Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
brigade
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
151.
▲
by
brigade
3y ago
I think US cable companies have to offer the option indefinitely because of CableCard, but they’re certainly moving to drop legacy TV delivery as much as possible.
152.
▲
by
brigade
3y ago
For why it's symmetrical now specifically: killing QAM TV delivery frees up a lot of bandwidth for DOCSIS. Until that was a possibility, DOCSIS competed with 200+ channels for cable companies to allocate spectrum to.
153.
▲
by
brigade
3y ago
Custom function ABIs. Though the maintenance overhead really isn't worth the reduced cache footprint, especially since asm writers are allergic to leaving behind any comments. Also instruction scheduling. Low-end Cortex will probably b
154.
▲
by
brigade
3y ago
It's technically correct; FFmpeg has a tiny amount of NEON intrinsics for no particularly good reason. (well, if it had a lot the good reason would have been to avoid writing everything twice between A32 and A64...) Despite all the oth
155.
▲
by
brigade
3y ago
It doesn’t account for humans staying alive through technology, yes. But the earth is still likely to be more habitable than anywhere else in the solar system for the next billion years even as the sun increases its output 10%, the oceans e
156.
▲
by
brigade
3y ago
Both Eero and Apple only support 160 MHz channels in 6 GHz
157.
▲
by
brigade
3y ago
For wall penetration/distance, it's because 5G has better spectrum allocation and allowed transmit power than WiFi. For speed, it ultimately boils down to that massive, expensive antenna array (up to 256 antennas) enabling a highe
158.
▲
by
brigade
3y ago
If there's one thing infosec types universally do not trust to improve security, it's sandboxing. Just look at all the discourse spilled over the recent WebP vulnerability.
159.
▲
by
brigade
3y ago
AV1 video support was already a baseline since Youtube and other websites really want to use it, so AVIF didn't add any attack surface with the decoder. (this is also unlike WebP vs VP8, since WebP added a lossless mode which is quit
160.
▲
by
brigade
3y ago
It is interesting to see how different the reactions are now than last week
161.
▲
by
brigade
3y ago
Are you asking what the full exploit chain is here specifically? Because that obviously hasn't been written up yet. But for the last one, it's the difference between https://googleprojectzero.blogspot.com/2021/
162.
▲
by
brigade
3y ago
Yes parsers are already sandboxed, and violating the sandbox boundary is where the actual valuable exploit is. Parser vulns are near worthless without the rest of the chain building on it, and the millions of man-years it would take to re-w
163.
▲
by
brigade
3y ago
The nature of stability literally is the discussion; I replied to a post that blamed the lack of a Linux equivalent to Win32 on dynamic linking. That the stable ABI you get on Linux requires you to bundle literally every dependency you ha
164.
▲
by
brigade
3y ago
No. The userspace/kernel boundary is explicitly not stable on Windows, and binaries that try to use that interface have broken when it changed.
165.
▲
by
brigade
3y ago
What you mean is: the only stable ABI in Linux is the Linux kernel’s itself. Windows is the opposite: the only stable ABI is the dynamically linked user space ABI. So yes, it’s perfectly possible to have a stable dynamic ABI across a dynami
166.
▲
by
brigade
3y ago
The work needed to maybe get it past outsourced customer support is not at all worth the effort for any site I don’t actually need to use
167.
▲
by
brigade
3y ago
WebP was I believe the first image format that supported lossy compression with transparency. You could argue that you can just use PNG if you want transparency, but allowing lossy compression if you need alpha is more like 10x smaller, not
168.
▲
by
brigade
3y ago
Gas cars have had destination-less range estimates for decades, and as far as I know have never based it off of EPA fuel economy. Instead, they use the car’s own recent fuel economy. Which concept more useful for drivers, the concept of r
169.
▲
by
brigade
3y ago
States are not required to register cars just because the federal government allows them to be imported or sold. They can add requirements consistent with their own laws, as California infamously does causing 49-state cars and parts. Whethe
170.
▲
by
brigade
3y ago
They put the cone on the hood, it's not an obstacle on the road. The car could try to shake it off, maybe, but is shaking it into the middle of the street and creating an obstacle really the right move? Should self-driving cars be requ
171.
▲
by
brigade
3y ago
That’s because CPUs and GPUs are useful for more than just matrix multiplication. TPUs aren’t; they assume highly regular data movement in and out of the ALUs.
172.
▲
by
brigade
3y ago
Probably search results showing items with prime shipping before items without it. Which... as a consumer is exactly what I want.
173.
▲
by
brigade
3y ago
If it's a false dependency, then it doesn't matter what's in the inactive lanes and a simple unpredicated instruction will break the dependency just as well as a zero masking one. Which is exactly what you said the compiler a
174.
▲
by
brigade
3y ago
Zero masking doesn't merge. If you're discarding lanes with a separate merge op, it doesn't matter what the discarded value was.
175.
▲
by
brigade
3y ago
All that solves is changing the merging instruction from a masked merge to a maskless OR.
176.
▲
by
brigade
3y ago
Masked instructions vs subsequent merging are both branchless and have no implicit data-dependent timing relative to each other.
177.
▲
by
brigade
3y ago
> This always benchmarks slower in loops Really? It's a forced read-write dependency on the destination register. Which makes sense for cores with limited superscalar. But for ops with >1 cycle latency or >1/cycle through
178.
▲
by
brigade
3y ago
I don’t think I’ve ever seen a 40gbps capable cable that didn’t support 100W, and Apple isn’t the only one selling such cables with certification. Where are you finding these odd cables? 5/10/20gbps cables, those I’ve seen lim
179.
▲
by
brigade
3y ago
There was some hope that someone would make a USB4 expansion card that didn't support PCIe tunneling, which wouldn't need those extra pins. If this card isn't it, then literally the only interesting thing about it is that i
180.
▲
by
brigade
3y ago
The fractal brokenness of the Thunderbolt ecosystem in Windows is... something. I was hoping that USB4 would at least make it a little nicer by confining failure modes to PCIe not working while the USB tree worked as normal since the USB tr
More ›