That one upset me and made me nervous.

A huge issue with the DragonAge game was that they spent a massive amount of development time trying to turn it into a 'live service' game, then pivoted to a classical single-player experience. Putting the sociopolitical messaging aside (and whether it was 'real' or 'perceived'), a development cycle that completely pivots that sort of underlying, fundamental paradigm shift is going to undo a massive amount of development work...and EA still found it necessary to have the devs do a bunch of last minute 'crunch' and 'ship now patch later'...which meant that the early reviews reflected some of the rough edges, and then the sociopolitical messaging accusations at the height of "anti-woke" sentiment was just the icing on the cake...and thus, the flop.

So, my concern was - and still is - that the suits at EA are going to grab their Excel spreadsheets and see that FIFA still makes a mint, while single player games don't...and ignore everything else and blame DragonAge's bad numbers on the absence of microtransations and lootboxes and season passes.

They've been real quiet this year on the Mass Effect front; I'm nervous that they're pivoting again because of the DragonAge failure, rather than looking at Elden Ring or Baulder's Gate or Cyberpunk 2077 as evidence that single player games can, in fact, still make money.

I did, at least to ChatGPT. I summarized here; I'll give you the transcript if you like.

Because that's what Microsoft is pitching. I gave ChatGPT a hell of a lot more specific information regarding equipment, already-attempted procedures, and intended outcomes, than about 90% of the people at work who call me for support provide...and it STILL waited until after I bought what it said to buy, to tell me it was the wrong controller. It could have said, "this is one possibility, this is another, depending on the exact specific LED strand you're working with; I can't conclusively determine that based on what you've provided..."

It's what the general populous is going to demand in exchange for their keyboards and mice. If the goal is for the computer to understand what the user wants, even when the user is being vague, then there's no way that Microsoft is going to get there in five years or less.

I don't care that ChatGPT was wrong in this particular instance; it's a learning experience for a hobby. No problem. I *do* care that Microsoft thinks that it will take a very short amount of time to get from that state to Jarvis, or the computer in Star Trek, in about as much time as it takes to graduate high school.

I asked ChatGPT to help me get a string of LED lights to work. I spent half an hour following the instructions; I tried the Tuya app and the first party app, made a dummy account, tried the AP-mode instead of the BT-LE mode...never, ever got them to connect properly. I gave it the exact model number on the back of the unit, I gave it links to the exact product, I told it the quantity and color of wires in the lead, and I was still on the version 4 model. It helpfully recommended the QuinLED Dig Quad board, a super cool ESP32-based controller to replace the craptastic Tuya garbage that came with it.

I waited a week for the board to arrive, and I connected it all up...spent an hour of faffing around with no ability to control color or brightness...only to find out that after ALL of that, the Dig Quad was the wrong board because these were analog LEDs rather than digital ones.

...So now, Microsoft wants to tell me that they're going to totally overhaul Windows to use *so much AI*, that it will basically be able to read my mind and do what I want it to do by me giving it vague parameters, and then being accurate? They're pitching Jarvis as something they'll have working properly in four years, to the point will actively want to be talking to their computer (along with everyone else in their adjacent cubicles), and it'll be desirable...but today, the models can't accurately assess which LED controller to recommend when given EVERY piece of information that an informed CSR would provide?

...Given that very few people use Copilot by choice, and given that previous attempts to overhaul Windows have been niche at best (almost 15 years into touch-based computing and *how many* Windows users leverage a touch screen even 20% of the time?), and given that existing models are useful but far from indispensable, and being that there is already a growing resentment from the sheer volume of "AI Slop" that's making the internet even less desirable to use for many.

...Occam's razor is telling me that this is just Nadella trying to avoid the stock price from cratering by give some sort of assurance to shareholders that the bazillion dollars they've spent on GPUs weren't wasted.

I know that Slashdot's unofficial tag line is "tomorrow's news next week"...but as a Nothing Phone 2 owner, the NP3's floundering is completely unsurprising.

They hyped up the reveal, as they tend to do, and when release day finally came...the entire comment section was like, "wait, wut?"

I don't *necessarily* think the $800 price point is the dealbreaker, at least in a direct sense. The bigger issue is that the NP3 shifted from the Glyphs - a unique function that all of their other phones have - and moved to the 'Glyph Matrix'. This wasn't a completely bad thing, but they took away the regular glyphs to do it. Now, I understand that the Glyphs come across as gimmicky for most, but it's a very visibly distinct element of their phones, so messing with it was a gamble...and it seems that Carl Pei is learning that a niche vendor needs far fewer people disliking their product for a company to have a 'New Coke' moment.

To continue, the press releases and announcements leading up to the NP3's release seemed to focus on their AI functions. I'm not convinced that was wise, because I think it's much more difficult to compete in that space. For the folks who *want* AI in their phone, a Pixel 9 costs the same as an NP3; I don't think Nothing is going to favorably compete with Google on those merits. Similarly, the 3(a) and 3(a) Pro releases spent so much time focusing on how amazing the camera was...it almost seemed like they were selling a camera that incidentally had a phone bolted onto it. This is a selling point for many, granted, but Samsung and Google and Apple all have fantastic cameras on their phones. Again, these devices seemed to compete in the most crowded of spaces.

Meanwhile, the areas Nothing *could* be competing in, they aren't. Batteries aren't removable. Headphone jacks aren't present. MicroSD slots are unavailable. Bootloaders are still unlockable (for now), but LineageOS only *just* got official support on the NP2, two years after its release. They don't offer any first-party alternatives to the Google services; no custom e-mail addresses or a variant of F-Droid for their custom apps and Glyph-enabled ringtones.

I can appreciate the desire to scale to the point where they compete competitively with Apple and Samsung, but pissing off the base isn't a worthwhile way to do it...and I think enough people are voting with their wallet that Carl is getting the message. My guess is that they'll try marketing it a bit more for the next month or two, and if they don't sell, they'll bring the price down to $599 to clear out inventory and fast track the NP4...which will probably have an $899 price point, but at least this time the people in the focus groups who say "don't do that" will probably be given a bit more credence, so the phone might actually be desirable.

...instead of having enthusiasts work on their hobby if it matters to them, Drew proposes that libraries - notorious for funding constraints and limited staffing - offer those same self-hosted services? ...And, I assume library staff is going to provide tech support for this, right? And he wants backups, right? And 100GB/patron means that it'd take less than 10% of patrons to exhaust the amount of storage most libraries would budget for...AND, are libraries supposed to keep data if users move out of the district? Also, my last discussion with my library's IT folks indicated that they got rid of their local servers and it's all on AWS and/or M365 now anyway...

And let's even assume ALL OF THAT was solved...he didn't like the function set...what's the library going to offer that will solve that problem? If it's software who's functionality he didn't like, that has nothing to do with self-hosting. If the library software solves the problem...what would be the problem self-hosting it since he's got the gear anyway?

Ultimately, it's super unclear what the point is here...except that, apparently, he wants to externalize hosting onto taxpayers instead of having to make the purchases. To his point, self-hosting is a hobby, and it's not for everyone...but it's super unclear how he's suggesting "a future we should be fighting for". People that don't care about privacy or transient availability of data are served by Google and Microsoft and Amazon. The people that *do* care about those things aren't going to trust their local libraries much more than they'll trust the big tech. So...why is this a better future? It's still trusting one's own data to someone else's computer...

Because touch screens are generally easier to break, and has a chicken-and-egg problem with desktop software - relatively few software titles have design layouts that lend themselves to touch screens, which means there isn't a huge demand for it, which means there isn't much of a benefit to redesigns. Most people who I've seen with touch screen laptops only really use them for scrolling anyway, if they use them at all. On a tangential note, I had a client recently *demand* a touch screen for a new desktop; I attempted to dissuade him since a screen that fit the bill was going to cost more than the computer he connected it to...and he was only using it to pinch-zoom PDFs...but he paid for it, and he was happy.

This is exacerbated in Apple World, where the major reasons to buy Macbooks are Xcode, Logic, and Final Cut...none of which are terribly improved by the existence of touch screens, and which cost more to repair/replace under AppleCare.

So yeah, they're kinda like the CoPilot key - a handful of people actually-use-it, but it's marketed as something that is mainstream, which is technically-correct, but in practice, a whole lot of people who have one, wouldn't miss it if their next laptop didn't include it.

So, I get what you're going for, but I would submit that there were more than a few missing steps here.

While the Nokia 3310 and similar models were the 'first phones' for many people, between about 2003 and 2007 there was a whole lot of 'feature phones', and they all had 'fun' names, like the "LG Chocolate" or the "Samsung Juke". They were frequently defined by a keyboard that allowed for full-blown texting beyond the T9-based input you're describing. Early iterations of smartphones, like Blackberry and many of the HTC phones running Windows Mobile, were also popular among consumers and enthusiasts. Teens frequently exchanged BBM PINs, and IT departments were a fan of BES, arguably the first MDM.

As for Nokia itself, it had a pretty solid following in Europe with its Symbian phones. They weren't as popular in America, but a solid foothold, they had.

What Nokia *really* lacked, however, was an analogue to iTunes. Nokia actually had some pretty good phone management software over the years, but it was tucked away as an afterthought, rather than Nokia treating it as a first-class reason to have a Nokia phone. Without it, moving from a Nokia 3310 to a Motorola Razr was just a matter of moving over contacts, rather than shifting ecosystems. With no ecosystem of their own, Nokia had to complete on features to retain loyalty...and then the whole Steven Elop thing happened, and that was pretty much curtains for Nokia as a consumer-facing brand.

...Because they didn't do a lick of marketing.

I had a Nokia 6.2, 7.1, and 7.2 phone...and they were all fantastic. They included one management app, a few wallpapers, and a few custom ringtones...but after that, they were bone stock Android phones without all the extra gunk that Samsung adds. This is a blessing and a curse; I appreciate that they didn't attempt to reinvent the wheel, but they also had nothing notable to set them apart. The Lumia phones were at least visually unique and had solid cameras for their day, but while I appreciate Nokia showing some restraint with the shovelware, it also meant that they were slightly-cheaper Pixels.

They had a handful of other issues that make me completely unsurprised they didn't make meaningful inroads. First and foremost, they weren't sold through carriers. They were exclusively retail/aftermarket phones. I got mine at Microcenter, but that's because I made it a point to ignore any of the free-upgrade or installment-purchase offers from my carrier. Most people get their phones from the carrier; the absence of that option severely cut down their potential customer base.

AT&T also screwed over the handful of users who had them back in 2022 by mandating VoLTE, which the phones didn't support. In fairness, this also caused issues with a number of slightly-older Samsung flagships as well, but that didn't help, either.

Finally, it was ironic that most people's recollection of Nokia phones were that they were indestructible, the 7.x and 6.x phones I had scratched easily, and had screens that were more delicate than other contemporary phones at the time. To add insult to injury, there were far fewer choices for protective cases - they existed, sure, but they were almost never available retail; Otterbox only had options for a subset of Nokia phones, even for mail order.

So yeah, it's completely unsurprising that a revival of the Nokia brand didn't work out well.

This problem was solved decades ago: private servers.

Anti-cheat mandates for public servers? Makes sense. But the way to allow online multiplayer without invasive anti-cheat is to allow users to host their own game servers, so the anti-cheat is good old fashioned community pressure - when everyone knows everyone and games are password protected, it's trivial for a community to self-police to a level they're comfortable with - including servers that expressly allow for cheats, if desired.

I'd almost go so far to say that in a few years, it'll be the only way to combat cheating - there are already methods of cheating being developed that perform their cheats via screen captures and input emulation, rather than memory sniffing. Once the tech gets a bit better, it'll be basically impossible to detect, because the actual cheating is happening on a completely different computer than the one the anti-cheat software is running on...and even consoles won't be immune.

So...small communities that can expel suspected cheaters is my solution to the cheating problem. It worked then, it'll work now...but it would be one more thing for the game companies to support, AND it would remove their ability to incentivize the X+1 release of the game because they couldn't remove the old servers, AND it would remove a good amount of the draw to the PSN and XBL, so I don't anticipate it happening any time soon...

Ahh, yes...because it was *top* priority for the Biden administration..or even got a mention on Harris's "four years of JOY!!!11" campaign trail that followed her four years of being VP where she could have attempted to get a subcommittee together in the Senate.

Or because Gavin Newsom or Kathy Hochul or Maura Healey have made it any level of a priority for businesses in their respective deep-blue states.

The Republicans certainly couldn't care less about the issue at all...but let's not pretend that the CFPB cracked the top 20 of priorities for Democrats.

So, these are a few things off the top of my head; I tend to limit my usage to only smaller installs, so consider this more of a "stuff to Google for clarification" list than a definitive set of information...

I think the biggest thing is that there is no analogue to vSAN. It'll mount iSCSI and NFS targets, and its ceph implementation is at least on par with VMFS, but larger installs that depend on vSAN tend to be underwhelmed.

The Proxmox Datacenter Manager, which allows for live migration of VMs between hosts, is still in an alpha state. I've had it work pretty well; it's quite polished for something being described as being in its alpha stage, but the functional equivalent of vMotion is still lacking.

Meanwhile, support is not quite at VMWare levels. Obviously, post-Broadcom, VMWare support took a nose dive, but Proxmox does not offer direct phone support, instead depending on resellers to do so.

Beyond that, I personally found the UI to be rather unintuitive, for example storage is defined at a 'datacenter' level, rather than at a 'host' level. PCIe Passthrough can be a bit...special, compared to VMWare having that be a trivial matter. Also, more sophisticated networking configs are more "Linux-y" in Proxmox than in VMWare, which uses a more traditional switches-and-ports paradigm that's much easier to understand and visualize. I also found VMWare's storage to be simpler, in that a datastore can have thin-provisioned VMs, thick-provisioned VMs, and installer ISOs all sit next to each other in harmony, while Proxmox gets more...particular. For example, an LVM-Thin volume will thin provision any VM stored to it and won't allow QCOW2 virtual disks to be added to it. ZFS storage is more flexible on that front, but it handles snapshots differently...

I say all of this as someone who either has moved, or will move, all of my VMWare clients to Proxmox. The feature set is more than enough for all of them, and as an added bonus, Proxmox is a *lot* less picky about hardware; I've got a box full of 10GbE NICs that got a new lease on life because VMWare decided the PCIe 3.0 Qlogic cards were 'too old', while Proxmox will still send traffic over a 10-Base-2 BNC network through an ISA card if I gave it one.

Of course they do; it's just more insidious. Try installing software without an iCloud Account...it's getting more and more difficult to get downloadable DMG files anymore; even open source Wireguard doesn't distribute a client for OSX independently of the Mac App Store.

Now, once one has the almost-obligatory iCloud account tied to the Mac, the nags come when the storage runs out...because while us PC folk can install 4TB internal SSDs, possibly two of them or even more, there is no way to upgrade the internal storage on a Macbook.

So, the choice is to either walk around with a USB external drive forever...*or*, capitulate to the conveniently-placed notices about how iCloud can seamlessly put 2TB of data in iCloud for you and manage it automagically, so you never run out of space, for $10/month. ...so yes, Apple has its nags; they're just more transparent because storage isn't the primary sales pitch for iCloud+...and while Microsoft is assuredly worse about the nags and notices, Apple isn't innocent.

Because - and I can't believe I have to explain this - if you have the generated value so that you can validate your identity to the computer, there's no way the computer can verify whether I paid you $50,000 for that code. There's no way for the computer to validate whether the data exports I perform are for backup purposes, or to extort the company.

So, whether we're dealing with metal keys, or 8192-bit SSL certificates combined with a 24-character password and an iris scanner...the human holding the means of access can use that access for good or for ill, and there can be motivations for both of them, including bribes. Whether one does so for money in the context of a "paycheck", or does so for money in the context of a "bribe", is fundamentally impossible for a computer to ascertain before it gives you access.

There are many issues with the movie industry at the moment...but I think that many of them can be addressed simply by making cheaper movies.

If a studio commits to spending half a billion dollars on a movie, they're going to enforce extremely rigid parameters. If Gunn wants to get half-billion dollar budgets, he's going to have to deal with the fact that he won't be the director - the Board of Directors will be.

The more practical approach is to pursue less expensive movies - $10M-$50M is a much better ground to work with, because the suits won't be as rigid on their direction. More creativity can flourish because there won't be as much pressure to be a paint-by-numbers film. There won't be expectations of making a billion dollars; if five $20M movies get made and one makes $200M and the other four only break even, the studio covered their costs and still doubled their money.

So, yeah, apparently James is discovering that big piles of money come with rules. News at 11.

Ordinarily I don't reply to AC's, but let's unpack this...

I'm sure you do! But I'm also sure that it doesn't involve the official Cricut app, since they only have a Windows/Mac download on their website. You can probably use the browser-based version of Cricut, or perhaps you've gotten WINE to work with it, or maybe there's some sort of OpenCricut solution, and maybe there's even a bash script running around the Cricut forums that require little more than downloading and running the script to have every dependency met and have it 'just work' the way it does for Windows. I'm sure I could tailor my examples a bit better, but fundamentally, one cannot get a Cricut working on Linux purely by following the Cricut-provided documentation, without using the browser (i.e. dumb terminal mode).

Okay, maybe it's gotten better. It'd be nice if Disney+ worked with an out-of-the-box browser on Mint or Ubuntu the way it does with Edge or Safari. Maybe I should have used Peacock as my example, or maybe my point was that most streaming services work inconsistently on Linux due to the DRM components that are unfortunately necessary in order to use them. Personally, I prefer Mint/Ubuntu shipping without the Widevine components and prompting at first setup, but there have been multiple articles here, on Slashdot, over the years, regarding the back-and-forth compatibility drama with various streaming services; I'm sorry I picked another imperfect example because D+, apparently, works at least some of the time.

Never said selling on Etsy was a problem; it was just a means to an end to describe a common scenario where a non-technical person might have a need for something more than a browser and a word processor.

Never said it wasn't possible, I went out of my way to indicate that drive syncing *is* possible on Linux (Mint, specifically), and that it *usually* works, but Google and Microsoft both mess around with the underlying protocols that sometimes cause those services to break. I put the blame SQUARELY on Google and Microsoft for this, but it's still something that isn't a problem on Windows.

Never said it couldn't...I literally pointed out it's possible to do so on a browser, and that Clementine works, but it'd require a whole process to import an existing library that involved python scripts, rather than an "import library from iTunes" button, for the users that still leverage that workflow (and I know *plenty* who do).

And, cue the broken record, I *said* that scanning works. I said that a specific function that is present on Canon/Epson/Visioneer drivers, which is a huge timesaver, isn't present on most iterations of Linux-based scanning software.

I'm sure you haven't. I'm not saying it can't be done. My *entire* point was that one will either run into higher barriers than Windows or OSX require, or the onramp is made easier because nearly all of the data and actual software has gotten externalized onto cloud-based solutions. Each user is going to have a different amount of tolerance for those barriers, but the draw to Windows isn't Windows itself, it's the laundry list of applications that Windows runs, and Linux doesn't. To switch is to have enough of a principled stance that one is willing to deal with data conversions and migrations, subscriptions and externalities, and workflow shifts in order to get back to the starting point...and that Linux's value proposition, on a desktop, unfortunately doesn't always even out.