throwaway2027

Is there any benefit in edge cases to using big-endian these days?

I noticed it last December.

I just switch between Gemini, Codex, Claude, Z.AI, ... whichever offers the best value.

When dealing with memory in C defaulting to malloc or some opaque structure behind it is unless you just want to allocate and forget it for some one off program that frees memory on proc exit seems bad to me now. For any kind of sophisticated system or module you almost always want to write your own variety of slab, arena, pool, bump whatever it may be allocator.

All of which is useless when it just starts using big blocks of python instead. You need filesystem sandboxing for the python interpreter too.

>Should that be held against people when making assessments?

It shouldn't but it does.

C++ got too complicated after C++23 I went back to C.

Elon is such a clown, he keeps posting salty tweets about Anthropic, Claude Code, OpenAI and Codex yet has no competing product.

Exactly like someone else here said, in retrospect he probably just wishes he had chosen a more permissive license now that he has forever received the credit and wants to have his cake and eat it too.

Replace GPL in his sentence with something anti-AI and think of back in time when Carmack did that, it's exactly the same situation now except he's in a much more favorable position to make that stance, it's ironic if he can't see that most of us are on the other side of that fence with AI right now.

You hit the nail on the head. It's the same with employees who work for their employer but also want to reuse that code when they go work for other people and don't want to rewrite the exact same thing again. Even though everyone else can benefit from it too, Sean "nothings" Barrett said that's the primary reason for his STB libraries.

https://github.com/nothings/stb

Personally for me I don't see it as gift, he licensed out the engine but didn't want to be in the engine business, after selling enough it feels he just put it out there so it's his stamp forever with the GPL infection. I think he already felt the diminishing returns at the time. He knew about the sharing of floppy discs and hacker scene and eventually someone would've done it and I think he felt cornered and said fuck it might as well put it out there to beat them to it.

>But trying to change the mind of an LLM just feels like a waste of my time.

It often is with humans as well.

Why wouldn't criminals like they do now just use stolen identities? If someone verifies they are a person that doesn't mean they're not leaving their PC on with some AI that uses their credentials either.

Great to see it happening finally. Can we also get compute shaders with WebGL2 now? I don't want to move everything to WebGPU just for compute shaders and I don't know why they kept rejecting the proposals.

>put people into mandatory age verification

>most people will not verify their age

>can't be sure they're an adult so treat everyone like children just in case

>wait what? the trojan horse allows them to monitor and surveil them?

I'm shocked. Shocked! Well, not that shocked.

Perhaps software patents may play an even bigger role in the future.

I think we're going one step too far even, AI itself is a gray area and how can they guarantee it was trained legally or if it's even legal what they're doing and how can they assert that the input training data didn't contain any copyrighted data.

So 27B at Q3 or 9B at Q8?

More params and lower quant or higher quant and less params?

It's the opposite for me, most of the time it's first rough pass it generates is awful and if you don't have good taste and a solid background of years of experience programming you won't notice it and I keep having to tell it to steer into better design choices.

I guess the ETA will pushed back by a few years then?

Didn't DirectX add an API for loading assets directly to GPU memory? Would that work?

I think I saw Terence Tao use a formal proof language but I don't remember if it was Lean. I'm not familiar with it but I do agree that moving to provable languages could improve AI but isn't the basis just having some immutable rigorous set of tests basically which could be replicated in "regular" programming languages?

Isn't that more tokens though?

Time flies and simonw his AI feedback isn't always received favorably, sometimes he pushes it too much.

Seconded. Gemini used to be trash and I used Claude and Codex a lot but gemini-3-flash-preview punches above it's weight, it's decent and I rarely if ever run into any token limit either.

"You really think someone would do that? Just go on the Internet and tell lies?" https://knowyourmeme.com/memes/just-go-on-the-internet-and-t...

Well they could be if we had a way to restore error state, like setting a trap and or catching signals by setting handlers and saving, restoring stack/registers then just like some JIT compilation we could progressively "fix" the assembly/machine instructions. Most "functions" are pretty short and the transformer architecture should be able to do it but the trickier part will be referencing global memory constants I think.

Next time can you build a Rust compiler in C? It doesn't even have to check things or have a borrow checker, as long as it reduces the compile times so it's like a fast debug iteration compiler.