setopt

> are there solutions for communication when international links are cut?

How far can you scale wireless mesh nets?

On both my Fedora laptop and on my MacBook, I have to enter a password at boot to unlock the hard drive, and both had similar complexity to set it up (in Fedora I clicked a button and entered a password during the install, on Mac I clicked a button and wrote down a backup key after the install).

I think this is specifically for the TPM2 integration, which isn’t required for FDE. But I assume it makes it smoother. System updates that requires multiple reboots for example are more annoying on Fedora, since every reboot requires re-entering the password, while on Mac I just login after the final reboot.

Thanks to the OBS devs!

I work at a university and occasionally make video lectures for my classes. I used to do it in Zoom, but it sometimes had weird glitches where it just… deletes the videos I recorded. It’s happened several times that I lost an hour or more of recordings and had to redo it, once it deleted all the recordings that I did that week. Also, it was a bit of work to position screen sharing vs my webcam feed in a way that looked decent. I then tried a bunch of other apps, most of them had issues with latency, laggy video, or stuttering, despite playing around with the settings a lot.

Finally tried OBS again, I bounced off it in the past because it seemed overcomplicated at first glance. But it’s FOSS, makes excellent video quality, and after some initial setup of my scenes it is very easy to make new videos. And most importantly, it has so far never lost any of my data.

That makes sense actually. Re-sub videos when they cross a million views for example.

AI != GPT

> I associate numbers with taste (2 and 5 are sweeter, 3 is salty, 7 is bitter and so on)

That sounds like textbook "synesthesia" to me? I don’t associate numbers with any taste, smell, or color, and would wager that this is more common than the opposite.

Within Europe, you mean? A quick Google suggest there were native rabbits also in the Americas, Asia, and Africa.

I don’t use Helix myself, but I’ve tried it before.

I believe I’ve heard many users requesting a plug-in system (there were some plans for Scheme?), and that the devs themselves considered many features as more appropriate for plugins than first-party support (code folding and non-LSP snippets, if I recall correctly?). I don’t know how Helix developed the last couple of years, but if features like that didn’t materialize yet, that sounds like sufficient impetus for a fork to me.

That’s not just a US boomer thing, you’ll see the same in many Arabic countries for example, across generations.

> Always showing vital elements like the scrollbar isn't a skeuomorphic choice, it's always a good idea, they only started hiding them because it looked 'cleaner' to some UI designer.

There’s a lot of modern design in not happy with, but this is not personally one of them. If I can’t autohide scrollbars I try to turn them permanently off instead of permanently on. For some reason I find them very distracting, and since I anyway use a scroll wheel or trackpad or keybindings, I don’t need it to scroll. (If it’s on autohide, I use it as a progress bar when actually scrolling.).

But there’s a lot of these weird things that seemingly everyone else accepts that just really throws off my focus. Blinking text cursor is another one, I just can’t get any work done until the blinking is off. If an app or OS can’t do that, I won’t use it.

Also with FastMail. (I don’t use them but they’re on the top of my list next time I decide to switch.)

https://www.fastmail.com/blog/push-email-now-available-in-io...

Not sure what the difference is then. I use iCloud Mail and didn’t have problems with search on iOS or MacOS, I did however have plenty of issues with search when syncing my work Outlook email via Mail.app, so I assumed there was something special about iCloud (similarly to how Mail.app offers push email only for iCloud and FastMail). I’ve imported decades of email into my account so it’s not about a small email volume either.

It depends on your provider, unfortunately. If you use their iCloud mail service, search works fine – it appears to have some way to do server-side search for their own email.

I don’t know what most do, but I certainly prefer the default Mail.app over the vendor-specific mail apps.

If you’re already using one of the major providers (GMail, Outlook365, etc.), there isn’t much to configure; you go to Mail Accounts, click your provider like Google or Microsoft, and login on their website to authenticate.

I think Flatpak is a good choice for most users. The main issue is that Ubuntu is perhaps still the largest Linux distro and doesn’t come with Flatpak installed by default.

One good alternative is AppImage. Works on every distro out of the box, but doesn’t have the permission system of Flatpak.

Thanks for the pointers, appreciate it and will look into it!

Yeah, my understanding is also that working with sparse matrices inherently requires heavy branching, which translates badly to GPU. I usually do everything I can via dense matrices via CUDA, but switch to sparse matrices on CPU if I run out of VRAM…

Interesting. Parallelism and profiling are both strong selling points.

I’ve done pretty low-level C++ and Fortran in the past, but these days I’m honestly mostly used Python with either NumPy or CuPy for calculations, so not very low-level at all. My code is pretty performance-sensitive though, so unless the main operations can be framed neatly in terms of NumPy or CuPy primitives, one had to drop down to low level.

Do you know how the Go library support is for typical scientific computing stuff, e.g. matrix diagonalization, sparse matrices, or handing off such calculations to CUDA (or other GPU frameworks)? Is there a strong numerics community in Go these days, or would you have to implement most of what you need yourself?

I can imagine that for web dev. But as someone doing numerical physics simulations, would you really recommend Go over either C/C++/Fortran for low-level, Python for high-level, or Julia I guess for both categories?

I sometimes forget the meaning of data in English. In my language, a computer is a «data machine», so a «data center» is obviously just a center of computing, which would have fit.

Off topic, but where’s the border between a "small large language model" and a "small language model"?

FWIW, I enabled that feature now, and it still answer based on discussions in other projects. (It’s pretty obvious given the very different contents of each project.) Perhaps it might be a caching issue, that it somehow doesn’t rebuild its memory just because you flip a switch, but just stops learning new things from other projects?

TIL. There is indeed a switch in the settings, which for me is set to "default memory".

Not sure if it always asked for this and I just forgot, I made all my current projects back when the feature was new.

Not sure how it works in Claude, but ChatGPT projects certainly leak memories between each other.

Regarding lazy mode, I recall ChatGPT sometimes almost refusing to do a web search despite me asking explicitly for it, instead replying with speculation about what the search results likely would tell us. If I pretend to be angry that it didn’t search the web it would however do it. Haven’t noticed this in a while either.

Good examples.

When I switched to Mac in ~2020, the Mac users I knew all told me to avoid Time Machine – at least two had issues where they needed their backups, only to find that the entirety of their backups had been silently corrupted. (I used Arq then based on online recommendations, now I use Vorta on my Mac just because that’s what I use on Linux.)

Spotlight was near-unusable when I switched as well. Often 5+ sec delays from pressing Cmd-Space to the thing actually popping up, and it would sometimes reorder entries right before I clicked Enter, causing me to open the wrong thing. Alfred uses the same Spotlight index AFAIK, but lightning fast.

Those are both things that I agree should… Just work.

To be a bit critical, haven’t Apple software and hardware quality been getting worse for years even before the new CEO?

Apple Silicon was the exception, that was truly transformative. But other things, like how polished and snappy the OS feels, seems to have been getting steadily worse.

I’m not convinced that was the issue in this case, as the battery still lasted all day with my normal usage, and the phone became slow overnight after a major OS upgrade.

I had my iPhone XR (bought 2019) until last year. At that point an upgrade made it so slow it was barely usable, while the lack of further updates made it impossible to install or upgrade apps after the device and OS went EOL.

Shouldn’t the difference be like manslaughter vs murder, in that intent matters? Accidental hacking on this scale is a somewhat new problem, no?

There is also incentives to adapt the message to the outlet. If you send that data to peer review and say we found 40% increase, the reviewers would reject it, so they have to moderate themselves. But if you send a summary to the university’s outreach outlet saying that we found something or nothing we don’t know, then they would also reject putting it out. So even for the authors, the incentive is to send a careful conclusion to peer review and an overblown one to popsci.

Btw, I also think a 95% confidence interval is just the wrong statistic to look at given that data, and that they could probably have analyzed it better.