SnowFox

@V_Mags_Fox@meow.social @eniko@mastodon.gamedev.place LLM pushers seem to think the coding bits are boring too! The "interesting" parts appear to be coming up with the idea and taking credit for the output of the plagiarism machine.

@MachineLordZero Also the decades of the media going "sperm counts are going down and nobody knows why!" but the moment a study links it to microplastic? Crickets. It's not something they can blame on "individual responsibility" (fast food, lack of exercise, whatever is considered "unhealthy"), and will instead require governments to force industry change, so suddenly it's no longer important to address. Somehow.

@mast0d0nphan@beige.party @NanoRaptor@bitbang.social Nah, that's nothing to be ashamed about.

@phooky@hexa.club @NanoRaptor@bitbang.social All the more reason to do it for a different reason!

@ana Perhaps related to the long tradition of referring to boats/airplanes/cars as "she"?

@0xabad1dea@infosec.exchange AIUI a variant of this is essentially why "strafe jumping" works in Quake engine games: Your velocity is capped in the "forward" direction, but by alternating "forward" between forward-left and forward-right you can theoretically accelerate to almost √2 of your normal run speed. (And the jumping lets you touch the ground briefly and switch directions while in the air, and I guess the turning lets you tune the forward-left/forward-right angles.)

@DosFox@tech.lgbt I wonder how big a logic analyser you would need to reverse-engineer the GLUE? IIRC it mostly does stuff like operating chip selects and implementing bus protocols so you might be able to replace it with a few GALs.

@r Is it giving the board designer the middle finger?

@colinstu Haven't used them, but the NiMH ones are often 7x1.2 V, whereas I assume the ones with a USB port use LiPo and possibly a boost converter. (Not sure if there are LiPo ones that work on a standard NiMH charger.) It also wouldn't surprise me if the space used by the USB port/charge circuitry was less than the space wasted by/between alkaline/NiMH cell packaging. What's the use case? I haven't looked into them because the things we have that take 9V (multimeter, smoke alarm) benefit enough from low self-discharge of alkalines that rechargeables don't seem worth it.

@NanoRaptor@bitbang.social Tangent that's been bugging me: Do you ever wish governments would make the forms they expect to receive internationally 210 × 279.4 so they print nicely on both A4 and US Letter?

@annarcana @pikhq IME that depends on the company (namely the type of work/how expressive they let you be). That said, I found it helped to have personal projects that were exciting and completely unrelated to what I did at work. AIUI one of the marketed features of Java (to management, not programmers/users) was to be able to treat programmers as interchangeable — sadly I don't have a citation, but it explains the lack of features like goto and operator overloading. With that in mind it's unsurprising that despite a decent amount of experience with Java for school/work don't find it enjoyable. (On the flipside, people who do programming as "just a job" shouldn't be looked down on, but even if LLMs weren't unethical, LLM-generated code lacks the professionalism I'd expect.)

@mikemccaffrey@drupal.community I fundamentally disagree with the idea that we should not give users autonomy over their own devices based on our own preconceptions of "average user". There are many, many ways to encourage people to automatically update (or nag them to manually update), iOS does a lot of them (big button to enable updates during setup, nag when an update is available, nag some more periodically, nag to enable automatic updates after applying an update), and as much as it annoys me it *works* (Apple's update rates were incredibly high when I last checked) and still gives people the option to say no. And it doesn't result in people jumping through hoops to disable updates entirely. And frankly, for a lot of Windows users, the main thing they need to keep updated is their browser which can be done without a reboot. (And Notepad, apparently, but that can also be done without a reboot.)

@foone@digipres.club I was annoyed (as a non-Windows user!) when I learned they added this, which (ironically) was because I was reading the features of https://schneegans.de/windows/unattend-generator/ : "Prevent Windows Update from rebooting your computer: This creates a scheduled task that periodically moves your active hours, tricking Windows into thinking your device is in use all the time." Maybe that's an option?

@whitequark And (checking the datasheet) apparently a log2() accelerator? That seems like an oddly-specific thing to accelerate.

@stereo @pikhq The thing about the latter case is that stuff like (IIRC) "a short sequence of x86 instructions that does the same thing as a slightly longer sequence" or "matrix multiplications in one fewer scalar multiplication" are things you can formally (or not-so-formally) verify without a bunch of testing, and I think the results would be much less publishable if they were merely tested (unless the input space is small enough that you can test to exhaustion). AIUI those are also the type of things that don't require (and possibly don't benefit from) a NN trained unethically.

@foone I've seen some rubber figurines like that: Deep grooves between areas of different colour. I'm guessing that's to hide the join between colours somewhat so inaccuracies are less visible (you can still see it on the mouse, especially how not all the lines are "drawn" in black). The problem (when used on the figurines) is it makes a weak spot so extremities break off easily.

@cholling@bytes.programming.dev @NanoRaptor@bitbang.social @pileofgarbage@mastodon.social Not quite! The function keys are still physical keys.

@foone@digipres.club You mean a good pun.

@mikemccaffrey@drupal.community @foone@digipres.club That would be fine if it was opt-in, but it's opt-out and the officially supported opt-out costs money.

@CodingItWrong@oldbytes.space @NanoRaptor@bitbang.social https://ultrasciencelabs.com/lab-notes/why-we-are-still-using-88x31-buttons speculates that it might ba have been the sample size of Netscape Now buttons (the official button size was supposedly 88×32), but there's not that much to go on. (If that's the case, then I'd guess it was sized around the existing app icon and the text they wanted on the button?)

@nicolas17@social.treehouse.systems @iBleedIn6Colors@mas.to @rmondello@hachyderm.io @siracusa@mastodon.social That assumes the design is sound and that the attacker can't guess your passcode or defeat HSM security measures (both of which could be done by a government-funded attacker with access to CCTV or billions to spend on attacking HSMs). Or even: If I go into an Apple store and need to passcode-unlock my phone then they likely have it on CCTV, and then Apple absolutely *can* read my passwords. It's a billion times better than Google/Chrome's "password manager" (at least without a "custom sync passphrase") but it's still something I would prefer not to have to trust. It is not unreasonable to want to keep passwords somewhere that would require a search warrant in the owner's jurisdiction to access.

@robinhood@yiff.life Related: Going through *all* of the settings after installing new software, and periodically afterwards/after upgrades.