there are eight billion of us. Trying to be significant is only gonna make you unhappy. The most you should be doing about fame is preparing for it a little bit on the off chance it happens to you, the same way you prepare for a car crash by putting your seatbelt on.

More important to me is that most code people write with LLMs should not be written, and instead an abstraction should be developed for it.

@idkhow I came home a month ago, I'd been applying throughout the year before that

if you don't care that much and your derivative work isn't important, you can risk that a mature hobbyist who publishes under copyleft may just let it slide and not sue you.

The only difference between copyleft and PD is that PD can make someone else a lot of money without my consent. That whole sentence. My own licensing doesn't prevent me from issuing separate licences to anyone who cares enough to send me an email with a pitch.

The first job I applied to in Hungary, after a quick and basic online test I was hired. In the interview we pretty much just chatted about my interests and the project, and negotiated the salary.

I spent a year applying in London before I gave up and fucked off back to Hungary last month.

@Demolishun Between CEO and HR, I'd expect HR to be even less informed about the technology of the company.

OAuth is an awful solution, but it's the least complex SSO / third party auth system that is actually safe, so if you want to offer SSO it's your best bet. That's why so many auth providers use it even though it doesn't make them interoperable or hot-swappable.

hmm no, the context did not make it any less weird

@Lensflare Thanks!

Also to clarify, the first M is the last digit of the month and the second M is the first digit of the month, because the clusters carry place value information

YM/DY/MD is the best date format

mfw cryptobros use my face for money laundering

@joewilliams007 It's probably not stupid expensive, but at one time they had a lot of traffic and I think they lost interest before the traffic dipped, so I expect there's a lot of expensive scaling infrastructure sitting on the lowest tier, not completely unused but not justifying its upkeep either.

in truth, since DevRant++ subscriptions stopped working, dfox will almost certainly notice the uptick in server costs, but hope dies last.

has the official app been removed? It would be so funny and cool if the platform outlived the app!

no I agree, reflection is scary, even if you wrote it.

and I bet that there's a soft requirement in the C version to manually initialize the buffer before calling another library function because it's full of Infinity and NaN by default.

You may argue that my comparison is unfair because initialize_* is a convention just like new, but

1. it's not a language-wide convention, only one of many conventions, new is a language-wide convention.
2. no tool will tell you that the function exists, you have to read the library source or hope it's written down somewhere in REAL plain English.

You can also argue that the sole pointer argument to initialize_* is obviously supposed to be uninitialized, but the function could equally likely expect the struct to be initialized and instead initialize the floats in the buffer.

@nike plain English would be "a layer stores a sequence of real numbers initialized to zeroes".

The rust version says "a layer contains a vector of float. The main way to prepare a layer is by providing a size, in which case the layer will contain that many zeroes".

The C version says "a layer contains a reference to a float and a size. initialize_layout is a function that accepts a reference to a layer and a size, replaces the reference to float in the layer with a reference to new uninitialized space for the specified number of consecutive floats, and replaces the number in the layer with the size."

There's no structure in C at all. Until you read the function body you can only suspect that the pointer is supposed to be pointing to an uninitialized layer, that the pointer inside a layer is an array, or that the array is owned by the layer. Even after reading it, you still don't know whether there are other ways to initialize a layer and whether the array can be null.

@K-ASS your orange peeler didn't peel my banana right.

Are you using a full python parser, or did you handroll the parts you need?

1L paint of a colour they like neatly packaged in 100 1cl baggies, and a shot glass set

I get inspired to clean / repair stuff when I see other people do the same.

What are you porting from?

Skyrant broke for 20 minutes and I had to experience your retardation again, but I remembered that you're a troll and every word spent on you is wasted.

My ambition is to be secure enough that I can ask back "Why do you want to employ me?".

I don't mind coming up with the most compelling aspect of the specific job and I think that's a useful way to answer the question, but it just feels incredibly entitled.

I mean, when you account for all the redundancy, genomes aren't actually very big for how much they describe. I would expect nature to aggressively compress information whenever possible.

it makes me wonder though; as a cat matures, the -b±sqrt(b²-4ac) ÷ 2a in their head has to be calibrated to the actual specimen's body. Since these numbers are weighted against gravity to produce instinctive notions of things like "too far", wouldn't this also be a convenient coefficient for how scared this particular cat should be of an opponent of unknown size rustling in the undergrowth?

@RageExpress English words are on average 6 letters. Your tags are probably more specific than that, but if you were a brilliant architect, the XML files were important enough to build your entire nomenclature around them, and your domain lent itself to metaphors, your tag names should be 6 characters on average.

I mean, it's not like they're particularly good at solving quadratic equations, so it ought to be a hardwired ability