|
Why I am I getting blamed for this third party AI company's gsps not storing in dicom server? 5 minutes later... Oh no these assholes used group 1337 for their dicom private tag group, this is going to go great. 5 minutes later... Oh jesus christ not only are a million type 1 tags missing, they're not validating the lengths for any of their VRs etc etc.
|
# ? Feb 13, 2024 17:46 |
|
|
# ? May 11, 2024 11:40 |
|
no golang code allowed itt
|
# ? Feb 14, 2024 01:41 |
|
Bruegels Fuckbooks posted:Why I am I getting blamed for this third party AI company's gsps not storing in dicom server? Count yourself lucky it was 1337 not 1488
|
# ? Feb 15, 2024 13:00 |
|
Look, if you want to use Forth, just use Forth, instead of doing whatever the hell this is
|
# ? Mar 1, 2024 13:17 |
|
https://x.com/ataiiam/status/1765089261374914957 its so clever a human couldn't write it
|
# ? Mar 6, 2024 20:27 |
|
"VERY few engineers out there could write it from scratch" [citation needed] It's not even that complicated of a type. C++ programmers are writing similar things with template behemoths and SFINAE all the time (god help them and us all). They'd probably kill to have it as easy as just a bunch of nested type-level conditionals with clearly defined and understandable conditions.
|
# ? Mar 6, 2024 21:01 |
|
maybe copilot said "no human should write this from scratch" and they misread it
|
# ? Mar 6, 2024 21:06 |
|
JawnV6 posted:https://x.com/ataiiam/status/1765089261374914957 That's a lot of useless comments. Be nice if it said what the type is for and how it's used.
|
# ? Mar 6, 2024 21:07 |
|
Jen heir rick posted:That's a lot of useless comments. Be nice if it said what the type is for and how it's used. You don't get it, only an AI is capable of writing so many white-noise comments
|
# ? Mar 6, 2024 21:34 |
|
QuarkJets posted:You don't get it, only an AI is capable of writing so many white-noise comments Have you never met freshman CS students?
|
# ? Mar 6, 2024 22:01 |
|
OddObserver posted:Have you never met freshman CS students? In highschool we had to write comments on every line, even it it was stupid poo poo like "i = i + 1; // add 1 to i"
|
# ? Mar 6, 2024 22:09 |
|
I don't know that language but that guy is very impressed with his own amazing cleverness in getting ai to write uh, whatever that pile is. looks like some metaprogramming wank
|
# ? Mar 6, 2024 22:10 |
|
Hammerite posted:I don't know that language but that guy is very impressed with his own amazing cleverness in getting ai to write uh, whatever that pile is. looks like some metaprogramming wank It's code that given a schema object generates the types defined by that schema. I've seen very similar code (quite a bit more complex, actually) in SDKs for services that let you define your own data shapes, like CMSs or e-commerce platform. Like, you setup your ecommerce so your products have a barcode and a dictionary of nutritional values. At build time, you download from the ecommerce a json metadata file that looks something like this (omitting quotes because I'm pooping): { barcode : string | undefined, nutritionalInfo: { type: object, attributes: number }}. You put that into a type, feed that to MappedParameterType<> and you have a reasonably strongly typed representation of a product that will update at compile time when you change the ecommerce configuration. e: it's just a very basic xsd.exe but built into the language's type definition syntax NihilCredo fucked around with this message at 22:31 on Mar 6, 2024 |
# ? Mar 6, 2024 22:29 |
|
NihilCredo posted:It's code that given a schema object generates the types defined by that schema. So the traditional way would be to have a compile step that generates types from a schema right? That's how I've done it, it can be kind of fiddly, but you can manually define a custom schema map/template in the tools I've used. If this thing messes up, I think you're just hosed. Schemas can get pretty complicated.
|
# ? Mar 6, 2024 23:02 |
|
Jen heir rick posted:So the traditional way would be to have a compile step that generates types from a schema right? That's how I've done it, it can be kind of fiddly, but you can manually define a custom schema map/template in the tools I've used. If this thing messes up, I think you're just hosed. Schemas can get pretty complicated. Pretty much. But I think you're selling it a bit short. I've definitely seen very lovely schema compilers that would just crash with a helpful "Invalid schema" error and leave you to figure out where the problem was by elimination. A TS recursive type definition looks scary (it definitely scared me the first time), it may not be much less code than a hastily-written schema compiler, but it's definitely a LOT less code than a good schema compiler with helpful error messages and useful options. If you're working with some niche or closed service that doesn't have a large user community, you'd much rather write this than a code generator. That is an acceptable tradeoff for the cost of not having a plaintext version of each generated type (and I'm fairly sure you can still print them out if you really want, the tooltips still come from tsc.exe at the end of the day). Though most of the time the guys running the original service are not doing anything fancy and there's no real reason they couldn't have just used XSD or JSON-Schema or any other schema definition language with an existing ecosystem.
|
# ? Mar 7, 2024 12:54 |
|
NihilCredo posted:It's code that given a schema object generates the types defined by that schema. thanks for the explanation. My opinion re: that twitter guy's very high estimation of his own/his AI's impressiveness is unchanged.
|
# ? Mar 7, 2024 13:42 |
|
https://twitter.com/dankuntz/status/1765445935407235252
|
# ? Mar 7, 2024 14:26 |
|
Hammerite posted:thanks for the explanation. I imagine there must be a way of writing the mapping in a way that's actually readable and reviewable?
|
# ? Mar 7, 2024 14:30 |
|
TIL about python prepopulating integer objects for every value between -5 and 256 every time you run a program
|
# ? Mar 7, 2024 14:36 |
|
shame on an IGA posted:TIL about python prepopulating integer objects for every value between -5 and 256 every time you run a program there was no other way for them to increase performance
|
# ? Mar 7, 2024 14:38 |
|
lol
|
# ? Mar 7, 2024 15:12 |
|
Htbh, but I would not approve this in a pull request because it clearly doesn't work if you pass a 0(which is even). AI still has a long way to go before it can replace a real developer
|
# ? Mar 7, 2024 15:24 |
|
GABA ghoul posted:Htbh, but I would not approve this in a pull request because it clearly doesn't work if you pass a 0(which is even). AI still has a long way to go before it can replace a real developer or any negative numbers, despite signed input
|
# ? Mar 7, 2024 15:27 |
|
leper khan posted:or any negative numbers, despite signed input Stuff like that is why I always use Peano numerals. Leads to a beautiful recursive definition of IsEven, too!
|
# ? Mar 7, 2024 15:29 |
|
JawnV6 posted:https://x.com/ataiiam/status/1765089261374914957 Hold by beer, I got this, JavaScript code:
|
# ? Mar 7, 2024 16:21 |
|
leper khan posted:or any negative numbers, despite signed input I think there is a very clever solution for this that could save a lot of work code:
|
# ? Mar 7, 2024 17:12 |
|
GABA ghoul posted:I think there is a very clever solution for this that could save a lot of work theres code like this in my current project and i dont have enough pull to get the person who wrote it on PIP
|
# ? Mar 7, 2024 18:00 |
|
GABA ghoul posted:I think there is a very clever solution for this that could save a lot of work Doesn't look like anything to me!
|
# ? Mar 7, 2024 18:50 |
|
I have literally seencode:
|
# ? Mar 7, 2024 19:10 |
|
code:
|
# ? Mar 7, 2024 19:25 |
|
lmao this one is great
|
# ? Mar 7, 2024 20:07 |
|
someone in the replies offered an improved version
|
# ? Mar 7, 2024 20:22 |
|
Not shown: this part. code:
|
# ? Mar 7, 2024 20:31 |
|
Long before I was aware of modulo (one of the all-purpose tools in a programmer's toolbox), I came up with what seemed at the time like a foolproof method for determining divisibility: 1. Divide a number by its possible divisor (e.g: 172 / 5). 2. Convert the result into a string. 3. Iterate through the string to look for a decimal. 4. If no decimal is found, the number is divisible. e: I didn't read Qwertycoatl's post closely enough. Guess I was on the right track if variations of this trick have been seen in "real" production code. F_Shit_Fitzgerald fucked around with this message at 22:22 on Mar 7, 2024 |
# ? Mar 7, 2024 21:35 |
|
code:
|
# ? Mar 7, 2024 21:57 |
|
code:
|
# ? Mar 7, 2024 22:03 |
|
Qwertycoatl posted:I have literally seen I love this so much, especially because it doesn't even work if you execute it on a German Windows So here is the correct version that is actually safe to deploy internationally code:
|
# ? Mar 7, 2024 22:04 |
|
GABA ghoul posted:
fail PR. use "." and the invariant cultureinfo
|
# ? Mar 7, 2024 22:21 |
|
leper khan posted:fail PR. use "." and the invariant cultureinfo whoa, buddy, what kind of thread do you think this is? We don't look kindly on that kind of code in here
|
# ? Mar 7, 2024 22:29 |
|
|
# ? May 11, 2024 11:40 |
|
GABA ghoul posted:
lol our company several years ago bought another tiny company that had a valuable client list but was run by three old guys who all wanted to retire. two of them were self taught programmers. one of the in-house pieces of software they used operated in two different modes, because the two of them had different preferences. it decided which mode to operate in based on whether there was a folder on the C drive named "hp", because one of them had a hp laptop. (i might be misremembering, I think it was hp, it might have been some other company.)
|
# ? Mar 7, 2024 22:49 |