"Source code file"
-
This just confirms Twitter is a 1M line BASIC program.
Think of all the GOTOs
-
Is using it without paying him anything fine?
naw. the smog producing generators run exclusively to power grok. the drinking water is being pulled from the ground to cool grok. using grok is the problem, not whether you pay for it.
-
This post did not contain any content.
sounds like a very productive team
-
Notice how it's suggested to cut instead of to copy the file.
Yeah duh, you get the fixed file so you don't need the original anymore obviously
-
Hey Grok, take this one file out of the context of my 250,000 line project and give me that delicious AI slop!
wrote on last edited by [email protected]Perfect, groks context limit is 256,000, and as we all know llm recal only gets better as context fills so you will get perfect slop that works amazingly
/s and more info on quality drop with context increase here https://github.com/NVIDIA/RULER
-
EVERYTHING'S COMPUTAH!
I love teslerr
-
You'd have to say it on the platforms normies use, make it short and attention-grabbing, a drone shot, a couple of tits...
Like this? Unrelated but similar in concept.
-
Like this? Unrelated but similar in concept.
That's amazing. Moving to Canada
-
From the genius and his whipped minions that brought you MechaHitler after some "fixes and improvements"...
...Comes the same AI that offers to "fix your code."
Wow. Simply astounding. A magnificent feat indeed, to waste so many finite precious resources over such garbage.
Kindly Ctrl+A and shift-delete "All your code files", sir.
LOL the macos dictionary entry/quicklook entry for "MechaHitler" gives you Grok's wikipedia page
Which happens because wikipedia redirects "Mechahitler" to Grok
-
please don’t use grok.
elon is poisoning the air over memphis with exhaust from the generators that he’s using to power it while using drinking water from the memphis sands aquifer to cool it.
please don’t use grok.
Why do these places even make sense to build datacenters in?
-
Perfect, groks context limit is 256,000, and as we all know llm recal only gets better as context fills so you will get perfect slop that works amazingly
/s and more info on quality drop with context increase here https://github.com/NVIDIA/RULER
250,000 lines is way more than 250,000 tokens, so even that context is too small.
-
That was clear when they made Bootstrap.
Shots fired!
But honestly, for the time, bootstrap was great.
-
What if this is a plot to literally steal your code? You give it to the AI, page refreshes and clears your clipboard. Your entire "source code file" is now theirs, and probably part of the training set for Grok 5.
The first part is a joke, but the second part...
Ctrl+Z undoes the cut.
-
No, that makes them slower because theyre leaning into the wind. You want to try to mostly use A's, because they're the most aerodynamic, and anything else should be formatted as subscript to keep code size down and reduce drag. C should be avoided at all costs because it's just going to catch the wind.
Can you use T's though, once you have enough As to redirect the wind overtop them? Capital T's of course, lower-case would just get all tangled up in the turbulence.
-
Does he really don't know the difference between cut and copy?
It’s just a mistake. He’s an idiot but he means copy and paste.
-
Think of all the GOTOs
More like all the On Error Resume Next statements. Which wasn't even the worst thing about Visual Basic! VB also had the On Error Resume statement. On Error Resume Next at least moved on to the next line after an error occurred; On Error Resume just re-executed the error-generating line, I guess on the assumption that something would eventually change there.
-
My company actually desires that we use AI. If you fix a problem quickly and it works perfectly but you've done it manually... it's somehow worse than a colleague who took twice the amount of time and had a bad solution, but "generated by AI!"
JFC and I thought having to work in an open-office plan was the worst thing a company could do to me.
-
Presumably anybody that does this gets their code integrated into the training data right?
Is there a downside to training your AI on broken code? Can't be!
-
Have you used GenAI for art? It can do a reasonable job for the entire composition, but the extra fine detail have to be inpainted with specific prompts, and there are special routines that make sure faces and hands look human.
The same with code. It can outline the entire project but details need to be inpainted and specific errors handled.
The context window can't handle an entire codebase, never mind multiple projects.
Actually, I have my entire code base documented in obsidian, and I literally tell cursor to refer to the documentation. It works amazingly well, and then I have it draft documentation for the new features it's creating. I can do in a day what I used to do in a week, and it's not because it's doing anything advanced, it's just takes care of so much of the brain draining tedious tasks.
-
More like all the On Error Resume Next statements. Which wasn't even the worst thing about Visual Basic! VB also had the On Error Resume statement. On Error Resume Next at least moved on to the next line after an error occurred; On Error Resume just re-executed the error-generating line, I guess on the assumption that something would eventually change there.
There would just be one of those, at the top of the 1 file…
