As one on that side of that argument, I have to say I have yet to see LLMs fundamentally improve, rather than being benchmaxxed on a new set of common "trick questions" and giving off the illusion of reasoning.
Add an extra leg to any animal in a picture. Ask the vision LLM to tell you how many legs it sees. It will answer the same amount as a person would expect from a healthy individual, because it's not actually reasoning, it's not perceiving anything, it's pattern matching. It sees dog, it answers 4 legs.
Maybe sometime in the future it won't do that, because they will add this kind of trick to their benchmaxxing set (training LLMs specifically on pictures that have less or more legs than the animal should), as they do every time there's a new generation of those illusory things. But that won't fix the fundamental that these things DO NOT REASON.
Training LLMs on sets of thousands and thousands and thousands of reasoning trick questions people ask on LM arena is borderline scamming people on the true nature of this technology. If we lived in a sane regulatory environment OAI would have a lot to answer for.
It looks like most of the margin differences you complain about are due to icons not filling out their square canvas/bounding box depending on the depicted object. This is generally how icons work, unless you go Android and slap a circle/squircle background on all of them.
Which isn't to say things shouldn't be better, of course.
There's currently a big multi-year community effort underway to overhaul how KDE's visuals are made, including adopting a modularized design system, Figma/PenPot and a new theming engine (Union) designed explicitly to serve the designer's needs. Along with the KDE Linux distro these are our two main next-gen efforts where we want a capability step change in our practice and product.
The screenshot from the article is for a single application that doesn't really have settings. The system settings panel does have a sidebar with separators, including group labels and contrast/color to make it easier to find things instead of a big panel of gray on gray. So... better than GNOME. Here's a screenshot I found:
You can also see that the KDE settings screen can fit more than 3 options at a time without scrolling, which is appreciated.
That said like I said I do sometimes feel that their more modern themes have random extra spacing that would be nice to be able to remove. Better then all the other modern DEs though (except for power user window managers like i3), so in that way they've done a fantastic job. KDE is vastly more usable than the alternatives, including commercial.
>The content has a bigger right margin than the left margin.
That's the thing that always made me feel KDE is ugly. It was never about the themes, the colors (as in palette choice), or the icons. It's that. Everything feels chaotically misaligned in ways I don't see in other software. The fonts don't look nice. Text never has the "right" alignment, or spacing between lines. Elements don't have enough different classes of color attributes (look at a screenshot of Dolphin and how the toolbar and sidebar have the same exact shade of background color and no transition between the two UI elements, not even a separator bar ala _________. Thunar doesn't do this, Gnome Files doesn't do this, Windows Explorer doesn't do this) You can't "fix it" just by installing an alternative theme.
It may be a "people willing to volunteer for KDE" problem but it's not a general open source volunteer problem. The average GTK program these days is almost perfect in that regard. Gnome has a very polished look and so do most apps written by its users/developers.
>When compared to not existing at all, being slow wasn't all that awful.
slow is relative to the use, anyway
I remember the first time I saw the real time chat function of ICQ, where people could see you typing with not that much delay, I was utterly fascinated that such a thing was even happening
normal web pages not filled with animated gifs were not unbearably slow either
"slow" is what happened if you tried to use Real Player and saw that dreaded "buffering" every 5 seconds of video
Add an extra leg to any animal in a picture. Ask the vision LLM to tell you how many legs it sees. It will answer the same amount as a person would expect from a healthy individual, because it's not actually reasoning, it's not perceiving anything, it's pattern matching. It sees dog, it answers 4 legs. Maybe sometime in the future it won't do that, because they will add this kind of trick to their benchmaxxing set (training LLMs specifically on pictures that have less or more legs than the animal should), as they do every time there's a new generation of those illusory things. But that won't fix the fundamental that these things DO NOT REASON.
Training LLMs on sets of thousands and thousands and thousands of reasoning trick questions people ask on LM arena is borderline scamming people on the true nature of this technology. If we lived in a sane regulatory environment OAI would have a lot to answer for.