Anyone that just plugs a random usb in their pc should be fired anyway

One rule: never drink alone.

That single rule can protect you and make you realize when you start having a real problem.

@whimsical I just wanted to start a new Svelte+Vite project with a basic dashboard. I provided a detailed description and said to only focus on the frontend part and not even worry about any logic yet. The files looked fine, but the packages it used were busted, one outright didn't exist, another broke the entire deployment and not even fresh reinstall helped. Giving the errors to gpt5 just resulted in it recommending the same solutions over an over, none of them worked.

I gave up on even trying any of the assembly stuff I wanted to test with it.

For what It's worth, the code itself looked fine, but I never got to run it because of It's dependencies.

Claude does this better, though even cloude starts screwing up projects once they grow to a few thousand lines l and couple of files.

Not impressed overall

For me it still doesn't do the things it didn't know how to do previously. So no big update

Oh lol, yeah, I hate that too xD

Don't worry, Chat GPT5 can't replace you in treating women like shit, your job is safe

You also need to format as FAT32 and mark the partition as bootable. And if It's EFI I think there's another special thing to do according to the spec, not all mobos follow the spec but some do. Can't remember what it was

It's not always about selling data. It's also a common method to reduce spam and bots.

Configure your torrent to only allow encrypted connections

Junior/Medior/Senior is a function or experience, attitude, apparence and performance.

Im not sure what the function looks like exactly, but you can identify anyone as any of those based on those 4 attributes.

Someone can actually have the years to X someone can actually have the mindset of X
Someone can act as if they know X
Someone can show the actual results for X

Companies boil it down as years of experience because It's the simplest to quantify and there is some* correlation. But as a dev looking for work you also apply to the levels where you feel comfortable in (usually the level is mentioned in the posting) so you also grade yourself to an extent... Lastly the interview processe will do some basic screening, but not everyone will tell you if you're over-qualified so It's a bit unfair.

That's where the system mostly reached equilibrium because there's no much more you can do without probing human brains.

The people that though emoji-yellow is chinese are the same people that though git-master is a racial slur. But oh well

I'm not a fan of windows and microsoft, but you really can't blame the build quality on them can you? Unless it was a surface laptop

Yeah, hash is about as reversible as recovering X from

X % 3 = 2
recover X....

You can use attacks on some weaknesses and dictionary comparisons if you know the data is at least actual language (or some other known final form) but that's about it

Remember, we want AI to automate the creative tasks, so we can finally do laundry! Or something like that :)

@kiki the entire toolset and standard libs to compile and run C# stuff. Microsoft created a version for linux (I suppose to compete with java) and it works, but whenever you update it using a package manager it pulls like 1GB+ of files... Not to mention mono itself is huge.

@jestdotty pretty much. Have you ever tried to use Mono on linux?! Worst tooling and package repo I ever seen. It has microsoft written all over It's slow downloads and abusive storage usage. :D

And this is obviously not limited to MS, but It's funny that they happen to be part of the bloat problem on linux eco too somehow

@SidTheITGuy yeah, the money was grabbed and the people were scammed, so the "experts" left. The greatest of these grifters being fucking Sam Altman, who will, no doubt, mention how they already achieved AGI for the nth time and release yet another identical model.

Truly, anyone can act like an "expert" these days -_-

The fact that alpine doesn't inherit the issue you mention proves that the problem isn't core linux, but the individual applications that distros package. The core is solid as fuck

What about it you can't believe? This is literally the adoption curve for every hype and buzzword. Absolutely predictable. Many of us were expecting exactly this since day 1 of major adoption of generative AI. People always get overly excited about things they don't understand, then they learn about it and gain experience using it, which removes much of the smoke and mirrors, then they grow desillusioned and finally everyone slowly accepts the thing for what it really is. Some take longer than others, but eventually the trend always follows the standard adoption curve -_-

Sometimes, always sometimes

Me

You're right, Im not cool because I use linux

I use linux because Im cool

As with all things, the truth is somewhere in the middle

@retoor See, the thing is, I have very specific issues, usually with very low level code and usually trying to test novel algorithms or implement interesting papers... So LLMs usually have very little useful insight for me.

It's great if you tell it "generate the same website template that you've already seen a billion times when scraping the web"

but for anything actually interesting, it's a rubber ducky at best. Sometimes it generates some interesting insights or approaches I didn't know about previously and can catch obvious coding mistakes and edge cases... but it's not really capable of generating novel solutions unless you hold it's hand... And at that point I'm the one providing the detailed spec and design already, it just converts it to subpar, but usually functional code...

which is I'm pretty sure it can't actually replace an engineer at any step of the solution creating process. It can replace code monkeys, but not engineers

@retoor I mean, I'm not going to pay for a month of something I'm going to test once.

sticking with the free option, which btw is not much better than ChatGPT tbh, didn't help. I solved the issue myself in the end again. Turns out it was an issue in my debugging assumptions and previous computations, but it didn't catch it, even though it calculated tons of numbers for me, it never even considered that some of the values I provided were suspicious and mostly agreed with my obeservations (sycophancy is a massive issue for all LLMs, probably unsolvable since it comes from the alignment training and shitty user-centric data)

@retoor, @BordedDev

Im gonna try Claude, but god protect you if it also sucks balls. :)

Try to use it for anything semi-difficult and you'll know what we all think.

Sidenote:

I've been trying to use chatgpt to help me optimize a cuda kernel for the past two days, and the only think it achieved is make me explain to IT why all of it's improvements are moot or outright wrong...

Essentially it worked as a rubby ducky at best, which is valuable, but not a replacement

Instrumental, LoFi, Coffee shop ambiance, retro chiptune.

I can handle lyrics if I don't need to be too focused and Im just in the flow, but if I need deep focus then lyrics are very distracting to me

All signs point to 'yes'

Your first post and you waste it on the weakest most basic rage bait. Im not angry, Im just disappointed....