comex

18 ms·

241.
by comex 1y ago
i686-pc-windows-msvc and i686-unknown-linux-gnu are still tier 1. Any bugs that show up in i686-pc-windows-gnu but neither of the other two are almost certainly MinGW-specific jank, rather than anything that could have security implication
242.
by comex 1y ago
The question is how best to send the modulus, which is a much larger integer. For the reasons below, I'd argue that base64 is better. And if you're sending the modulus in base64, you may as well use the same approach for the exp
243.
by comex 1y ago
We can only speculate why the employee decided to merge the PR, but my guess is that they did it as a joke — which would mean that at least one person with merge permissions does view the whole affair flippantly, and was comfortable broadca
244.
by comex 1y ago
> There was no standardization of parts in the probe. Two widgets intended to do almost the same job could be subtly different or wildly different. Braces and mountings seemed hand carved. The probe was as much a sculpture as a machine.
245.
by comex 1y ago
> This is exactly what Fil-C does. Okay, I just checked. It does not. I wrote: "There must be something ensuring that what the kernel interprets as a pointer is actually a valid pointer." And sure enough, your runtime manual
246.
by comex 1y ago
> C and C++ are often chosen because they make systems programming possible at all due to their direct access to syscall ABI. Surely Fil-C cannot provide direct access to syscalls without violating the safety guarantee. There must be so
247.
by comex 1y ago
But does that mean much when the training that produced the original model was not distributed?
248.
by comex 1y ago
FYI, the Copyright Office doesn’t enforce copyright law or determine its correct interpretation. Courts do. The legal analysis in this report is really just a suggestion, and judges probably won’t give it too much weight. As for illegal a
249.
by comex 1y ago
The executable format defines a number of segments, which must be on separate pages because they have different permissions, and can’t be slid relative to each other. So the page size can’t be larger than the segment alignment chosen at li
250.
by comex 1y ago
Isn’t the biggest issue that compiled binaries have segments aligned only to 4kb, and those segments must be on different pages (because they have different memory protections) and can’t be slid relative to each other? I’m not sure how iOS
251.
by comex 1y ago
According to Linus, “the ‘various compliance requirements’ are not just a US thing”: https://lore.kernel.org/all/CAHk-=whNGNVnYHHSXUAsWds_MoZ-iEg... I do not know the details myself.
252.
by comex 1y ago
From what I’ve seen, Stack Overflow answers are much more reliable than LLMs. Also, using Stack Overflow correctly requires more critical thinking. You have to determine whether any given question-and-answer is actually relevant to your pr
253.
by comex 1y ago
That's true, but it illustrates a point about 'jagged intelligence'. Just like there's a tendency to cherry-pick the tasks AI is best at and equate it with general intelligence, there's a counter-tendency to cherry
254.
by comex 1y ago
> and I struggle to think of what would lead one to the urge to implement a key=value config file parser in 2025 C/C++ culture never changes. As many new build tools and package managers as people come up with, the ‘default’ environ
255.
by comex 1y ago
1000+, though the last two generations have included only a subset.
256.
by comex 1y ago
Hmm. Still has the biggest issue I’ve observed with AI songs in the past: the lyrics don’t have a consistent or pleasing rhythm. It seems like the lyrics are generated with little regard to rhythm, then the melody is generated and tries t
257.
by comex 1y ago
At least for federal courts, the records are available on PACER but they cost money. Who is paying?
258.
by comex 1y ago
It’s just a hallucination, same idea as o3 claiming that it uses its laptop to mine Bitcoin: https://transluce.org/investigating-o3-truthfulness I doubt the model was trained on Street View, but even if it was, LLMs don’t r
259.
by comex 1y ago
Sam Altman wrote in February that GPT-4.5 would be "our last non-chain-of-thought model" [1], but GPT-4.1 also does not have internal chain-of-thought [2]. It seems like OpenAI keeps changing its plans. Deprecating GPT-4.5 less t
260.
by comex 2y ago
Some feedback: On mobile, why is there a custom keyboard? I found it pretty annoying. I kept getting missed taps, and at least one wrong guess due to a typo I didn’t catch before. It doesn’t seem like the system keyboard’s spellchecking
261.
by comex 2y ago
IOKit C++ is running in the kernel, so it's not really related to any of the technologies you mentioned which are all userland-only.
262.
by comex 2y ago
Today the kernel does create the swap file itself. I don't know why it behaved differently in the past, given that the version of dynamic_pager I linked is only 355 lines of code, not obviously complex enough to be worth offloading to
263.
by comex 2y ago
At the risk of nitpicking, there are a bunch of things that are not quite right. Nonexhaustive list: - Discussion of paging mixes together some concepts as I described in [1]. - Mach port "rights" are not directly related to en
264.
by comex 2y ago
The claim is inaccurate and mixes together multiple different things: - The Mach microkernel originally supported true userland paging, like mmap but with an arbitrary daemon in place of the filesystem. You can see the interface here: htt
265.
by comex 2y ago
So how does the 10M token context size actually work? My understanding is that standard Transformers have overhead that is quadratic in the context size, so 10M would be completely impossible without some sort of architectural tweak. This
266.
by comex 2y ago
> I think they have said somewhere that Headscale is actually revenue positive for them. I have the same memory. But they may not feel that way forever. Many a company started by attracting developers with a generous free tier or open-
267.
by comex 2y ago
> Of course we don't know what kind of information the model encodes in the specific token choices - I.e. the tokens might not mean to the model what we think they mean. But it's probably not that mysterious either. Or at leas
268.
by comex 2y ago
The paper significantly overstates the scope of PAC (pointer authentication codes) on Apple platforms. To quote the paper: > This is effectively what is done in Apple software, which uses special ARM hardware support to also check point
269.
by comex 2y ago
The post states clearly that the foreign server only gets private posts if one of that server’s users is authorized to read the posts. How else do you expect it to work?
270.
by comex 2y ago
Update: I tried a different version of the prompt and it's doing really well! Well, so far it's gotten out of its house and into Professor Oak's lab, which is not so impressive compared to ClaudePlaysPokemon, but it's a

More ›