-letter acronyms. It can be very helpful but it won’t help if one does not use the learned stuff.
To me, freedom is about the mode of production. I think people would be actually free if the very act of creating something were fulfilling on its own because of its creative manner. In that case you wouldn’t need anything in exhange, and distributing your work for free wouldn’t be a sacrifice, so there would be no problem if somebody decided to sell it.
I think exactly here is the crucial difference to the GPL and the rights it is concerned about: The GPL is concerned with the rights of the users. The reason for this is that closed-source and non-free software turns into a means of control that affects the sphere and rights of the users. A few examples:
these problems are what the GPL and copyleft licenses address, and the reason why systems like Linux are much more user-friendly.
Oh, and in respect to the artists: Yes, many do art because they need to do that. And this is all the time blatantly exploited by companies. And companies try to exploit open source developers in the same way.
GPL forces all contributions to stay open-source […]
I am somewhat tripping here over the word “force”.
Do you want to say that non-commercial software developers should give away the fruits of their work for free, in general?
But companies not?
Do they have a kind of right to that? What would that right be based on?
To make a picture: Let’s say we have a farmer who goes to the farmers market and sells his stuff. Does his offer to sell his vegetables violate anyone’s rights? Or freedoms? Is there any right to go and take his stuff away without paying?
Now, let’s say the farmer realizes that there are poor people which are starving, and because he is from a culture in which sharing has a very high value and people help each other all the time, he decides that people can take some of his vegetables for free.
And then he observes that there are companies which take all his free stuff and sell it again to poor people. And he thinks “wait a minute”, and writes a contract stipulating that anyone who takes his free stuff, agrees to not sell it to other persons, but to only give it for free, too. Anyone who takes his free stuff must sign that contract first.
Does existence of this contract make others unfree? Or forces them to do something ? In exactly the same sense in which you were using the word “force” above?
And still, we are only scratching the surface of the issue, since the GPL is not at all about the rights of software authors - it is a contract which uses these rights - but actually it is about the rights of the software users.
I use Zim Wiki + git. Or gollum wiki, which also uses git. Works both very well and can not only be synced but automatically merged.