Roblox is a game engine and social platform. The basic idea is, you create a game in their engine, using their development tools (such as Roblox Studio), and then you market/promote the game on their website for others to play together with their friends. The game runs on their servers, and you don't have to worry about your own infra if you don't want to, in exchange for the money players spend on Robux (their virtual currency, and now the only one), which they can in turn spend on developer products and other paid items that you set up in your games. Then once you make enough Robux from player purchases, you can cash it out through a process called Developer Exchange which essentially makes you a W-2 employee of Roblox (or whatever financial partner they use, Tipalti?)
The idea is that people like playing with their friends and when they can take their friends and make new friends across hundreds of thousands of games they stay for a long time and (they or their parents) make lots of purchases. The social features of Roblox are a huge part of the appeal, even though I was mostly interested in the engine.
Related, I've seen a lot of misidentification of Aspie writing as being LLM-generated lately. You seem Aspie to me (and parent does as well) so it makes sense that you'd also see the similarity.
I wouldn't agree that a lack of autism diagnosis is definitive because that diagnosis is primarily based on need for treatment. (Notice how the diagnostic criteria are exclusively variations of "finds it difficult to be normal".) I only see the tokenized / scratch-blocks writing style with Aspies. (This is likely what comes off as LLM-ish to others.) Non-autistics tend to be very sloppy/imprecise in comparison. ADHD primarily has to do with behavior (misbehavior in fact) and so wouldn't solely be responsible.
I have a few friends who are definitely at other places on the spectrum and yet not diagnosed. The funny thing is Aspies are the most likely to be diagnosed because their early development can be most obvious, but it's still not always caught. (This is supported by recent research identifying the presence of a finite number of distinct genetic phenotypes in the autistic spectrum.)
I really hope Apple does not start anthropomorphizing even more of their interfaces. I don't want to talk to a simile of a human. I want to talk to the machine.
I once ran out of disk space while Chrome was running and, despite me clearing the space again shortly after, the damage was already done and Chrome had already decided to wipe all my local storage and cookies. It didn't keep it in memory to save again once there was space, it just deleted it all permanently.
I really want something between Sequoia and Tahoe. (Probably mostly Sequoia, but with targeted applications of Liquid Glass.) I don't like how Tahoe treats everything as floating on top, as if properly dividing windows into sidebars and panels is wrong... There's so much extra padding and rounding now, I hate it. Everything's lost the depth, detail and cleanliness it used to have, replaced by this bubbly mess. Like, sheets don't even slide out anymore, they overlay like on iOS. The charm, expressivity, and, well, Mac-ness is gone.
I love Liquid Glass - the blur and refractive effects are so pretty and technically impressive - but it should be used tastefully instead of this nonsense. I feel like Tahoe in general is straying way, way too far from the battle-tested Cocoa foundation and into this total top-down crap. Liquid Glass feels like some sort of shareholder-enforced enshittification.
macOS is supposed to be defined from the bottom up; it always has been. There has always been importance in having a solid base; a robust foundation for developers to build on. HIG, Cocoa, CoreGraphics, all of that is in service of this. The user experience and vertical integration is a result of this and couldn't exist without it.
There's so much wrong with Tahoe that goes against everything Mac has ever been. We don't want to dumb down the interface; that has never been the goal. The goal has always been to make the interface intuitive enough that anyone can learn it. macOS and iOS are fundamentally different platforms with fundamentally different design constraints and considerations.
Icons being able to escape the squircle was supposed to be a reflection of the fact that apps on Mac are less contained than apps on iOS. They have more expressive power and more advanced capabilities. You're working closer to the metal and in a less controlled environment. Because of that, you can do more and you're not constrained to the flows of the system.
iOS always hasn't been this. The constraints of touch are different than the constraints of the desktop. Steve Jobs spoke about this a lot back in his day, about why iOS is so much more locked-down than Mac.
But Mac has always been a platform for freedom and control. And Tahoe strips the soul of that.
> But Mac has always been a platform for freedom and control.
My impress has always been the opposite: MacOS is "opinionated", and the user can either accept the Apple way of doing UI or can take a hike.
MacOS has offered token customization, such as allowing the user to change the color of menu bar highlights, but any substantive change required 3rd party intervention, which would inevitably cease to function at the next upgrade.
These days the OS is even more locked down, making it all but impossible to modify OS files.
Catalyst was already sort of a death knell, since it's an admission that it's ok to port over iPhone/iPad HIG to mac. Maybe swiftUI too, since it's replacing appkit and all its various affordances.
Can't speak for GP but I got the feeling that after Apple embarrassed itself shipping almost none of the Apple Intelligence features announced at WWDC 2024, they scrambled to get something drastic out the door to show they're still "innovating" and "doing big things"
I assume the subtext is something like: "Customers are being abused to create the short-term illusion of improvement, to satisfy myopic investors in the financial markets and the personal compensation incentives of executives."
Shareholders want to maximize stock price, therefore they choose psychopathic CEOs willing to do literally anything to achieve that. People who view reputation and goodwill as just capital to be spent. Giving out free service to get people hooked then turning the screws on them is a proven strategy.
None of the siblings got it right. By 'shareholder-enforced enshittification' I meant when shareholders (or, generally, anyone from the top) enforce a direction that doesn't align with what's natural of the foundation. So the system ends up being stretched to afford it, corners get cut / shortcuts get taken, and then that becomes the final shipping version.
I think what they're saying is the methods used today are faster but have a lower ceiling, and that that's why they quickly took over but can only go so far.
I'm not closed to it. You can check my comment history for frequent references to next-generation AIs that aren't architected like LLMs. But they're going to have to produce an AI of some sort that is better than the current ones, not hypothesize that it may be possible. We've got about 50 years of hypothesis about how wonderful such techniques may be and, by the new standards of 2026, precious few demonstrations of it.
Quoting from the article:
"Within five years, deep learning had consumed machine learning almost entirely. Not because the methods it displaced had stopped working, but because the money, the talent, and the prestige had moved elsewhere."
That one jumped right out at me because there's a slight-of-hand there. A more correct quote would be "Not because the methods it displaced had stopped working as well as they ever have, ..." Without that phrase, the implication that other techniques were doing just as well as our transformer-based LLMs is slipped in there, but it's manifestly false when brought up to conscious examination. Of course they haven't, unless they're in the form of some probably-beyond-top-secret AI in some government lab somewhere. Decades have been poured into them and they have not produced high-quality AIs.
Anyone who wants to produce that next-gen leap had probably better have some clear eyes about what the competition is.
I really hope Ternus gets the next CEO spot. I haven't seen much of him, but I watched part of an interview with him and he seems very reminiscent of Jobs, in the good way. Reportedly Tim Cook is preparing runner-ups from all the top-level divisions, not just engineering:
> But Mr. Cook is also preparing several other internal candidates to be his potential successor, two of the people said. They could include Craig Federighi, Apple’s head of software; Eddy Cue, its head of services; Greg Joswiak, its head of worldwide marketing; and Deirdre O’Brien, its head of retail and human resources.
Personally I wouldn't like to see any of these. Especially services needs to calm the fuck down and stop giving into the temptation to milk their captive audience; that's absolutely not what Apple is supposed to stand for.
The idea is that people like playing with their friends and when they can take their friends and make new friends across hundreds of thousands of games they stay for a long time and (they or their parents) make lots of purchases. The social features of Roblox are a huge part of the appeal, even though I was mostly interested in the engine.
reply