LLMs have improved my education, work life, and general knowledge search. I get more time to spend living my life and less trying to find one dude’s stacked change post from 2009 that fixed my problem.

They allow me to access information the way I learn and operate in a way that textbooks, college education, video courses, or online classes have never allowed me to do in my entire life.

That being said, the general AI buzz and buzzwords need to die, the real positives need to be celebrated.

Really curious to see how the modem stacks up, but I also guess reviewers given demo devices were told to not talk about or play down the modem performance, rather than fix the problem.

Hold on now, let us not start using logic when referring to computer programming here. It’s all magic and bumbling around on a keyboard.

All of the data you mentioned, voicemail audio included, would be about 10 megabytes.

I’d like another choice, both platforms have become terrible metric-collecting, dark-pattern-infused data-harvesting trash. (iPhone users, don’t believe the commercials claiming privacy.)

Because they love and believe their corporate daddies they so worship, especially in the US.

I probably would consider them, but their phones tend to have lackluster US carrier/band support and lack of security updates. Coupling that with the high price tag, no go.

I was trying to think of a mental image that was internet-safe to imply something like you’re stuck in deep mud that is pulling your boots off but at least you survived. Or less internet-safe, you’re trying to get home and have no ride and for some reason the weather turned to a storm of actual shit and you’re walking home with shit raining down upon you but you must survive.

It definitely wasn’t scientifically accurate.

Hey, friend. Definitely not saying Windows is better.

That being said, in this example, a UI open dialog that will blindly consume any file fed into it with behavior bad enough to crush the entire computer into absolute uselessness is… let’s just say, not something one would want to run into on a bad day just trying to deal with life stuff.

Now, if one were piping /dev/urandom into the framebuffer expecting to load Firefox, that’s another matter.

Hey, good to know. It was one app that was only available as a Snap. Will look for alternatives.

And jumpers and terminators and make sure your SCSI IDs don’t conflict!

No diss. I read that slot comment above you and went, “yep, yep, yep, makes sense. Man our standards are often dumb.”

Laughed with joy at your comment, because I totally get how foreign this shit is to so many people. It’s like if I walked up to a building engineer asking how they know that iron beam is safe for another 50 years via their skills and I’d just be like “…do what now?”

You know. It’s interesting. I’ve been trying out Debian 12 with KDE Plasma. It actually has been a joy and feels like what Windows 11 should have grown into, had Microsoft actually been designing software with the customer in mind.

…but then there have been times where things so easily critically break until you fix them. Don’t get me wrong. I’ll go mess with kernel code if I have to, so I’m comfortable, but… I just want my computer to work. Windows, for all its shittiness, still keeps working through it like a slow cargo train pushing through a park piled in millions of pancakes.

I had one event the other day where I was installing a Snap app for the first time. Decided rather than installing the Snap package manager because I wanted to avoid Canonical if possible, I’d just manually put it in /opt. Figured out how to edit the KDE “start” menu to add the app using the included GUI tool. Wanted to use the app’s icon. The snap app had an icon embedded in it that Dolphin file manager recognized and displayed.

So I went, “ok, sometimes applications can parse out images from binary files. I’ve seen this work for decades,” so I tell the menu editor to ingest the snap binary for the icon, to see if it will scrape the icon. No icon showed up, so I found a a svg online and assigned that to the icon.

Then I went and saved and launched another application.

GUI slowly started not working and eventually the entire OS locked, even the alt text consoles would not load. Ctrl+alt+backspace was dead, caps lock died, which was when I knew, “he’s dead, Jim.”

Tried rebooting, tried launching that program again, (bearing in mind, not the program I manually added to the “start” menu) and every time the whole OS freezes up. Tried launching apps in different order, launching from command line, etc. When the one app launched that wasn’t the one I created a launcher icon for, same thing. Freeze. (It is possible that the bug is in fact time-based or boot-sequence-based, and since I was trying to reproduce the bug rapidly, the other app had nothing to do with it.)

I go remove the start menu link, hoping that, what I assumed was part of Plasma was trying to load this binary as an icon even though it should have checked the file, recognized it as “no I can’t parse this,” and done nothing or displayed an error or parsed it and showed the icon. Especially after I assigned it another image. I just hoped whatever screwed up would be connected to the code executing that app launcher icon config, and deleting the config for that application would delete whatever mess that was created, and hopefully was created discretely.

Shit you not, the computer became rock solid stable again after that and one more reboot. Hasn’t glitched since.

It’s shit like that that makes me proooobably give up on this experiment and end up on a commercial OS like MacOS again despite the cost and downward trend they are also suffering in a lack of innovative energy.

https://www.livescience.com/archaeology/remains-of-7000-year-old-sunken-stone-road-discovered-off-croatias-coast

Nah

They’re made very aware of things on social media, they internally have tools that scrape social media to surface hot issues. They should be leveraging them. They likewise have instrumentation to see how users use every element of the operating system. If they see that widget being installed->deleted, they should have the metrics to drive “something is fundamentally wrong with this thingy.”

The poor cartoon duplo UX is a fundamental design problem. You’re just buying into the whole corpo “hey we do a beta so you get a voice (but actually we’re just lazy and laid off a bunch of QA to crowd-source their job for the lulz)” experience. It should not be the job of the end user to (unpaid) tell the corporation how to do their job.

Posts like OPs are just surfacing an, “I’m annoyed at this dumb thing” user experience. Venting to others is cathartic, corpo should be smart enough to gather this intel. Filing a bug report that could lead to 60 follow-up questions from the corp that they expect you to keep submitting more information and will likely not lead to any meaningful change is just pointless (because it’s unpaid) work that their engineers should be doing.

Although if a person likes to donate their time to the company they already spend thousands of dollars at to improve the product that shouldn’t have been shipped broken, that person should go ahead and do it.

Having left the platform for 5 years and migrated to Android, been considering migrating back due to Android not showing a promising future.

Playing with iOS again to see if I can tolerate it, it feels like everything is giant font giant icon Duplo Brick UI. Which could very well be why they’re so terrible at space management. Do they realize some people have eyesight? That widget has a terrible amount of wasted space, but then so does most of the OS.

“Please do Apple’s UX work for them, unpaid, of course.”

Smartwatches are a really interestingly sad storyarc.

I got into smartwatches early on with Pebble. It was the correct balance of battery life to functionality. Then Big Tech accelerated to, “let’s run a phone OS on a watch” - which came with terrible battery life and sluggishness. Still, the OG Moto 360 was actually “not bad”. The LG Watch Sport added a SIM card slot and a cellular modem. Now we’re cooking with gas! I’d trade off bad battery life to have a parasite phone on my wrist. Also Google acquired and killed Pebble, because of course they did.

At the time of the LG Watch Sport, T-Mobile also released DIGITS, which made it so I could cobble together a parasite SIM card that receives my calls and texts on the watch and build out what “modern smartphones + smartwatches” do without the high bill and vendor lock-in. It also had cellular antennas built into the strap, so you couldn’t replace the strap, but you at least had decent RF.

Apple’s watch came out, and showed promise, but to this day suffers from a few critical bugs that they’ve never completely fixed.

Bugs, namely:

• A dependency on keeping iMessage turned on to send/receive SMS from the watch. The watch can’t do any messaging directly, it has to use Apple’s cloud via data. A watch is a perfect use case for simple text messaging!
• The biggest: there’s a continual glitch where WiFi calling and/or cellular calling will get screwed up. You won’t know it until you’re away from your phone and suddenly have to place a voice call, and can’t. This is a core feature of having cellular on a watch. There is no way to resolve this other than backing up the watch and resetting to factory defaults and then restoring. This will happen every few months, you won’t know when it happens until you need the feature.

Then they all started throwing health features into the smartwatches. Likely to try and vendor-lock you into a platform. I tried some Withings watches for a while, and their hybrid (what I always call “dumb-smartwatch”) was a refreshing take back to the Pebble days with a bit of style. Unfortunately, Withings saw the sweet sweet candy of medical industry money, and their smartwatch line has really stagnated while the app rots on the vine.

I’ve been maintaining periodic cross-links to maintain health sync so I’m not vendor locked in to one set of data, like Samsung->Withings or Garmin->HealthSync->Somewhere else or the most convoluted at one point was like Samsung->Withings, Withings->Fitbit (with a donor old Fitbit used just to get the app up but then not carried) then Fitbit -> whatever health app I was using at the time. So that whole thing is a topic itself, that getting your health data around is a complex chore that nobody should have do deal with. Yet the vendors make that health data so constantly in your face! Time to sit, time to stand, time to breathe, time to drink water, DANCE PUPPET DANCE! On Apple’s platform, their health app does make cross-sync easier-ish, but also in a lot of smartwatch forums, there are many posts of duplicate data or data from the wrong user cross-syncing, so something is funky there too.

Samsung had one good smartwatch as far as I’m concerned and it was the OG 46mm Galaxy Watch with cellular, running Tizen. It had great multiday battery life, cellular capability, enough storage to put a few playlists in it, the physical rotating bezel to select UI items with a click where each click meant one menu (throwback to the old BB 8700g what what!), all the notes of being a device on your wrist that lasts a few units of time and works on its core function. It even had a barometric pressure sensor on-watch so you could see if a storm was coming without Internet.

It seems, especially with Big Tech all having AI hardons now, that they don’t know what to do with the watch lines now. The chipsets really haven’t accelerated like they should. Qualcomm took entirely too long to get their watch chipset power requirements down. The 3GPP spec for 5G IOT is mostly finished but that doesn’t mean chips exist, there will be many years until the chips start showing up in watches. They also also really haven’t nailed down thermal issues. I was once at cell edge on GW 5 Pro, and 10 seconds into placing a voice call, the modem became too hot and the watch went into thermal throttle mode where it sleeps everything until it cools. How could that ever be depended upon?? (That was actually the line for me giving up on caring about a watch with a modem. If you can’t call 911 for more than 5 seconds, what’s the point?)

Then, since carriers have always forced vendor-lock for pairing of smartwatches now, and smartwatches no longer have SIM card slots, you can only use Verizon post-paid or AT&T post-paid to pair a watch, forcing expensive post-paid plans (except the weird outliers like Visible + Apple Watch, or Fi with Samsung watch) and now they’re raising the rates to $15-20/month for a watch that might use 20KB of data a month!

Now Google’s watch can’t even be repaired? These companies want this tech to die.

Instead they could have been looking at/heading towards wrist cuffs like something out of Death Stranding that fully replaces the need to carry a pocket computer. Which they would hate, of course, because then you’re not buying 5 devices, you’re buying one.

Garmin might be the only company doing smartwatches right these days. They focus on their core functionality and iterate. They tried LTE, realized it stunk, and gave up. They have solar charging to boost battery life, low-power tech like memory-in-pixel transreflective displays, and great multi-day battery life. They don’t have all the bells and whistles of other brands, but, they seem to actually want the product line to succeed…and they’re not trying to nickel-and-dime users with monthly fees.

Good theory, I like it.

I’ve wondered for quite a while if there was ever any truth to the rumors that Meta apps “listen” to people. It doesn’t make logical sense, as the OS should expose they are doing it. Unless they had API access to allow microphone access without triggering the microphone icon. These companies have had API access in the past, like when Uber had full screen display capture access, and Meta definitely has some agreements with the fruit company to access some kinds of data. Or, when iOS first introduced the location tracking symbol in the status bar, I was able to write a program that allowed gathering of location access without actually triggering the icon.

Most of the time, the events can be explained away by knowing how adtech works, like, I was drinking a beverage, a friend asked what it was, the next day I started getting ads for it.

In that case:

• They were on my wifi network
• They picked up their phone when they were asking about it and did an Internet search
• So once the GeoIP was cross-referenced across ad providers, the IP started being targeted for those ads, makes sense

Some stories I’ve heard are more strange.

It does make me wonder if it was true the whole time, but they now have to ask for permission.

Don’t use the platforms myself or I’d try and set up a test experiment.