So...I still stand by the fact that the article is problematically-scarce on details regarding exactly how the ex-husband was using the Family Group functions in a harmful way...I get that Kate didn't like it, but there was no indication that he caused direct harm to her or the children.

That being said, the article DID make clear that there WAS a court order for him to disband the account, and even if he was using in all the right ways for all the right reasons, not-complying with a court order is extremely problematic.

I'm not all the way on Kate's side here, but the ex is clearly in the wrong for leveraging the Apple account after the court issued its verdict.

So, I went back to the article and re-read a quote from it...

Looks like I missed the very-important fact that Kate *did* have a court order to disband the group.

While I still stand by the fact that the article is extremely sparse in details, the fact that a judge issued a ruling on the matter does alter things significantly. I still think Kate isn't telling the whole story here, but I *do* think that if the ex-husband was unwilling to comply with a court order, that he needs to be held in contempt, and I also think that Apple SHOULD comply with such a court order, and that Apple DOES need a system in place to comply with court orders of this nature.

My apologies for missing this critical detail.

You seem pretty adamant about this stance, so if you're cool with having a discussion about it, I'd appreciate some perspective - I'd like to engage in a discussion on the topic...

So, first off, I'm not completely convinced this is abuse. The father tracked the kids' location, sure, but every indication is that this is a joint-custody scenario, not a sole-custody scenario. The article gives no indication that the father has done anything harmful to the children as a result of the data he has access to. It gives no indication of physical harm of *any* kind, or any actions he has taken for non-physical harm. so if it's simply the location tracking, is there any statute that indicates that a secondary guardian is expected to forego location tracking?

I'll grant that he should have respected whatever screen time limits Kate wanted to set for the kids, during the time she had custody. Fine. Granted. No contest. If Kate wanted them to have three hours a night on her days, it should have been three hours a night on those days, full stop. However, for this to work, it hinges on the stance that denying screen time to children constitutes abuse. Please keep in mind that the article has no information regarding the father's side of the story; for all we know the father was this way *because* one of the kids got in too deep with someone they met on Roblox and the father ended up having to dispute credit card fraud, or worse.

Maybe - MAYBE - I could find some common ground on this quote from the article: "Kate recalls her own children faced constant aggressive questioning about their movements, social interactions, and activities based on data served up by Apple Family Sharing.". There certainly may have been some overstep here; I make room for the possibility that the ex-husband was helicopter parenting the kids at an unreasonable level. He may have been asking about things in an improper manner or time, and this amount of access certainly makes abuse easy to enact.

In this *particular* case, based on the information expressly stipulated in the article, there is a grey area between "parenting" and "abuse". The article mentions nothing about a restraining order or sole custody, so why does Kate have a monopoly on making sure her children are where they're supposed to be, or who they're talking to? It sounds like there's no winning here - Kate obviously wanted the kids off the father's account, granted, but it speaks to a much deeper issue - what would it take for Kate to consider the ex-husband a good father, to the point that she would respect his wishes in the way she wanted hers respected?

I'm not convinced that this is the case. The 'organized system' would likely need a court order, because if Apple allowed children to be moved from one account to another without the consent of the primary account holder, that's a pretty obvious recipe for disaster. The article gives no indication that Apple was served with a court order with which they were noncompliant, so it's unclear that Kate attempted to get that court order to compel either the ex-husband *or* Apple to do it.

As an added bonus, the article *also* didn't indicate whether the account was set up prior to the marriage, or after the divorce. Nothing is said in the article, and the part that describes the potential loss of photos is part of an aside that was not a concern specifically attributed to Kate's case. If the father bought the phones after the divorce and pays the bills for them, the mother is fully within her right to say that the phones stay at dad's, or buy different phones they can use when they're with her. Again, I'll concede that he should have been more reasonable.

So then why did Kate allow the kids to keep the phones on them when they weren't at their dad's? If it's because she wants to be able to contact them, Walmart has phones for $20 with $15/month service...did the court mandate that the father has to pay for the phone the mother uses to reach the child?

Abuse, certainly not. Biased, one-sided description of parenting who's closest description of abuse was "limiting screen time", "asking questions about who the parent's children are talking to", and "tracking a child's whereabouts", as performed by a person who has at least partial responsibility for the well-being of the children? I want more information before I pick sides here.

And which policy should change? "Parents" should be able to call Apple and request that "their" child be transferred from one family account to another, without the consent of the current account-holder? It'll take less than a day for that to go sideways. If it's that Apple refused to comply with a court order, the article makes no indication that such an order was given. So...again, a sincere question - what policy would need to change, where Apple can allow Kate to have her childrens' phones transferred to her account, independent of a court order, such that it wouldn't put the children at even more risk?

And which customers are these? Should the father be able to have the accounts moved back to his account without Kate's consent? Because he's an Apple customer, so Apple only has one satisfied customer in this set of people, regardless. From Apple's perspective, why does Kate get to be the happy one?

And why would antitrust law apply here? Nobody is stopping either Kate or the children from buying whatever phone they want, and tying it to whatever account they want. There are certainly other reasons why Apple may well be worthy of antitrust litigation, but I'm having trouble understanding how Apple's unwillingness to transfer sub-accounts without either a court order or the consent of the account holder proves that they're abusing a monopoly position.

1. Use something like iMazing or Photo Transfer App to pull photos and videos from the phone to a desktop/laptop, then replace the phones and apps. I'm really unconvinced that this is some insurmountable or prohibitively expensive solution.

2. What's the expectation in terms of changing how family sharing works, because they all sound like they'd be even worse. By definition there can't be two primary accounts, because that wouldn't solve the problem - she'd extend screen time, he'd retract it, and they'd just go back-and-forth with the kids caught in the middle. If Apple does what she wants and makes it possible to do a hostile-takeover of a primary account position...then he could just call Apple and do a hostile takeover if he's still on the account...And, Apple is stuck wasting valuable CSR time being caught in the middle of these two bickering as each calls to have Apple do the cha-cha-slide.

3. Devil's Advocate here...Kate's story is extremely light on details.

...okay, but there is no indication even in the TFA (yes, I read it) that this was used in a quantifiably harmful way. The article mentions nothing about him doing anything with this information that involved kidnapping or threatening behavior. The article seems to ignore something pretty important - they *are* his children, too...and especially if he was the one paying the bills for the phones (again, unclear, but if Kate was paying the bill, it seems like something the author would have enumerated), why is he a *bad* father for enforcing rules? So, there have been a hundred Slashdot stories about how screen time is bad for kids, and when a father requires that his children account for their screen time so that they don't spend massive amounts of time doomscrolling, that's 'draconian' now?

I'll grant that he should have been more reasonable about usage limits during Kate's days...To be fair to Kate, it does sound like things were uneven and he *should* have been more willing to alter the account parameters during her parenting days without having to go to court/Apple/the press. That said, it definitely sounds like there's room for there being more to the story than Kate's telling...especially because the article includes this chestnut:

So...not only does the article show its bias by calling this a "happier ending", but it seems that weaponizing children against their other parent is okay when she does it? It goes on to say:

Wow...it's "parenting their own parent" when a joint-custody situation has friction, apparently. Instead of buying new phones and creating new accounts and re-buying apps, and instead of going through the traditional court method of compelling the accounts to be transferred or to have the account better reflect her wishes on her days, and instead of telling the kids to accept the suboptimal situation and learn to work within the unfortunate contexts...she wants Apple to come up with a means for someone to forcefully migrate childrens' accounts from one account holder to another without the consent of the primary account holder? Anyone else see next year's WIRED headline about an actual-psycho convincing a vulnerable child to migrate away from their Family Group, and Apple doing nothing about it?

I'm sure the ex-husband was no saint, and yes, I understand that the amount of control here *can* be misused in a way that *is* harmful and certainly *has* been at some point...but the free option is having the kids leave their phones at the father's house when they leave. The best-for-Kate option is undergoing a set of phone purchases and a weekend's worth of account setup and app re-purchasing, which would cost $3,000 at the *very* most (likely less than half that). The fact that the problem was solved through incessant nagging makes question whether this specific case was one where the father was being dangerous. The article makes zero indication that WIRED even reached out to the father to get his side of the story, so I'm hard pressed to believe that there is a need for Apple to redesign the system in a way that could allow even more dangerous abuse.

I'm not so sure. Spamming can use the law of large numbers due to its low barrier to entry and easy automation, so even a 0.01% success rate can be profitable. With ransomware, most of the easy stuff has been picked off already - basically nobody has RDP open on 3389 anymore, pretty much everyone has SSH locked down, lots of people do their work in SaaS products which are more difficult to ransomware, and for all the obnoxiousness for Microsoft shoving OneDrive down everyone's throat, it *does* do versioning remotely, which can mitigate the extent that ransomware is a problem. Meanwhile, insurance companies are more adamant about requiring 2FA, virus scanner companies are better at detecting ransomware, and while vulnerabilities continue to be found, the overall trend of those which can be exploited for ransomware is decreasing overall.

As deployment becomes more and more difficult, and it requires more and more skill to deploy, the amount of effort increases, while the probability of payout decreases. If it takes four hours to deploy with 50/50 odds of payment, that's probably a better paycheck than the IT guys who are keeping the ransomware out. If it takes ten hours to deploy and 1/4 odds of payout...still viable. If it takes 20 hours to deploy while keeping the 25% odds of payout, automation/AI might help keep it even, but some of the smaller gangs may throw in the towel. If it takes 60 hours to deploy with the same 25% odds, the squeeze continues. If the odds go down to even 20%, but with 100 hours of deployment, it becomes a niche.

Now, if it's made illegal to pay, all that's going to do is to create a market for a middleman. The company hit is paying a consultant, and the consultant is just buying an NFT, which is perfectly legal for a company to possess as an asset. More to the point, making it illegal to pay the ransom means that the likelihood of payout is, ironically, higher, because even getting hit is going to trigger an investigation, so it's ransom *and* hush money in one fell swoop.

As easy-to-hack legacy systems get replaced with more-secure-by-default systems, either through upgrades or through closures and mergers, the process becomes more difficult, and the continued pressure from insurance companies reduces the likelihood of payout, starving the industry over time. It'll still happen, but it'll become an asymptote. Give it a few years, and it simply won't be profitable.

I was with you in the first half, but I don't think the "because money" argument is the result of wanting government handouts.

My grandfather was an Italian immigrant, who came to America right before the Great Depression hit. He only got a 9th grade education (some variants of the story say 6th or 8th; suffice it to say he didn't graduate high school), and then he worked for a defense contractor doing machine work. On one income (admittedly 80-hour work weeks were the norm for that income), on a 9th-grade education, he was able to raise three children, have a custom-built house in a then-semi-rural part of the country (that still had public transit to get to and from work), and still save for retirement.

How much money does it take to raise three kids in a custom-built home and save for retirement within two hours of a major city, and how many jobs make that much money reliably, and how many of *those* jobs can be had with a 9th grade education?

Sure, my parents didn't get designer clothes, and only had maybe three family vacations between birth and adulthood, and they didn't eat out frequently....but while there are certainly areas that twentysomethings can be more prudent about, the median family income in 2024 was $83,730, and that usually involves two incomes. 50 years ago, it was $6,600, or $69,150 inflation-adjusted - roughly a 21% increase. Meanwhile, the median home cost $20,264 in 1964, or $212,314 inflation-adjusted, while median house cost in 2024 was $414,361 - a 95% increase.

I haven't even considered the costs for everything else - Even if you want to argue that families usually had only one car instead of two, they cost more than twice as much now. Even if you want to argue that food is different now than it was then, it's also twice as expensive. Even if you want to argue that stay-at-home moms *were* childcare in 1964, the functional need for two working parents means that child care is a requirement that wasn't a need at the time.

So...yeah - I'm not a government-should-pay-for-everything, socialism-but-real-socialism-not-Cuba-or-USSR-socialism leftist, but the math simply makes it completely impossible for a couple in 2025 who isn't in the top 10% of earners to have the life that our grandparents and parents did - even if such a couple is willing to live in a 1600 square-foot home, drive a single car, go on infrequent vacations, live in an average region of the country, and have a daily diet of rice and beans - and if homes were $45K in 1964 against $6,600 median incomes, you can bet that there would be a lot less people in this country, too. ...and yes, I understand that my numbers are USA numbers while TFA is about Finland, but I'm hard pressed to believe that their numbers don't have a simliar trend.

1. Windows LTSC isn't sold retail; one has to acquire it either part of some complicated business licensing deal, or on the high seas. No shade in either direction, and I do ultimately agree that it's the best solution available, but either way is a massive barrier between this solution, and people moving away from Windows...to also use Windows.

2. The user is still, ultimately, running Windows, with all of the issues involved with that - licensing drama, patches and updates, principled freedom elements, privacy concerns, and even if the malware concern is limited to ruining the one or two Windows applications, that's still going to be a nonstarter because those applications were, by definition, irreplaceable.

3. Virtualbox is solid, but depending on the nature of the application, may still be a problem. I don't think it supports PCIe passthrough, so any hardware interfaces would need to be USB. Even at that, some devices are more picky than others. I remember having a massive problem when I first got my Phase, that it took some back-and-forth with support to realize that my troubles were caused by having VMWare Workstation installed - the USB Arbitrator Service was causing detection issues, and that was on the bare metal, not even after USB Passthrough. I've had inconsistent success with bank-issued check scanners and USB licensing dongles with USB passthrough; some have worked fine, some haven't.

So yes, we agree that for 2/3 of normal computing, Linux is just fine. Unfortunately, the last 1/3 is both the hardest to replicate, and the most critical to replicate. I've always been of the persuasion that if I won the lottery, I'd put some money behind ReactOS, because I think it's the closest solution to the Windows problem. I love Linux for what it's great at - my Proxmox servers and web servers and router and Immich server and Mail/PIM server and NAS all run Linux, and it'd be laughable to do any of these things on Windows. There are certainly some cases where something like Virtualbox or Codeweavers are able to bridge the gap, but the much smaller ecosystem and absence of a true support system will always be a liability that keeps that gap firmly entrenched.

If it's a principle thing, I obviously can't help...but I will say that I've been thrilled with dbPowerAmp.

It fully automates the ripping process, especially with multiple drives. A friend of mine gave me 500 CDs; I had them all ripped within 2 hours, properly tagged, consistently named, and to my specifications (I had five SATA drives in a custom-built tower for the purpose).

If he's really got 10,000 CDs, I'm reasonably confident that a stack of $17 USB disc drives and dbPowerAmp can get the whole shelf ripped in a weekend, maybe two. ...again, I understand that he may have other reasons to avoid doing it, but if it's purely a time thing, dbPowerAmp really helps cut the time down.

Not a bot or anything, just an extremely satisfied user with personal experience on ripping large numbers of CDs, who also found the task prohibitively daunting and was able to find a solution.

The good news is that it's closer than ever. The bad news is that the exodus is no better.

People don't run Windows because they like their OS or spend their days enjoying the operating system by itself. Operating systems are a means to an end - they run other software, and people choose operating systems based on the software they need to perform other tasks.

For all the millions of apps on the iOS App Store, Windows has a hundred applications. They do everything from weather modeling to hardware programming to lighting control to document management, from door and gate access to games, from software-defined radio usage to tax preparation and accounting, and the amount of obscure use cases in hospitals and manufacturing would surprise anyone who hasn't had to service them.

Now sure, anyone looking for a word processor, a spreadsheet, and a browser, has had options for quite some time...but let's look at the landscape. Windows 11 is trash, especially when it comes to privacy...but ChromeOS is no better. Windows 11 does plenty of phoning home that it shouldn't, but Android does just as much, if not more.

MacOS won't allow running on hardware that's not from them, and even if they did, the Mac App Store is increasingly required to get software, and it's super convenient that they offer more storage in iCloud than they ship with their computers, which conveniently don't have a mechanism for hardware upgrades. iOS is no bargain on any of these fronts; they saber rattle about privacy a bit better than the others, but let's not pretend that Apple hasn't actively fought against anyone trying to run unsigned code, or that one must own an iPhone or an iPad to run it...and that's somehow better than Windows?

The obvious answer to this is "uhm...Linux?" And yes, Linux has some massive advantages over these systems - no phoning home, no restrictions on what it can be installed on, and it's free....solving the most fundamental issues of the other. The problem goes back to the original premise: Linux is very inconsistent with what software it will and will not run. Sure, LibreOffice and Chrome run everywhere, the LAMP/LEMP stack is the default for internet services, and plenty of software development happens just fine on Eclipse...it's everything else that's the problem.

Linux has many equivalents for many things, but they are inconsistent - there's no viable replacement for Quickbooks on Linux, and even if one used WINE/Proton to emulate it, and it works, it's a case of "it works until it doesn't", and Intuit is the epitome of being unhelpful. This applies to the millions upon millions of Windows programs that do boring, invisible, critical work in industries where imperfect emulation is a liability. Even for software with viable replacements, moving to Linux typically means a whole process of migrating data and relearning lots of software simultaneously. The absence of an equivalent of the Apple Store, where one can schedule a visit with an actual-person to help solve problems, or actually-useful phone support, means that the user is on their own to figure all of this out.

So, yes, millions of people said 'enough' to the worst elements of Windows...and got an iPad or a Chromebook or an Android phone/tablet. Depressingly, the concerns about data privacy are so abstract to most people, who never truly felt in control of their own data to begin with, that it's truly a lost cause to pitch moving away from Windows, MacOS, or ChromeOS in terms of privacy. It needs to be something else, and the value proposition needs to be better than "you own your data"...especially if all of their actual-work is done on the web; there's no net benefit for privacy if someone runs a Linux computer and then proceeds to upload their photos to Google Photos and edit documents on browser-based Office365.

If a Reddit post amounts to a human-summary of a StackOverflow disussion, which itself is a complilation from a forum posts on a discussion board and a Wordpress blogger, who got *their* information from man pages and error outputs...who do you cite? Each of them validates the others in order to minimize the amount of "SEO Blogger Spam" that also ended up in the meat grinder somewhere.

The problem with the meat grinder is that the whole point is essentially to make it impossible to trace sources to the point where the actual answer came from.

More to the main point...the bigger issue is that Perplexity is looking to make money off their training data. Nobody was given the opportunity to opt-in to sharing, and none of the creators even get Spotify-amounts of profit sharing. On the one hand, I hope Perplexity loses because they're the most creepy of an inherently-creepy industry. On the other hand, just because I don't like them doesn't mean it's fair that Perplexity gets slapped with a lawsuit while OpenAI gets off scot-free because they managed to scrape Reddit first.

honestly, that's probably all it needs to be. If If they can make a GPU with 16GB of VRAM and 5060 performance, and give it a $99 MSRP, even if it involves recompiling to whatever-their-answer-to-CUDA-is, they'd give nVidia a run for their money. If they made it readily compatible with vGPU tech with no licensing fees, it would definitely be disruptive.

Temu's existence shows that plenty of people are fine with "good enough". Nvidia's threat isn't going to come from China being able to DIY their own 5090Ti, or even an A2000....it's from China being able to give people more than half the performance at a price that makes it worthwhile to just buy three of them. Nvidia will still have its market, but I don't think it's unreasonable for China to make a good-enough card that takes the bottom 20% of Nvidia's customers.

The unfortunate problem is that this is technically true, but practically...less-true.

If HAL9000 understands what an HR hiring filter is looking for better than a person, then the person using AI has a higher probability of getting the job, because the process of *getting* a job has been so heavily abstracted away from knowing how to *do* the job. "Doing well on an interview" and "Doing well in the trenches once hired" are similarly distinct skills. Those with charisma are, as a rule, more likely to be hired if they are interviewed, than those without charisma who are a savant as the skill the job requires.

The problem with the premise that those doing the work without AI will outperform those who rely on it, is that the premise assumes that job performance is a company's metric of success. It *should* be, but more that a particular company is concerned with appearances and relationships and market capture, rather than the ability for a product or service to meet the needs of the customer, the less the skill of the worker is relevant to the worker's success.

See, there's just no winning. If we move (back) to cardboard, the argument becomes about trees. If we move to that biodegradable quasi-plastic that some drinking straws are made out of, then the argument is that the change disproportionately affects the poor, since that stuff is somewhere around triple the cost of plastic. If we move to glass, then the transport of the containers becomes far more prone to pollution because of the significantly higher weight of everything. If we eliminate one-time packing entirely, then we deal with health concerns and chemicals to combat those health concerns.

As much as the plastic industry loves to lobby, let's not pretend that it's the only barrier.

Yes, because we all love walking half a mile a day in the rain...carrying groceries in paper bags...or in the cold...or in the heat...or transporting 20-kilo items...or making multiple trips...or are we just ordering everything from Buy-N-Large and no longer in-person shopping?

To the topic at hand, PHEVs are fantastic INTERIM solutions. The charging infrastructure isn't as pervasive as gasoline and diesel, so a solution that both encourages the use of charging stations while enabling the use of existing gas stations is a helpful way to handle the transition. As we get to the point where EVs can get 1,000 off a charge and/or 200 miles of range out of a 5-minute charge, and as the number of charging stations continues to increase, and the grid adjusts to compensate, PHEVs will be less desirable as their reduced EV range will start to become a liability as gas stations decrease in number.

Maybe you don't like the fact that intermediate solutions are compromises by definition, but you won't get a whole lot of folks on board with the expectation that the solution to a global problem for everyone to relocate to walkable cities.

As someone who considers Mass Effect their favorite game, all the way to actually-liking Andromeda (I read the novel that replaced the Quarian Ark DLC)...ME4 is something I'm keeping an eye on, but have zero hope about.

We're coming up on four years since the release of the teaser poster, and three years since the trailer...and by all accounts, the game is still in pre-production. *PRE* production, for longer than it took to make ME3, and a year less than the time it took to make ME:A.

This leads me to believe one of a few things: first, that the original push was likely similar to that of Veilguard, namely that EA wanted to make it a casino game, er, "live service"...then, pretty far into the production process, the MBAs were somehow convinced that turning Dragon Age into a live service was going to fail as spectacularly as Anthem, causing a massive pivot back into the single-player narrative game the audience wanted all along. It would shock me more to know that the ME4 story didn't follow a similar chain of events.

Second, I would completely believe that there's a whole lot of internal bickering within the team about how to progress on...basically everything. Amongst the stories about why ME:A was such a disaster was because the team was forced to use Frostbite, which was made for FPSes but not RPGs, leaving the team to waste time building systems for Frostbite that already existed in Unreal...well, it was announced years ago that ME5 would use the Unreal engine again, so 'building the infrastructure' doesn't factor in this go-round. Art assets certainly need to be created, but the use of Unreal means that a good number of assets from the remaster can be utilized, so it's not that sort of baseline stuff.

Let's talk about ME:A for a quick minute...what made it such a letdown had nothing to do with the crappy animations or the fact that there was some transgender NPC in the game...the story itself was the problem. First off, the game's runtime was massively padded with fetch quests - lots of driving around and...retrieving these rocks, scanning this structure, getting this plant...if the missions were limited to unlocking the vaults and fighting Architects, the game would be ten hours long. There *were* pieces that were interesting, but they were few and far between. In terms of the actual narrative...there were two alien factions, the oppressor and the oppressed, and we sided with the oppressed fighting the oppressor...no sign of the nuance or complexity that made one stop and think that the Salarian Dalatross might *possibly* have a point, that curing the genophage might not *actually* be a good idea...just an enemy who wants to destroy everything and a Fern Gully/Na'vi, everyone-gets-along-with-everyone underdog race...that's it. Oh, and dialog choices that never *actually* matter. Either ME4 is spending so much time in pre-production to overcorrect for these issues (I concede that it takes a LOT of time to do either branching outcomes or multiroute missions like the garage pass in Noveria), or it's a bunch of people going to work and arguing all day about how to do things, and the game will be built on a stack of compromises that will please no one.

The fact that they're five years into production - longer than any other game in the series' development cycle - and they still aren't into the actual-production phase - leads me to believe that the game is going to absolutely wreak of compromise at every stage. If the game is in development hell at this stage, I fully expect the release to be about as smooth as Cyberpunk 2077's...and I *don't* expect the sort of stability patches that made ME:A playable in the months to follow.

Really, my pipe dream is that EA will sell the franchises to other developers and stick to a publishing exclusivity deal...they could probably make some decent money doing that with some of their older franchises if their plans involve sticking to FIFA and Battlefield...but I have no hope that Saudi leadership is going to do any better with ME or DA or SimCity than the former owners did.

16GB of RAM, I'll grant is fine for standard use...but Apple really needs to come up with some sort of solution for storage expansion beyond "bag of USB accessories" or "2TB of iCloud storage"; most Mac owners end up with both.

Sure, 512GB is fine for Apple Chromebooks, but video editors easily end up with either an external storage array or having to do "the project shuffle" of data management that is an absolute chore. There are more than a handful of PC laptops that offer multiple NVMe slots, so 8TB of internal storage is an option on less expensive laptops, that isn't possible on a Macbook, at all, for any price.

Sure, it's not everyone...but soldered-RAM and soldered-storage means that a nontrivial amount of the Mac market (which is disproportionately musicians, photographers, and videographers) is stuck buying the other half of their computer in pieces, including a port replicator to plug them all in simultaneously. This isn't an argument that Apple needs to design some Clevo monstrosity with interchangeable GPUs, only an acknowledgment that there is a sizeable market demographic that is artificially hamstrung.