brigade

29 ms·

301.
by brigade 4y ago
In addition to LiDAR, Tesla also bucks self-driving wisdom by refusing to use mapping data. Everyone else’s L3/L4 plans include a reference mapping of what they expect to see, with the car signaling a disengagement if it deviates too f
302.
by brigade 4y ago
> Tesla can't even guarantee that their cars won't barrel directly into a fire truck that's stopped on a freeway trying to address an existing accident. Isn’t this a design limitation of radar ACC systems, not FSD? Not tha
303.
by brigade 4y ago
Can't be, in CA you can't return normal plates, they explicitly don't want them back. And you don't really have to return vanity plates either. Rather, there's an online form for notifying them of a sale.
304.
by brigade 4y ago
If by "best" you mean "more job opportunities." That's always been true, and earlier migrants always hate the later migrants. At most, you could say that the preference of living in the city vs. its suburbs flippe
305.
by brigade 4y ago
Relative to VP9, https://arxiv.org/pdf/2008.06091.pdf is a good reference. The technical details on the official AOM site are pretty much just the raw specs and code.
306.
by brigade 4y ago
If you scale frequency down far enough then leakage power can easily dominate, and a decade ago it was bad enough that race-to-idle really was more power efficient than the DVFS most SoCs could do. Vendors were so bad at this that one of AR
307.
by brigade 4y ago
Within the same room at least, the big catch is that that’s with 16 spatial streams. A more likely 2 streams would top out at 5gbps link rate, and somewhere around 3gbps of real throughput. Then the only real caveat is that you’re using 2&#
308.
by brigade 4y ago
Just because a marginal cable works on one machine doesn't mean it'll work on another. I used a USB cable, but I retested with an active TB4 cable with the same results. And again with a passive TB3 cable to the same results.
309.
by brigade 4y ago
Getting USB 2.0 speeds strongly suggests that the cables somehow had horrendous signal integrity... Anyway I grabbed a 3.1 gen2 SSD and plugged it directly into my 16" M1 Pro (Sandisk Extreme), and I got 850 MB/s read/write w
310.
by brigade 5y ago
Well the claim is also that it isn't generally an issue for the M1's architecture until you scale up to 32 cores or more [1], which is beefier than any other current GPU outside of AMD and NVIDIA's offerings. So it's c
311.
by brigade 5y ago
...there are modern integrated GPUs that don't use virtual memory for all global resources? I expected that was a given to even just have an MMU enforce security when accessing RAM.
312.
by brigade 5y ago
Accessing tile memory doesn't need an address translation; it's not part of virtual memory. Accessing global memory does. I think the claim is that the typical working set can exceed what the TLB can map once the GPU scales to 3
313.
by brigade 5y ago
Part of being pulled over is not blocking traffic while stopped, so if there’s literally no shoulder/dirt to pull over onto then obviously you continue until there is. Just like a human driver is legally required to.
314.
by brigade 5y ago
Parking citations attach to the car (as does civil collision liability generally), but moving violations can only be levied against the driver. Which is how some red light camera fines are avoided, if the driver isn’t clearly identifiable b
315.
by brigade 5y ago
Same way humans do? It needs to recognize flashing lights and emergency vehicles to obey the law to get out of their way, and from there add code to fully pull over and remain stopped as long as there’s a cop car with flashing lights stayin
316.
by brigade 5y ago
I do wonder how accurate that is these days... Like it says that my user agent is significantly more unique than my monitor resolution, but Safari froze the user agent years ago. Which means that they calculate there are 21x more people tha
317.
by brigade 5y ago
Unless every single one of those CVEs was actively exploited at the time a fix was committed to the kernel, no. Just no. Even then, no.
318.
by brigade 5y ago
Ehh.... Audi's functions on demand does explicitly say that "the hardware for the functions that can be activated subsequently is installed in all vehicles at the factory" to reduce production costs [1], and the e-tron is o
319.
by brigade 5y ago
It's actually the high-end cars that are on the forefront of moving features to subscriptions, starting with BMW's CarPlay subscription (well they changed their mind), to remote start (Mercedes charges a subscription that Toyota w
320.
by brigade 5y ago
Manufacturing defects are not hardware bugs. Binning is irrelevant to hardware bugs.
321.
by brigade 5y ago
Even without anything plugged in, or any other displays downstream of the host thunderbolt port, XDR will preferentially negotiate one HBR2+DSC stream if the host supports it. (17.28gbps) Which yes, means that a hypothetical 6k120 could be
322.
by brigade 5y ago
Samsung G9 is for sale now, and it pushes 5120x1440 at 240Hz without subsampling through DisplayPort 1.4, which is the exact same pixel rate as 5120x2880 at 120Hz.
323.
by brigade 5y ago
Must have last checked a while ago, because the Vive is #4 these days at 7.3%; 47% of the headsets used with SteamVR are Quest 2, with the original Quest making up another 3.8%.
324.
by brigade 5y ago
The Windows ecosystem mess is why Apple says capabilities may vary. - Yes, plenty of people including USB-C advocates like Benson Leung really wish NVIDIA didn't drop it for the 30-series. Most new laptops have an appropriate port, how
325.
by brigade 5y ago
I think it's negative for every major codec thus far, except maybe VP9 (did any videoconferencing apps use VP9?) H.263, H.264, VP8, HEVC, and AV1 certainly were all used by various videoconferencing apps before hardware blocks reached
326.
by brigade 5y ago
And bitrate goes up for 4k tiers. Even if bitrate only doubles for 4x the amount of pixels, that'll probably look better with modern codecs (depending on where in the quantization curve you are), and most services use at least 3x the b
327.
by brigade 5y ago
The JPEG -> JXL conversion is reversible, so you can encode JPEG -> convert JXL -> convert back to JPEG as needed. You could potentially encode a JPEG-XL directly with the subset that can be converted back to JPEG, but it's no
328.
by brigade 5y ago
A lifting implementation of a wavelet is pretty low complexity, iirc the CDF 9/7 used in JPEG2000 was effectively a 5-tap filter. Implementation-wise, the biggest issue is that approximately no one bothered optimizing their memory acce
329.
by brigade 5y ago
Alternatively, if you're being confused by the quoted intermediate DCT precision, that's not relevant to the final output either. You cannot implement any transform, DCT especially, without having intermediate values with a grea
330.
by brigade 5y ago
I don't believe any platforms or browsers have enabled JPEG-XL support yet, so right now you'd need to ship the decoder yourself anyway. But, it's at least as much better than WebP as WebP was better than JPEG. And unlike HEI

More ›