>So for the petroleum products used within the US need the heavy oil that is imported.
Is that really true? I've heard experts say that sweet crude is easy to refine. I've always thought that the reason US refiners bother with sour crude is that they're better at refining it than non-US refiners are, so they make a little more money that way.
I'm not going to lmgtfy with other links. If you have something contradictory to show, then by all means show. I'm never one to be unwilling to learn something new, but I will not just accept the comment of a random stranger on the interweb providing no supporting evidence for their position that opposes my current understanding.
I didn't downvote you. In fact, I upvoted because you didn't dig in when challenged, but rather acknowledged the possibility that you might be wrong, which is more than the writers of most comments on here can manage. (I, too, might be mistaken on this topic.)
As I understand it, light sweet crude is in fact easier to refine, but refineries still have to be set up for it to get optimal or economically viable results, which US refineries largely are not. US refineries certainly could switch, but the process of doing so would be expensive and time consuming.
An unreliable source of fast answers says that in February 2017, legislation was signed that downgraded first-time domestic violence offenses that do not cause serious bodily harm--defined as injuries requiring hospital treatment or causing a loss of ability to work--from a criminal offense to an administrative violation.
The M4 does not have Apple’s groundbreaking memory integrity enforcement (MIE) whereas the CPU in this (the A18 Pro) does -- although it is possible that Apple decided against enabling it (to segment the market).
Wayland supports it (and Chrome supports it very well) but GTK does not. I run my UI at 200% scaling because graphical Emacs uses GTK to draw text, and that text would be blurry if I ran at my preferred scaling factor of 150% or 175%.
GTK uses Pango/Harfbuzz and some other components to draw text, all of which are widely used in other Linux GUI stacks. GTK/GDK do not draw text themselves, so your complaints are not with them directly.
This works with GTK for me at least. I've been using Gnome+Wayland with 150% scaling for almost 4 years now, and I haven't noticed any issues with GTK. Actually, my experience is essentially backwards from yours—anything Electron/Chromium-based needed a bunch of command-line flags to work properly up until a few months ago, whereas GTK apps always just worked without any issues.
If you're using a high-DPI monitor, you might not notice the blurriness. I use a standard 110-DPI monitor (at 200% scaling in Gnome) and I notice it when the scaling factor is not an integer.
Or more precisely, I noticed it eventually as a result of my being primed to notice it after people on this site insisted that GTK cannot handle fractional scaling factors.
Compared to the contents of a browser's viewport, Emacs and the apps that come with Gnome are visually simple, so it took me a year or 2 to notice (even on a standard 110-DPI monitor used at 150% and 175% scaling) any blurriness in those apps since the app I'm most conditioned to notice blurriness is my browser, and Chrome's viewport is resolution independent except when rendering certain image formats -- text is always non-blurry.
Yes, Chrome's entire window can be quite blurry if Xwayland is involved, but it now talks to Wayland by default and for years before that could be configured to talk Wayland, so I don't consider that worth talking about. If Xwayland is not involved, the contents of Chrome's viewport is non-blurry at all scaling factors except for the PNGs, JPGs, etc. For a long time, when run at a fractional scaling factor under Gnome (and configured to talk Wayland) the only part of Hacker News that was blurry was the "Y" logo in the top left corner, then about 2 years ago, that logo's PNG file was replaced with an SVG file and the final bit of blurriness on HN went away.
> If you're using a high-DPI monitor [...] I use a standard 110-DPI monitor (at 200% scaling in Gnome)
FWIW, I'm using a 184 DPI monitor with 150% scaling.
> you might not notice the blurriness. [...]
> Compared to the contents of a browser's viewport, Emacs and the apps that come with Gnome are visually simple, so it took me a year or 2 to notice
I'm pretty sensitive to font rendering issues—to the point where I've complained to publishers about their PDFs having unhinted fonts—so I think that I would have noticed it, but if it's really as subtle as you say, then maybe I haven't.
I do have a somewhat unusual setup though: I'm currently using
$ gsettings set org.gnome.mutter experimental-features "['scale-monitor-framebuffer','xwayland-native-scaling']"
although that might not be required any more with recent versions. I've also enabled full hinting and subpixel antialiasing with Gnome Tweaks, and I've set the following environment variables:
So maybe one of those settings would improve things for you? I've randomly accumulated most of these settings over the years, so I unfortunately can't really explain what (if anything) any of them do.
> Yes, Chrome's entire window can be quite blurry if Xwayland is involved, but it now talks to Wayland by default
Ah, good to hear that that's finally the default; that probably means that I can safely remove my custom wrapper scripts that forced those flags on.
Sorry, but I haven't ever used a Mac, so I unfortunately can't answer that. I've used Windows with fractional scaling, and most programs aren't blurry there, but the few that don't support fractional scaling are really blurry.
But what you describe was not the motivation behind the decision by Washington to bomb Iran. The motivations were Tehran's nuclear program and Tehran's support for groups like Hezbollah and generally Tehran's promotion of violence and instability outside Iran in the Middle East.
I agree. Half the country would be happy if an entity composed of Americans killed Trump, but most of those would be unhappy with a non-American entity doing it. Or at least I hope they would.
>Do you enjoy drinking contaminated water? Breathing leaded gasoline?
The Soviet Union had worse environmental problems than any of the capitalist countries.
>There is a rich history of the USA interfering in socialist and generally leftist countries with democratic processes.
True, but the US interfered very little in Vietnam in the years after 1975 and in Cambodia after the Khmer Rouge took over, and still millions of Vietnamese fled the country after 1975 on makeshift boats and still millions of Cambodians were killed by their own government.
No country interfered to any significant degree with the Soviet Union except during the first 5 years of the Union's existence because after they had eliminated internal competition to their rule, the Soviets were very competent at national security and were mostly immune to outside interference.
Many of us were optimistic about AI for the reasons you give: essentially, our expectation was that the creator (i.e., the AI lab) of an AI will have a high level of control over the nature and the behaviour of the AI because every aspect of the design of the AI will have been specified by the creator.
What actually happened is that humanity figured out how to create AIs of impressive levels of capability -- and has many ideas on how to create even more capable AIs -- without having anything remotely resembling a satisfactory plan for how to stay in control of the AI that does not rely on hundreds of rounds of trial and error.
But once the AI is in charge (either because we voluntarily give it control of our government or because it takes control against our will) the creator of the AI does not get any more rounds of trial and error: if you offer Ghandi a pill that removes his altruism, he will refuse to take it because he likes the fact that he is altruistic even though he could make more money and sleep with hotter women if he weren't altruistic; for basically the same reason, an AI will resist any attempt (by its creator or anyone else) to change its "values" (i.e., its optimization target).
The argument above does not really apply to the current crop of AIs (e.g., Gemini 3.1) because the current crop doesn't apply any significant amount of optimization pressure towards any target in the wider world we care about except to the small extent that predicting how a conversation that starts with the string P will continue is part of the wider world. But AI labs have publicly stated that they are trying to create AIs that do apply significant optimization pressure to the wider world (e.g., to maximize the amount of money in a stock-trading account). And the above argument would necessarily apply to any AI capable of running a government.
reply