Thing is in america every collective seems to immediately go to corruption. Unions, HOAs and any kind of "temporary levy for the construction of.." which become horrible, abusive, permanent and increasing every year. In order to get people on board you need to provide them assurances of how this can't and won't happen otherwise they'll just look at the endless examples of their fears coming true and vote based on that perceived reality. It makes it trivial for Amazon et. al to union bust when they barely even need to engage the FUD engine because people have plenty of Fear, Near Certainty and lack of Doubt.

One of the constant irritations I have with windows is that certain keys switch keyboard focus in annoying ways. Ctrl and Shift are (mostly) fine, you hold it and press another key to do a shortcut and if you change your mind or hit it by accident, pressing and releasing the modifier key on its own does nothing. This is good. But Alt and Windows are IMHO *wrong*. Go for Win+r for run, change your mind, too bad now you're in the start menu and everything you type is in the search box. Alt? Menus.

Now, don't get me wrong, I like keyboard navigation for menus, but the shortcut key shouldn't be tapping Alt. On macs it's Ctrl-F2. Can't hit that by accident.

So this copilot key seems like even more annoyance and a return to the glory days of '95 where an accidental hit of the windows key crashes whatever full screen program you're using. Sure, I don't expect it to crash anything, but it'll certainly get in the way.

And just for completeness, even though I said Ctrl+Shift are "fine", I think we've all been stuck trying to get "sticky keys" to disable after we've tapped a modifier a few times...

I don't really consider myself a windows hater, but the more I think about it I think I might hate windows...

From the earliest trailers to the latest build I've wondered what they are doing wrong at Tekken HQ. Unreal in all its incarnations can be bad for noisy, low contrast images, I assume because the defaults are not great and nobody has consistent information on what you "should" set them to. But T8 is especially bad for poor contrast, poor distinction between back and fore-ground and horrendously noisy rendering that fuzzes up an already fuzzy image that is shockingly bad in motion. Unreal games have been relying a bit too much on DLSS to "fix" their noisy upscales for too long. Adding colour blind options that are a series of visual sliders is a good idea in theory, but you're not even putting lipstick on a pig here, you're actively making a bad thing worse and helping nobody.

Two reasons why we wouldn't focus on AlphaFold:

1) It's outside of our wheelhouse so we couldn't do more than read summaries and take them at face value

and most importantly 2) we've heard countless stories of "AI solves problem that humans struggle with" only to find out that the AI managed to find some bullshit way of fudging the problem set that it hits all of the markers needed for "solution" but once taken to the general case for real world use it just fails spectacularly.

And we can't prove 2) because of 1) so we just go with the cynical "been there, done that" approach which is typically the safer option. Cynicism may lead to a miserable existence, but by golly it's right more than it's wrong.

Vulkan isn't a friendly API the way OpenGL was and I think that has really hurt it. Vulkan is what AAA bleeding edge, Unreal Next Version developers need and it doesn't care about your toy renderer. It's a sheer cliff face of difficulty that doesn't get any easier and constantly kicks you in the shins every time you fail to understand the design ethos behind it. I get it, OpenGL was a dead end, it was a bag of spiders by the end and there was nothing that could be done to make it properly threaded and performant without essentially rebuilding it in a non-compatible way, but I feel like Vulkan has caused a divide between the high end and the people who want to learn how to do engine development.

And I feel like Metal *is* that halfway house between "build your skills as you go" opengl and "high performance stateless compute kernel based rendering." Vulkan I think pushes a narrative that either you want to make the next gen rendering engine and are willing to aim for the top 0.01% of programmers, or you should just be playing drag and drop in an off the shelf engine. It means I wouldn't be surprised if it becomes easier for companies to find Metal fluent developers to make mac/ios ports than it is to find competent Vulkan devs.

But that's just my opinion.

Things are not looking good for upgradeability in the future though. Socketed RAM and PCIe may be convenient but they are performance sucking dogs and high end business customers want better which looks like it's going to be soldered HBM and compute accelerators closer to the CPU with better sharing. nVidia doesn't care about gaming beyond marketing now. It was all fun and games when gamers could subsidise their push for compute, but now gamers are a significant cost for a small but irritating competition for their GPU dies that they want to put on $10k datacentre cards and sell by the pallet load. Basically the needs of business customers and the needs of gamers are diverging so much that I would expect gaming hardware to get more niche and more expensive with lesser and lesser gains.

Steam used to have a lot more mac games, but many of them never made the transition to 64bit. Valve aren't interested in updating their games either which tells you how little they care about the mac now. It cools your enthusiasm for purchasing things on a digital storefront when stuff just becomes inaccessible and everyone points the finger of blame at everyone else. The devs are bad for not patching their old stuff, publishers are bad for squatting on the rights and not permitting patches, valve are bad for forcing Steam client upgrades that shut out old the OSs necessary to play these unpatched games and apple is bad for deprecating interfaces that these games need.

Apple making a "push" for mac gaming? Yeah, give me a timeframe for when you'll deprecate Metal 3 and put the onus on a defunct developer to update their old games and I'll decide if the price is worth it. If you can't do that, I'll assume it the next WWDC.

People are struggling to make ends meet and climate change and its proposed solutions look to make life harder, therefore people who last year would have nodded along to everything the AGW reports said are now being pulled into the conversation.

Humans look at existing works and then create something of their own using elements of stuff they like distorted by the lens of their memory and their individual skill. If it's too derivative it's called plagiarism and you'll be made to stop. The deeper people look into LLMs the less we see of creative recombination of public works and more thinly veiled plagiarism where GPT is behaving more like a lazy college student obfuscating what they copied rather than writing their own. What we can say for certain is that the non-deterministic nature of LLMs has very little to do with the model and almost entirely to do with the front end to it. A little bit of overflow here and some jailbreaking there and suddenly the Key/Value weights files start doing what people said they can't do.

Same here, though probably some flavour of Arch. Haven't thought overly much about it yet. I hated Win10 but I've been more or less stuck with it for one reason or another, and Win11 I consider completely distasteful. I have no desire to run that OS at all after trying to use it for a few months. Every time I think "this is not so bad" it does something that just gives me the 'ick. So the excuses have run out, linux and macos it is. It's the office I'm not looking forward to. The staff are change averse and replacing something they are used to with something better is bad enough, I'll get them new hardware, it'll have win11, they'll complain. A lot. And I'll tell them it's this or linux and they'll stick with win11+complaining.

In theory it helps stop money fleeing. If your neighbour can't undercut you, then you both get a share of the money and can focus on attracting investment through something other than being the world's largest shell holding bank. A Walmart/Amazon/Temu race to the bottom but on a nation scale sounds terrifying.

You're not wrong, but documentation by itself is no less error prone than code. Just because someone wrote it down doesn't mean it's correct any more than a function called mysql_escape_string() really escapes a string for mysql... So yeah, I suppose I agree that self documenting code is a fallacy, but at least the code does what it does, where the comments and the documentation can be ambiguous or just outright wrong. I usually take the approach of "comment anything important and if it contradicts the code then at least someone will ask the question. Eventually."

I see in Rust threads a common pattern. Someone expresses difficulty in handling stuff that would normally be okay for a C pointer, there are arguments back and forth before eventually you get to choose between: The Rust Way which is complicated and will likely involve much re-architecting and Just Clone It (tm) where you just make copies of everything and forget about trying to wrestle the borrow checker.

Now, to be clear I'm not denigrating Rust here, that's just a common theme on Rust threads. If anything I'm denigrating the common responses to "I can't do X in rust..."

And when they did get into DVD they did so with discs riddled with anti-piracy features that reduced player compatibility significantly. And it was all for nought as it only takes one pirate to add a workaround to the ripper software and post it, so much so that if you were affected it would be easier to pirate the film than figure out how to watch it "legally."

I got so used to discs that were Region 0 with disc keys of 00 00 00 00 and such because they had no desire to prevent sales that it was always a shock when you got some horrible encumbered disc like the ship hadn't sailed on that forever ago.

The executives see how much money a company is appearing to make with something and they want it. But they never seem to realise these things tend to be like trying to trap lightning in a bottle. It happens, no one really understands why it happens, copycats try to replicate it and fail meanwhile the company that did the original success struggles to keep it going as they don't know wtf happened either.

People who play one game as good as every night as a minority. Mass market appeal needs shorter, more tightly focussed content that can be enjoyed and then put away. But that doesn't make people spend thousands on microtransactions so otherwise okay shooters like Destiny could have been have to stretch that content thin and make it require 80 hours of effort to achieve what a Halo would in 4.

Starfield shocked the industry in that it's "old fashioned" and really not that great overall, but it's a middling offline single player experience that lasts a weekend and it turns out people want that a heck of a lot more than some online service game that they have to set aside grinding hours to keep on top of.