LLMs and compilers are a crappy, tired, and spectacularly wrong and misleading comparison, and I'm also not surprised Torvalds used it.

Jank, not yank, of course. Geez.

I think that this is conflating yank with slop, and that there's major differences between the two.

Piranha Bytes games and such clearly have passionate, artistic humans behind them that tried to create something great and different. What makes them janky is that they were overambitious and couldn't quite get to where they wanted with the resources they had. But you can feel the heart and sweat that went into them. That's why "eurojank" isn't even a terribly negative term and those kinds of games have many genuine fans (me included).

The slop variety didn't have anyone ambitious, creative, or dedicated behind it. Just someone who wanted to clone something. It's not janky because of overambition, but lack of ambition, and never having cared enough to put even a minimal amount of effort in it. And you can see and feel that at every corner.

It seems to me someone hasn't read up on Hamilton's career and accomplishments.

Being sour and envious on an obituary isn't a good look.

Apart from that being some of the worst LLM writing I've read so far ("it redefined what winning looked like" -- seriously? And human players didn't disappear -- is that so?) I've heard that a lot of pro chess players are very unhappy about what chess algorithms did to the playing styles used in high-level games.

I've been a professional software engineer for 20 years. I don't see it having changed much about the profession at all, except maybe for giving a part of my peers a weird form of losing touch with reality for some reason.

How much of your job was actually writing code? Timewise, 10%? Maybe 20%? In terms of mental effort, probably far less than even that.

Even if LLMs were good enough to automate all of that perfectly -- which they are so very, very, ridiculously very far from being able to -- that would still be a pretty meagre improvement to our efficiency, considering the investment.

The cost of writing code is moving towards zero. The cost of accumulating technical debt is moving towards zero. The cost of eliminating all knowledge from the company is moving towards zero.

Producing as much code as possible was never the issue. Maximizing technical debt was never difficult.

LLMs are a piss-poor "solution" to something that was never the problem of the bottleneck. So what exactly changed, apart from the fact that management types who haven't written a single line of code in their lives now think "coding is solved", and half the "professionals" in this field suddenly no longer have any qualms about collecting a paycheck for an absolutely abysmal imitation of work that they put zero thought, care, or understanding in?

I'm not saying those have no real-world impact. But what's changing isn't the nature of the job, or what we'd be capable of offering. It's about the threat of enough people being content with stolen unworkable unmaintainable slop garbage trash code and no longer caring about standing for and improving the work we do. I'm anxious to see where that will lead us, and the reputation of a profession that, by that time, I hope to have left far behind me.

It depends on where you live and what your job is. Non-DST in summer where I live would be total bollocks, because you're getting most of the extra hours of daylight the seasons give you at a time where, if you're roughly working standard business hours, you have no opportunity to do anything to enjoy them.

Fair enough, but desktop Teams is also just a website.

I'd agree in principle and can't stand people who still use "Micro$oft" post-1996.

On the other hand, Microsoft's public annoyal about the term "Microslop" has made it so, so much more fun to use anywhere their management might become aware of it.

It does, but it's also US-based so doesn't really solve the problem.

The tech people target audience might be more or less maxed out. Things like the mascot, or making it look pretty and "modern" can make a difference in the wider population. Trying to increase market share is important, because if it dips much further and Google stops being interested in paying, Mozilla might have to pack up.

It makes sense even on a regularly sized 16:9 monitor. Vertical screen real estate has always been at a premium since the move away from 4:3. Heck, most websites are styled to leave left ~25% and the right ~25% of the viewport empty, anyway.

They were already rounded before, but now they're even rounder.

I also don't get it. The rest of the new design looks nice, visually, but this hurts productivity. I had to restore compact, rectangular tabs with userchrome. They're called tabs for a reason -- they should look like them.

Maybe you have some wack extensions installed?

My freshly started Firefox takes less than 200 MB of RAM (Windows 10, almost 20 extensions). Less than a third of what a freshly started Chrome uses.

I suspect most of that is because people simply can't be bothered to even try another browser. Firefox is so blazingly fast and performant compared to Chrome, the comparison isn't even funny anymore. It's been like that for years. Chrome is a snaaaaill.

I know, "benchmarks". But most people's daily browser usage is not running isolated JS calculations. It's working with a few tabs on real-world websites, and that absolutely kills Chrome.

I also barely ever have any stability issues with Firefox, while Chrome crashes a few times a week of normal use. (I have to use it at work). And the compatibility issues are almost never fair to blame on Firefox. It's websites being developed for non-standards-compliant parts of Chrome.