kllrnohj

> given that we already had a working GUI. (Maybe that was the intention.)

Neither X11 nor Wayland provide a GUI. Your GUI is provided by GTK or QT or TCL or whatever. X11 had primitive rendering instructions that allowed those GUIs to delegate drawing to a central system service, but very few things do that anymore anyway. Meaning X11 is already just a dumb compositor in practice, except it's badly designed to be a dumb compositor because that wasn't its original purpose. As such, Wayland is really just aligning the protocol to what clients actually want & do.

> Most systems have a way to mostly override the compositor for fullscreen windows and for games

No, they don't. I don't think Wayland ever supported exclusive fullscreen, MacOS doesn't, and Windows killed it a while back as well (in a Windows 10 update like 5-ish years ago?)

Jitter is a non-issue for things you want vsync'd (like every UI), and for games the modern solution is gsync/freesync which is significantly better than tearing.

Rust is better here (by a lot), but you can still ignore the return value. It's just a warning to do so, and warnings are easily ignored / disabled. It also litters your code with branches, so not ideal for either I-cache or performance.

The ultimate ideal for rare errors is almost certainly some form of exception system, but I don't think any language has quite perfected it.

If you forget to handle a C++ exception you get a clean crash. If you forget to handle a C error return you get undefined behavior and probably an exploit.

Exceptions are more robust, not less.

> But how is it different from other tools like doing it manually with photoshop?

Last I checked Photoshop doesn't have a "undress this person" button? "A person could do bad thing at a very low rate, so what's wrong with automating it so that bad things can be done millions of times faster?" Like seriously? Is that a real question?

But also I don't get what your argument is, anyway. A person doing it manually still typically runs into CSAM or revenge porn laws or other similar harassment issues. All of which should be leveraged directly at these AI tools, particularly those that lack even an attempt at safeguards.

Pentile typical omits subpixels to achieve the resolution, so yes if you have an LCD and an AMOLED with the exact same resolution and the AMOLED is pentile, it won't be as sharp because it has literally fewer subpixels. But that's rapidly outpaced by modern pentile AMOLEDs just having a pixel density that vastly exceeds nearly any LCD anymore (at least on mobile).

There's RGB subpixel AMOLEDs as well (such as on the Nintendo Switch OLED) even though they aren't necessarily RGB strip. As in, just because it's not RGB strip doesn't mean it's pentile. There are other arrangements. Those other arrangements being for example the ones complained about on current desktop OLED monitors like the one in the article. It's not pentile causing problems since it's not pentile at all.

iPhones all use PenTile and nobody complains about fuzzy text on them. Early generations of pentile weren't that great, but modern ones look fantastic at basically everything. See also everyone considers the iPad Pro to have probably the best display available at any price point - and it's not an RGB strip, either.

RGB strip isn't really better, it's just what cleartype happens to understand. A lot of these OLED developments came from either TV or mobile, neither of which had legacy subpixel hinting to deal with. So the subpixel layouts were optimized for both manufacturing but also human perception. Humans do not perceive all colors equally, we are much more sensitive to green than blue for example. Since OLED is emissive, it needs to balance how bright the color emitted is with how sensitive human wet wear is to it.

The author is clearly aware of `error.Is` as they use it in the snippet they complain about. The problem is Go's errors are not exhaustive, the equivalent to ENOTDIR does not exist. So you can't `errors.Is` it. And while Stat does tell you what specific error type it'll be in the documentation, that error type also doesn't have the error code. Just more strings!

Is this a problem with Go the language or Go the standard library or Go the community as a whole? Hard to say. But if the standard library uses errors badly, it does provide rather compelling evidence that the language design around it wasn't that great.

Right but Mach 1.0-1.3 is all that "Boom Supersonic" is claiming to hit, though, so the paper is in line with the marketing pitch. The speed advantage of "up to Mach 1.3" might not be worthwhile, no, but that's orthogonal to the claims of "boomless" supersonic.

Now the article randomly pulls Mach 1.7 out of seemingly nowhere, and I have no idea where that came from or how that is justified. But the company isn't making that claim as far as I can tell ( https://boomsupersonic.com/boomless-cruise the "FAQ" section even specifically says: "Boomless Cruise is possible at speeds up to Mach 1.3, with typical speed between Mach 1.1 and 1.2.")

The original comment was "explicitly can't run Linux" which is explicitly not true. Not "it's not fully baked" or "it's not good", but a categorically unambiguously false claim of "explicitly can't run Linux" as if it was somehow firmware banned from doing so.

When things like this (or Vello or piet-gpu or etc...) talk about "vector graphics on GPU" they are near exclusively talking only about essentially a full solve solution. A generic solution that handles fonts and svgs and arbitrarily complex paths with strokes and fills and the whole shebang.

These are great goals, but also largely inconsequential with nearly all UI designs. The majority of systems today (like skia) are hybrids. Things like simple shapes (eg, round rects) have analytical shaders on the GPU and complex paths (like fronts) are just done on the CPU once and cached on the GPU in a texture. It's a very robust, fast approach to the wholistic problem, at the cost of not being as "clean" of a solution like a pure GPU renderer would be.

> Why any technical person would want a PC that explicitly can't run Linux I'll never know.

Huh? https://www.phoronix.com/review/snapdragon-x1e-september

The Snapdragon Dev Kit is canceled. Snapdragon as a whole sure as hell isn't canceled, and Windows on Snapdragon isn't, either. There's loads of Windows laptops using Snapdragon with more continuing to release.

Saying it requires or uses AI seems to be just... modern marketing bullshit. But the technique itself seems sound enough. NASA studied the phenomenon already: https://ntrs.nasa.gov/api/citations/20160007348/downloads/20...

If you live alone by yourself such that the concept of a shared gaming system is entirely unimaginable, maybe you're the one that needs to touch grass

> Sure it costs more, but your 2TB plan isn't a single harddrive, it is likely across several harddrives with RAID ensuring that when (not if!) they fail no data is lost, plus remote backups. When something breaks the subscription fixes that for no extra charge.

Well yes, of course. And for cloud compute you get that same uptime expectation. Which if you need it is wonderful (and for something like data arguably critical for almost everyone). But if we're just talking something like a video game console? Ehhh, not so much. So no, you don't include the backup system cost just because cloud has it. You only include that cost if you want it.

Depends on how much you play. geforcenow is limited to 100 hours a month, with additional hours sold at a 200% premium. This dramatically changes the economics ( https://www.techpowerup.com/344359/nvidia-puts-100-hour-mont... has a handy chart for this )

> TSMC owns 60% of the foundry market. So if China decides to invade Taiwan, that would likely mean ~60% of CPU and GPU manufacturing capacity permanently destroyed at once.

While it would certainly be devastating, do note that TSMC has fabs in places that aren't Taiwan. So their entire production wouldn't immediately go offline, and presumably China would still want to keep selling those products and would have an interest in avoiding destroying those factories.

If China suddenly decides it doesn't want to export electronics, though, then we're all super fucked. After all, what percentage of those TSMC chips flow through China to get mounted onto PCBs or need major supporting components from one of the "Foxconn Cities" in China?

Difference is subscriptions need to support IT staff, data centers, and profit margins. A computer under your desk at home has none of those support costs and it gets price competition from used parts which subscriptions don't have.

Cloud (storage, compute, whatever) has so far consistently been more expensive than local compute over even short timeframes (storage especially, I can buy a portable 2TB drive for the equivalent of one year of the entry level 2TB dropbox plan). These shortage spikes don't seem likely to change that? Especially since the ones feeling the most pressure to pay these inflated prices are the cloud providers that are causing the demand spike in the first place. Just like with previous demand spikes, as a consumer you have alternatives such as used or waiting it out. And in the meantime you can laugh at all your geforce now buddies who just got slapped with usage restrictions and overage fees.

They target different things. kmp/swiftdroid let you share business logic, but not really the UI. Although this is SwiftUI-like, it's not actually swiftui and doesn't behave as such. So you'd be doing platform-specific front ends, which isn't necessarily a bad thing but it's different from the promise of Flutter/React Native which is the same UI everywhere

> The only thing significant about /64 is that’s the smallest unit for SLAAC.

...which means you can't subnet it because you have to assume SLAAC might happen since that's the only thing ipv6 requires. Ergo, an ISP only giving you a /64 means you have to nat if you want subnets, and if you have to nat why wouldn't you use ipv4 instead where it's so much simpler?

I can't use vlans because my isp only gives me a /64.

So I either need to use ipv6 + kludges or ipv4 + kludges. ipv4 is obviously easier and more reliable at that point, it's a no brainer.

Any sort of hot spot / bridge faces the same problem.

Now RFC 9663 is supposed to help here but guess what? It's only like a year old and barely exists. Not 20 years.

It's not that change is painful, it's the ipv6's original design of a shallow depth network was just... bad. Bolting on RFCs to fix it is taking a long time.

It wasn't until the 25th amendment (which, you'll note, came after the 22nd) that the vice president was officially the successor to the presidency. So it would be weird for the 22nd to have a "what if" answer to something that wasn't yet itself law

There's a simpler one: Have Vance run as president with Trump as VP, then Vance immediately steps down on day 1. The Supreme Court will then ignore the intent of the 22nd amendment instead focusing on a narrow interpretation, make up some "this isn't a precedent" one-time ruling that allows it, and ta-da!

How is it an opportunity for Apple? They are a customer of Samsung and Micron RAM modules just like everyone else is. They aren't in any unique position other than their user base is already used to paying extreme markup for RAM. Now whether Apple just eats the cost in their profit margin or charges even more for RAM remains to be seen.

> If we really want to compare, I bet a refurbished Thinkpad would probably be about the same price as the typical Framework mainboard of similar performance.

And a used laptop is cheaper still! Which is all entirely irrelevant to the original point and the goal posts aren't even in view anymore. Finding cheaper products that aren't equivalent isn't exactly hard. If the singular thing you want is CPU performance then a laptop is a bad idea in the first place.

They don't?

Ryzen 340 mainboard: $450 https://frame.work/products/mainboard-amd-ai300?v=FRANTE0005

Ryzen 340 laptop: $1100 https://frame.work/products/laptop13-amd-ai300/configuration...

Yeah it's certainly the single most expensive component, but it's still cheaper than a whole new laptop and, more to the point, less waste than a whole new laptop

Framework as a company is only 5 years old, but they've already done several generations of upgrade parts (including webcam, keyboard, display, mainboards, etc...) and have large availability of repair/replacement parts. So unless their parts are uniquely bad in that they randomly fail after 5 years, it seems pretty clear that they are delivering on the long term ownership goals?

What were you comparing against? When I bought my framework it didn't really have a price premium to it relatively to comparable laptops. The main compromise was the thicker chassis and fewer options in terms of things like premium displays.

> Plus I prefer intel for TB support for an egpu.

lol you and nobody else prefer Intel in a laptop these days. But FYI framework has TB support on their Intel skus and AMD has USB 4 (aka, thunderbolt 3++)