>>3616
It is a modern trend to dislike C. People can't write good C code so they dislike it instead, pretending their C++ abominations are any better. But I guess if you use C++ in the first place then using C style arrays simply doesn't make sense because it probably won't integrate well into the rest.

>>3616
Meanwhile, in Rust it's

// 2d array let two_d:[[i32; 2]; 3]; //3d array let three_d:[[[i32; 2]; 3]; 3];

Post edited on 28th Jan 2025, 11:58am

>>3616
Just import a tensor library and do

Matrix<int, rows, cols>

. Easy!

>>3616
>C-style arrays are discouraged to be used for whatever reason
C-style arrays have a nasty habit of decomposing into pure pointers when passed into functions, thus losing vital type information: their length. (Yes, there are ways to work around this, but they're not elegant.) std::array and similar containers don't have this issue, along with the methods that you get "for free" by using these types. Plus, they integrate well with the rest of the standard library. But if you're aware of these downsides and want to use C-style arrays anyway, then more power to you.

>Of course I could just make an alias/typedef for it, but then I would have to include this in every program, that needs it.
If you're going to need multidimensional arrays this often, you might as well build or use a library that defines such types anyway. At the very least, cloning your own single header file into a project is hardly a hassle.

I don't like them based on my admittedly limited experience. Your choices are Apple, which has a horrible, locked-down software ecosystem, and a stupid setting layout, or an endless pick of Android ones with shoddy build-quality. There's also the Surface, but that's just a low-spec Windows laptop with a gimmicky form factor.

The whole concept is questionable, and the only real advantage it has over laptops is when you're using it without a hard surface, like on your bed. It encourages simplistic user interfaces with overly large buttons, which is annoying since that also affects the desktop.

If it has a good screen might do instead of a ebook otherwise I'd rather have my Linux(tm) laptop instead.

>>3308
Do let us know when it can generate a song on the level of The Gates of Delirium that won't sound like garbage.

>>3308
Hopefully. Regarding percieved standalone potential of present models (i.e prompting only, no assistive file to mimic) generative music has far more of interest to offer, and so piques mine. I don't think it will be capable of generating anything that tops good creative direction, especially for any genre which has a heavy focus on atmosphere, lengthy progressions or production nuances (with the exception of ambient drone I suppose). Similarly I think attempting to mimic vocal music (in a single pass, for everything) is a mistake. However, for genres that are instrumental, purely monotonous in tone but quite varied yet very similar in melody composition (of which there are many) I could see it being very satisfactory, for instance: most any traditionally influenced genre (Bossa Nova[!] and celtic especially), Math rock to an extent, and dnb/jungle.

It would be made more interesting if it was capable of producing tracker formatted music, given that format is far more efficient in file size, more configurable and is more structurally specified and so easier for it to produce, and easier to reference. Though I imagine the methods used aren't suitable for doing such directly as the database is certainly raw audio, so the best one could hope for is a post-convert...
Maybe one day I'll be able to run a bgm generator that either live generates and appends a running mod file in ram or simply generates an 18h file that I can listen to throughout the day, for only 50mb and maybe 2-3 minutes of peaked resource usage, at startup.

>So what do you think the future of AI is?
Lately I have been having this hope, that machine learning models might shake up copyright and intellectual property laws. There is plentiful people, who complain about AI "stealing" their IP, and so I think it's inevitable that in some point in time, new laws are being passed, in particular, because politicians of all sides seem to be eager to implement new laws for AI as soon as possible to appease the fears and hopes of the general population. More so, I don't think copyright law as is, is possible to further maintain with AI progressing. Not only does need data from all over the internet for training sets, which is one issue, but also needs to work with data, that it got from the internet, which is the other. It does so without regards to licenses or regulations and I think it's not possible to make an AI, which uses this data with regard to licenses. Even if they tried, it would be some half-baked solution, that would still not be enough to please the copyright lobbyists. Maybe in the future it will become possible to control for licenses, EULAs and ToS, but I don't think anywhere soon this will happen.

Of course it could also happen, that the laws passed, are just applied to AI specifically and not to anything else, but I hope this gets some ball in motion, that will ultimately make copyright law impracticable.

>>3609
>I don't think copyright law as is, is possible to further maintain with AI progressing
Copyright hasn't been a barrier to current models, and even if all copyright was lifted it wouldn't make current base models much better. It's fairly well accepted that frontier labs have switched to using synthetic and scoreable data instead (e.g. math problems, code) which both solves the data quantity issue and makes it easy to train with RL. Also video is still fairly untapped I think.

>>3596
never tried ccache?

>>3602
never compiled software from release?

>>3603
you dont compile often?

>>3604
once, then i install it and wait for the next release

>>3588
>used all-in-one for very little.
all-in one what exactly?

>>3589
PC. Like this one

edit: the cpu in this is better than a celeron despite being quite bit older
https://www.amazon.com/HP-EliteOne-800-G2-FHD/dp/B08B5HHD1J

Post edited on 10th Jan 2025, 11:09am

>>3587
>Then, you spent 230EUR on a used 12TB, I'd rather have a new 10TB for that price instead.
Well, I got two 12 TB drives for 230€ total, even if they're used. In the end I only get 12 TB of storage if I'm going with redundancy, but I'd like to have redundancy in any case regardless of whether the drives are used or new, since you can't trust any drives. I bought a new 5 TB external HDD for 110€ or so a few years ago, but I'm paranoid about it failing and I plug it in as rarely as possible.
>>3589
Don't get an all-in-one, 426$ is not cheap and all-in-ones have a myriad of problems that regular PCs don't. I'll give you the sacred poorfag advice, look up used business workstations on eBay, like ThinkCentre. You can get a semi-recent PC for less than 100€ that way. A used monitor plus some cheap keyboard+mouse and headphones/speakers shouldn't cost a lot. I recently got a decent functioning computer for a family member just like this, for less than 150€ total. It sounds too good to be true but it's indeed real.

Thank you for your advice. Actually, I was half aware of the possibility to get cheap but decent hardware like this >>3591 but admittedly have no experience.

One thing to note here, forever, is that I have no confidence at all about my residence place, ever. So I - I won't claim paranoia has not left its mark here - prefer to use a laptop because it allows me to relocate immediately while loosing only the comfort of an office keyboard I got for free from a random dude and a used monitor.

As for drive failures, I just play the russian roulette, really. Let the world burn! If I could afford it, I'd simply keep some drives for backup only, but as it is, I have no choice but to stop caring. Actually some years ago I had a fit or two of anxiety/paranoia over it, but survived.

You could try using ssh-copy-id to basically make it so the devices you use to log into the server is given automatic access whilst password authentication can be completely disabled which basically prevents anybody from brute forcing into your server using a password guesser script of sorts.

Oh and also, how you're even going to get started may be something to consider. For example, some people use Static Site Engines, which have features such as automatically making lists of blogs, more consistency across pages, writing articles and such in markdown etc.

But some people might prefer to write their site without one, and instead with the aid of shell scripts for both automatically generating html files to write in or again, making lists of blogs. This although can be done using locally run scripts, apache offer something called "CGI", which allows you to run shell scripts on your apache server, which could help with automating a few repetitive tasks.

https://crippled.media/free-speech-vps-providers-put-to-the-test
I was reading this later the day and realized, that those services are all good and all in terms of anonymity, but there really isn't a way to know how reliable they are. In particular when you pay a year ahead with Monero or something, who guarantees that they don't disappear tomorrow from the face of the earth? While they don't host anything illegal, they still are advertising themselves as for high-risk websites, or websites at high risk of censorship, or anonymous this and that. I'm very well aware of how hosting providers have that legal privilege in almost every country, that they can't be hold responsible for their hosting, but if 90% of their sites hosted are illegal, there probably is a case to be made against them (see those cases of "bulletproof hosting", who got taken down for that very reason.) Personally, I am not hosting anything illegal and don't plan to, but I also don't want to dox myself to my hosting provider.

Of those on the list, I've only ever tried Flokinet and Frantec/BuyVM. Flokinet only needs and E-Mail address, but they are quite expensive for a normal VPS (the cheapest option, the Romania VPS, stopped working for no reason, and when I asked they replied I should buy the other option, which was around 10 euros + 4-5 euro for IPv4. They refunded the money however, so that's good.). Within Flokinet, I tried both the web hosting and the VPS. The webhosting was okay, albeit I couldn't cope with the CPanel nonsense, which was tedious at times. In regards of the VPS, the VPS was fast and responsive and they had the newest ISO for Debian 12, which was enough for me (and shouldn't be taken for granted, as a lot of hosting providers have outdated images). An SSH-key could be added while setting it up too. A minor thing noticed, was that they added some auto-update daemon to it, which didn't bother me.

Eventually I discontinued Flokinet, because they were too expensive (around 15 euros monthly). Following that, I looked into Frantec/BuyVM. Unlike Flokinet, they require your full legal name, address, phone number. Apparently they also check those things manually, when you pay them too. Either way,

Message too long. Click here to view the full text.

>>3583
What I do is self-host from my home internet connection using an inexpensive, renewed mini-pc, but connect that to a VPS with wireguard, and point my domain to the VPS. That way you can protect your privacy while also having full control of your data. If the VPS provider ever goes down, I can easily switch without having lost anything. This guide explains how to set up wireguard, and it can work on any distro.
https://www.digitalocean.com/community/tutorials/how-to-set-up-wireguard-on-ubuntu-20-04

>Now there is Kyun, Privex, NiceVPS and the others on the list. Do any of you have experience with those?
I've used Privex and currently use Kyun because their options are more flexible. Both are good, but I think Privex is actually more reliable.

>when I asked they replied I should buy the other option, which was around 10 euros + 4-5 euro for IPv4.
You don't need to pay extra for IPv4. Just use http://v4-frontend.netiter.com/

For a domain, you can get a subdomain for free from https://freedns.afraid.org/

Post edited on 1st Jan 2025, 1:07pm

And here's my last one for this year. It's starting to get tedious.

>>3563
寂しくなるね
I might not continue this year either, because it looks like there isn't anyone else here who's still participating.

>>3564
This solution is incorrect in the general case. It happens to work for my input but it doesn't work for all valid inputs. For example, it fails on this input:

Button A: X+2, Y+2 Button B: X+1, Y+1 Prize: X=2, Y=2

The correct answer for this should be 2. Press Button B 2 times, spending 2 tokens in total. Pressing Button A 1 time is suboptimal because it costs 1*3 tokens.

>>3341
>h264 releases will likely keep getting made forever since its the lowest common denominator
GPUs could stop supporting it like with VP8. Sure h264 is way more entrenched, but streaming is only getting more common and faster. If you're encoding everything in AV1(or something else) to begin with because of that, there's no reason to encode things in h264 for the blu-ray release. I'd say this shift will happen within 25 years.

Post edited on 2nd Feb 2024, 11:47am

>>3342
>GPUs could stop supporting it like with VP8
I don't think vp8 hwdec ever had wide adoption in the first place? Unlikely that gpus would stop supporting h264, there is so much legacy media that it's worth it for that alone.

> If you're encoding everything in AV1(or something else) to begin with because of that, there's no reason to encode things in h264 for the blu-ray release
No one is encoding av1 but hobbyists and big streaming companies right now because encode cost is too expensive, and almost no devices support av1 hwdec. It's been about 20 years since h264 was released, now most devices can hwdec h265 and yet h264 is still popular. I doubt it will go away that easily.

I think discs are going to make a come-back, at least in the enterprise space.
https://en.wikipedia.org/wiki/3D_optical_data_storage
https://phys.org/news/2024-02-3d-nanoscale-optical-disk-memory.html
https://files.catbox.moe/kf9w0a.pdf

>For the first time, researchers have demonstrated that optical data storage capacity can reach the petabit (Pb) level by extending the planar recording architecture to three dimensions with hundreds of layers, thereby breaking the optical diffraction limit barrier of the recorded spots.
>The storage capacity within the area of a DVD-sized disk can reach up to Pb level, equivalent to at least 10,000 Blu-ray disks or 100 high-capacity hard drives.

>The dataset behind GPT, which includes 5.8 billion indexed web pages and occupies about 56Pb of text, would typically require a playground area of hard drives for storage.
>However, the three-dimensional nanoscale optical disk memory can shrink this space to the size of a desktop computer, significantly reducing costs. Moreover, the energy consumption of nanoscale optical disk memory is several orders of magnitude lower than traditional methods, and its lifespan can reach up to 50–100 years.

Post edited on 10th Jul 2024, 3:00pm

Nothing has replaced discs as the physical medium for software distribution, which is why even the latest Touhou game is sold on a CD. Flash drives are far more expensive per unit, so despite being even more widely compatible, they haven't filled that role.

Looking into potential successors, either they don't really do the same thing, or they aren't even close to being available. You could sell people a download code, but the only benefit of that is the developer not having to buy discs. It's mostly worse for the consumer, although if the code comes with some nice art or something, maybe uv printed, that could compensate for that.

Although pretty cheap, RFID cards don't have nearly enough storage capacity, and their longevity leaves something to be desired. The only other things I could find are material science experiments. Definitely not cheap or widely compatible.

https://www.nature.com/articles/s41587-019-0356-z
https://www.nature.com/articles/natrevmats201670

>>3476
Which version of man pages are you looking at, mine seems to say "mirror is a superset of recursive - "--mirror" option turns on recursion and time-stamping, sets infinite recursion depth and keeps FTP directory listings. It is currently equivalent to ‘-r -N -l inf --no-remove-listing."

Which seems fairly clear. (Other gnu projects do have unfortunately terse pan pages, for some reason Gnu hates man pages and prefer using "info" doc tool, which is pretty stupid.)

>>3477
>Which version of man pages are you looking at
GNU Wget 1.21.3 2022-05-14 WGET(1) on Debian 12.
I copied the relevant section and it reads as follows:

--mirror
Turn on options suitable for mirroring. This option turns on
recursion and time-stamping, sets infinite recursion depth and
keeps FTP directory listings. It is currently equivalent to -r -N
-l inf --no-remove-listing.

I archive Youtube channels like this

yt-dlp --embed-chapters --sleep-request 5 --download-archive archive (link to video tab of channel)

>>3378
thanks to Javascript-centric web design and Mossadflare, traditional tools like wget and httrack are becoming less useful.
There are tools like Selenium and Puppeteer that let you automate browsers, but they're harder to use.

>>3509
>Ideally, my perfect messenger is a mesh network
Aside from the extra latency, intercontinental communication is nearly impossible with mesh networks. On tc's irc, like a third of the active posters are outside North America.

>>3510
Can't I p2p connect from Europe to New Zealand? As for extra latency it is naught compared to drastically improved privacy.

>>3511
P2P protocols still use existing internet infrastructure, including undersea cables and data centers. A mesh network uses its own infrastructure, which is inevitably more limited in scope.

>>3512
I meant sorta zerotier but smarter. Actually think about making world wide network on top of currently existing layer 3 where nodes are addressable by their fingerprints. You could integrate any software in it. It could still make use of trusted nodes, though. I guess there were many attempts already, all failed. Alas. If it had at least 100M users across the world it'd never go offline. It would be sorta like ipfs but general purpose.

chromium's meh but their dev console is so good i can't switch to anything else. using ungoogled chromium atm

>>3443
How is it better than firefox's?

>>443
I sometimes use Links2 in graphical mode, but most websites look pretty ugly in it, if they work at all that is. Not even TC is fully functional in it. It has a fake_firefox option, but it doesn't do much besides changing the user agent, it's not really efficient against browser fingerprinting.

Post edited on 22nd Oct 2024, 10:38am

>>3507
I tried posting with links in the terminal the other day and my experience could be described as dismal at best.