wtallis

27 ms·

571.
by wtallis 1y ago
These days, if I'm feeling generous I'll spend a minute or two to see if I can get a promising Python tool to install with uv. If it's not going to easily submit to a `uv tool install`, then I move on and forget about it.
572.
by wtallis 1y ago
The source of funding matters. There are parts of the Federal government that are not entirely funded by annual appropriations from Congress. Funding from permanent appropriations or fee collection has not stopped, and agencies are allowe
573.
by wtallis 1y ago
In what way do you think that's relevant? Large portions of that standard had already been abandoned in the Windows 95 era. Nowadays, approximately nobody uses Shift-Insert for Paste, and most laptop users wouldn't even know whe
574.
by wtallis 1y ago
> And if your text exceeds what can fit in multiple columns on a single screenful, scrolling also becomes awkward, because you'd have to scroll exactly to the next screenful to have any consistency. That used to be a solved problem,
575.
by wtallis 1y ago
They've been aggressively removing or reducing hardware that's vestigial from the perspective of AI. NVIDIA's Hopper has no display outputs, no raytracing hardware, no video encoders, and only one of the eight GPCs has raster
576.
by wtallis 1y ago
> The M1 Ultra doesn't have 800GB/s because it's "integrated", it simply has 16 channels of DDR5-6400, which it could have whether it was soldered or not. The M1 Ultra has LPDDR5, not DDR5. And the M1 Ultra was r
577.
by wtallis 1y ago
With regards to latency: reaction time is the wrong thing to compare against, since input latency can be noticeable even when it is much smaller than a human's reaction time. And wanting to have your keyboard latency be no worse than
578.
by wtallis 1y ago
I've learned to be wary, too. Most desktop motherboard these days have several fan headers just above the memory slots. They also frequently use DIMM slots that only have one latching lever, usually at the top of the slot near the fan
579.
by wtallis 1y ago
Did you click on any of those links aside from the final one to the Consumer Reports article? All the other links are merely to listings of blog posts with the given tag. There's nothing specific or even human-curated about where those
580.
by wtallis 1y ago
This wasn't published in an academic journal. It was published on the News page of Harvard's School of Public Health. It's PR/publicity for the school. The purpose of the post wasn't to appease industry donors, but
581.
by wtallis 1y ago
There's real value in having a multi-threaded benchmark that doesn't ignore Amdahl's Law and pretend that everything is embarrassingly parallel.
582.
by wtallis 1y ago
> maybe there is an issue and you return it, so you are still more likely than a random person in purchasing one again soon I find it hard to believe that return rates are high enough for this to be worth the trouble. It's much eas
583.
by wtallis 1y ago
I think the advertising industry has a very long tradition of relying on and trusting bad data if it's the only data they have. As long as everyone plays along and believes that Nielsen ratings or circulation numbers or click count
584.
by wtallis 1y ago
> So far they’ve chosen to go with a simple feature flag but eventually it may require a pullout. Isn't that entirely your speculation, not supported by anything Apple has actually said or done? As far as I can recall, the most no
585.
by wtallis 1y ago
> And according to the European Commission, under the DMA, it’s illegal for us to share these features with Apple users until we bring them to other companies’ products. If we shared them any sooner, we’d be fined and potentially forced
586.
by wtallis 1y ago
The headline currently reads: "Apple threatens to stop selling iPhones in the EU". The blog post cites an article from The Guardian, with the headline "Apple calls for changes to anti-monopoly laws and says it may stop shippi
587.
by wtallis 1y ago
MeshCore seems to enforce a strict separation between client devices and repeater devices, which is a significant downside for many use cases.
588.
by wtallis 1y ago
128GB can't be the current or future limit for a 192-bit bus. There's a missing factor of 3.
589.
by wtallis 1y ago
They got cancelled after they started shipping, and even people who received the hardware got refunded.
590.
by wtallis 1y ago
I wish that was how it worked. Far too often, we have OEMs trying to build unique differentiating features without actually building much in the way of custom hardware, so you end up with some random thing hanging off GPIO pins or somethi
591.
by wtallis 1y ago
Yeah, it's a problem that's not new with systemd. But that never stops systemd from tackling a problem. It's weird to see that they went with a narrow, ugly-hack kind of workaround. And it's weird to see anyone talk abou
592.
by wtallis 1y ago
> This is obviously not about systemd, so what's up? It absolutely is about systemd. It's a project that's famous for having almost no limits to its scope, and is highly willing to break existing conventions in order to
593.
by wtallis 1y ago
> I'm lost. You're complaining that you'd be happy if you could do the thing I said you could do in my opening sentence as well as my closing? Sorry, but I think it is you who have No, please put a little more thought into
594.
by wtallis 1y ago
> There's no need to use systemctl edit to make or edit an override file, its just a convenience shortcut Do you think anyone who's reached this point in the thread has missed out on that? Yes, obviously those steps can be do
595.
by wtallis 1y ago
I think you missed the point entirely. I'm not complaining at all that there's a way to make override files. I'm pointing out that the need for making override files (even with the affordance of a special command to handle
596.
by wtallis 1y ago
> systemctl edit foo.service That seems like an unnecessary foot-gun. Needing a special command to safely edit what appears to be an ordinary config file indicates some other part of the system is trying to be too clever. In this case, I
597.
by wtallis 1y ago
Of course not. But it makes sense that a bug with a narrower scope is more likely to escape testing, and apparently something that changed between 15.6.1 (released a month ago) and 15.7 (released two days ago) affects the process of upgradi
598.
by wtallis 1y ago
From what the article says, this is specifically a problem when updating from an OS version released two days ago, to the other OS version released two days ago. That's not exactly the most logical or likely path for most users to ta
599.
by wtallis 1y ago
Building an array of five 1TB QLC drives seems like a really odd decision, like somebody started with the constraint that they must use exactly five SSDs and then tried to optimize for cost. The 4TB models obviously will hold up better unde
600.
by wtallis 1y ago
You must be using odd definitions for "properly" and "recently". Intel started volume shipments of 10nm-family parts for laptops in 2019, servers in 2021, and desktops in 2022. They've since moved most of their prod

More ›