Posts by sured 12 publicly visible posts • joined 2 Mar 2026
There is no RPi ecosystem, just various independent devices. That should equal choice and be a good thing, but modern RPi products seem like different desktop form factors. The 40pin is now there out of obligation and not application.
Today you can buy a Intel N150 (x86) and match the power usage of a RPI for any embedded project with the advantage that you can increase power consumption and performance with the N150 if you need to.
Much more sad than fumny.
It's becoming borderline wasteful when viewing the topic from Linux. I understanding wanting options but trying to support old 32bit programs isn't as productive as working on a qemu plugin for 64bit or a somerhing similar for current programs. 32bit already has a ton of support and finding a 32bit only program that doesn't work or have a better replacement is near impossible. This would be against Microsoft’s hard stance of helping Linux users, but it would help thier own Windows users.
Microsoft needs something and Ai isn't it. Windows basically has nothing anymore for developers, so much so that I truly believe that I could cover 70% of their offerings with the first VS .Net release from over 20 years ago.
A calculator is a reminder that genAI companies are making new programmers forget repeatability, A.K.A. reusable code. You won't see genAI companies go anywhere near this topic and you know why.
Also this article beats a really dead horse. The analytical types already know that mass scale genAI programming is only useful for the non-analytical types who are effectively software collagists. That's fine, that's who genAI programming is really for, so why pretend otherwise?
Generically, $RAM + $BIGGEST_PROCESS == $SWAP_SIZE
In short, I generically use the sum of the largest process the machine will handle plus the physical RAM quantity. My rationale behind that is that the machine has already ate all of its own RAM so the job is forced to swap which has caused a processing slow down too great to be overcome (losing race condition). If I'm desperate,, like with a cheap VPS, I'll set it to the size of the largest dataset. No matter what, knowing what software will be used should be considered, at least a little.
Once you hit swap it's no longer a RAM problem, it's a permanent storage + unknown data quantity problem. I've never been sure where the "RAM * X = ENOUGH" formula came from (maybe a swap file had to be a multiple/denominator of RAM way back when?). Does it matter how much RAM you _HAD_ when it's now become about how much filesystem you need,?
Do you mean standardize the terminology or in respect to the 'Z'? Moving from/to permanent storage or from/to RAM is the difference regardless of what the bytes represent. As for the 'Z' in ZRAM, it's basically for OEMs trying to lower costs by including less RAM. Now if you could also dedupe in RAM that might change things, but that would be a noticeable performance hit without new CPU instructions being added and possibly a tweaked PCIe (which the article doesn't mention).
After re-reading your comment, I think you're using 2 examples that use different context :-/. Of course it depends on your context of "swapping", but I feel like the article is using the context of 'swapon /swapfile'
You mean to surf the web right?
This 8GB advice has never been so poor and ironically it is extremely poor for anyone who pays attention to swap. For over 20 years software for CAD, 3D, 2D, NLE, profiling, etc. hasn't worked dependably with 8GB swap and anyone who uses these knows this every time the GUI freezes/seizes. Now that's just typical GUI based userland software but it's the same with software services. Anyone running a modest DB laughs at 8GB swap. Any VPS user knows that a modest GIT repo wants at least 16GB. A modest mail daemon minus the DB will eat 8GB swap with queue.
For the audience of this article, again the people who pay attention to swap, there is simply near 0 use cases where 8GB swap will be acceptable.
A “token“ has variable width and in fact time since a token might be garbage and this (you hope) the result is discarded. Really the practicality of the term is used in the same context as the “nth“, and not like CSV tokenization. This is of course O.K. for some things but IMO not for genAi _UNLESS_ you accept any output. Sadly the chances of seeing real metrics for result sets is never going to happen because companies don’t log the Plinko logic, which seems crazy but the financial cost of that really would be considerable.
I've been using AI for a little over 2 years for comments and tests and what AI generates isn't satisfactory. I rewrite 90% of them because they're exclusively (>= 98%) in 1 of 3 categories:
1. completely pointless, ie. 4 = 4 - 1 + 1
2. too vague, ie. "receives three unsigned integers and a double"
3. extremely misleading to the context, ie. "returns an integer" ... actually returning 8 bools in 1 byte
As for tests, I'm not a "professional" programmer but I feel for non-pros (or lazy types) that AI generated tests are better than nothing. If they're not great they'll pass anything which is essentially the same as not having tests. If they don't pass, then you partially rewrite them and make them partially legitimate or at least better than what AI came up with.
A switch Vs. if-else argument is kind of moot in any language as they both satisfy the logic you want and probably optimize identically.
Typescript being needed is a argument of syntax Vs. function. Is the syntax needed if you can use the benefits of Typescript without using Typescript syntax?
I do feel comparing C++ to simpler languages is legitimate for arguing the denial of features in simpler languages, same as your sentiment. Any language can be a role model.
IMO, the only people who start new Go projects are the people who are already tattooed by previous Go projects. However, for the applied areas Go currently has, Go seems a better choice than Java or .Net.
