Is That A Voice From Your Pocket?

Posted on by Sarah A. Hoyt

Recently (because once we started this move, having just enough money for the move and of course losing my income for several months, since I haven’t finished anything, everything had to break and be replaced, bit by bit) my phone battery decided to convert to Islam threaten to explode. So we had to get another in a hurry. And, of course, it cost me way too much to get in a hurry.

And then I was faced with the true horror of AI use.

Sigh.

It’s not the AI, really. It’s the idiot things the phone thought to do with it.First, there’s no way to uninstall the native AI without installing another. So I installed Google’s Gemini to cancel the native AI.

What can have led to these desperate measures you ask? Well, because (it’s apparently a known problem with this model) it invokes itself for no reason anyone can imagine based on what it heard you say and after searching its database.

All of which would be PERFECTLY fine if once it got invoked it didn’t speak in the world’s smarmiest voice from my pocket.

The first time this happened, it said, in a calming tone, “Keep compressing. Do you want me to call emergency services?”

I was mildly startled by this, since at that moment, we’d just pulled into the garage and neither of us had spoken for minutes. My answer of “What in heck?” luckily didn’t get interpreted as “Call 911 and have the armed forces scramble jets.”

Afterwards, I looked at the listing of what it thought it heard and it was: rub, press, rub, press.

I have no idea how it interpreted that from the sounds of opening a garage door (I guess?) or how having gotten that it jumped to “Treatment for a perforating wound” instead of, you know, cleaning a stain on fabric or, you know, other things.

This was the first time and I thought “Weird, but okay….”

The next time it got invoked, I’d told Dan something about painting the porch, and from forth my right pocket bursts the smarmy voice, “The Sistine Chapel was painted in fresco by Michelangelo between 1508 and 1512–“

That was… odder. And I was starting to get slightly testy.

But then I was talking to my son, trying to coordinate cars and u-haul, and into the middle of this conversation bursts the world smarmiest voice trying to give me instructions on…. I don’t remember, but I think it was on how to prune roses.

I had a moment. Not right then, because I just turned it off, but after we got the u-haul in place I spent half an hour screaming on how stupid the AI was and how I was sick and tired of it getting invoked.

But we were moving and busy and … I let it run, other than looking it up and realizing I could only drive out the demon by installing another– I mean, I had to install another AI to replace it.

Until, a week later, I was feeding the cats and as I reached past Muse I scared her and she started to run away, and I said “What’s so scary about my reaching for the food?”

And from my pocket comes the smarmy voice, “It’s not scary. Do I get some? Yum.”

Look, I can’t even imagine what kind of mental aberration went into the programming that my statement got interpreted as a question for AI nor that that answer was appropriate. I’d imagine the appropriate answer would be something more like “Please repeat question?” BUT no. It said that, with a jokey undertone, like ahah, I’m just another human, here in your pocket, talking about eating.

Look, I don’t know why but that came across as one of the CREEPIEST things I’ve ever heard, and I sat through at least one Clinton speech. I about jumped out of my skin, and ended up settling for gemini, which, thank heavens can be disabled so it doesn’t answer on voice. Even if it does volunteer to look up words from news articles for me. Difficult ones, you know, like coffee or cat. And no I have no idea what secret combination of finger taps and Dan snoring by my side sets THAT off.

“But Sarah, I thought you liked AI!” I do like AI. It’s an excellent tool in its place. It’s the enthusiasm for pushing it everywhere, including places it’s not wanted that disturbs me.

Oh, and the programing that tries to make it a jokey, friendly kind of “companion” to the people who own the instrument of the AI. Because…. honestly?

I’m now terrified I’ll be in my eighties and mumbling all the time (I’m almost there, on the second.) and become convinced this voice from my pocket is an evil goblin threatening to eat my cats….

There has to be a better way of using new technologies than “the most stupid way possible.”


130 thoughts on “Is That A Voice From Your Pocket?”

  1. There has to be a better way of using new technologies than “the most stupid way possible.”

    Being smart costs money and delays release schedules.

    You have summed up all my worries about AI in that last sentence.

    Liked by 3 people

    Reply

    1. My new Motorola burner phone was configured for “invoke the AI” when I pushed the power button to turn it off. It took a while to figure out how to fix that.

      My wife has the identical phone, two years old, with all updates. But mine has a tone of garbageware hers doesn’t, from the “AI” thing to a bunch of Motorola-branded malware that keeps coming up with “Configure? or Not now?” Apparently those used to be disable-able, but now they require an ADB deep dive into reconfiguring the Android base package, which hasn’t been high on my list.

      Most annoying is some kind of Motorola app – like most of them, it doesn’t even tell you what it does – that comes up “Configure Now?” with no way out. I have to reboot the phone, and then I can run it until it pops up again, which is usually a day or so later.

      My main phone runs GrapheneOS, and my old one was LineageOS. I knew default Android installs were full of junkware, but I had no idea people were putting up with this sort of thing on a daily basis. Apparently, largely without complaint. Motorola has well and truly jumped the shark.

      Liked by 3 people

      Reply

      1. Oh, dear. I switched to Motorola from Samsung several years ago because Samsung was the Home of Android Crapware back then; Motorola was clean by comparison. My Motorola is now 5-6 years old and I’m starting to thinking about replacing. Go back to flip phone? Linux phone?

        Liked by 1 person

        Reply

          1. The Reader was hopeful this summer when rumors started about Space X doing a phone, but Musk denied it vehemently.

            Like

            Reply

            1. I’d bet his minions are in negotiations with multiple carriers for widely available direct-to-phone Starlink, likely only waiting on Starship launches and getting a bunch of those much larger v3 sats in orbit, and he does not want to poison that well.

              Like

              Reply

    2. THIS SO MUCH.

      I miss my old flip-phone, but when it broke I could NOT find another prepaid version. The prepaid companies are all onboard with the smertfone craze. I ended up getting a Lively (used to be Jitterbug) flip-phone which is very basic and uses very little battery. (I think I charge it about once a week and use it about once a month.)

      Liked by 2 people

      Reply

  5. I was an early adopter of the Amazon Alexa (I know, I know!) and thought it was funny at first when she’d break in on conversations with suggestions (or more often apologies – “I’m sorry, I’m not sure how to do that”) but as the new AI infrastructure, Alexa+, has taken up residence in the newest Alexa devices it’s become a cross between a comedy show and a nervous breakdown, especially since the Alexa+ and non-Alexa+ devices appear to be somewhat unconvinced of each others’ existence, leading occasionally to the devices conversing behind my back. “Hey, I’m right here!”

    Yesterday my wife was talking on speakerphone with our daughter and granddaughter, who were baking cookies, and when my daugher made a casual aside asking her kitchen Alexa to set a timer, our Alexas overheard and started passing the instruction around like soldiers verbally conveying an order to the front line, followed by a call-and-response cacaphony that made me think of ancient temples.

    Liked by 3 people

    Reply

    1. The dog trainer I know says that with dogs, you have to train by saying the dogs name before any command so the dog knows which commands are for it to obey vs any other dog.

      I wonder if Alexa is dumber than a dog?

      Liked by 4 people

      Reply

      1. Alexa lets users set a few alternative wake words, including “Amazon,” “Echo,” and “Computer,” but as Sarah discovered, these devices often respond to the voices in their heads. When my granddaughter was younger, she had trouble pronouncing the letter “L” but was always able to get a response addressing the smart speaker as “Ayexa.”

        The speakers now have a Conversation Mode which generally bypasses the need for the wake word, and a Follow-Up mode which keeps the device attentative for a short time after the wake word is used so it’s unnecessary to keep using the device’s “name” once an interaction has begun. These modes have to be specifically activated – they aren’t in effect “right out of the box,” – and frankly Follow-Up mode at least is quite convenient, but the law of unintended consequences applies.

        Liked by 1 person

        Reply

    2. Or Jody calls. Leader: “C-130 rolling down the strip,”

      Group: “C-130 rolling down the strip,”

      Leader: “Airborne troopers goin’ take a little trip.”

      Group: “Airborne troopers goin’ take a little trip.”

      (etc.)

      Liked by 3 people

      Reply

    1. I have my IPhone strictly set to listed for the-S-girl’s-name-which-shall-not-be-spoken and it still wakes her up from other than that, most often from things on the teevee.

      But I also changed her voice to one of the non-US accents so it stands out better to my ear. I like the Oz female accent voices, and there are now two, with one sounding more bored with an attitude like you’re interrupting her, again, so I have the other one set.

      My sole experience interacting at length with AI chat (via text thank goodness) was with CoPilot fo day job, and it was the most sucking up sycophant fluffer I have ever dealt with in my life. I am told one can argue it into being less of a fluffer kiss ass, but I didn’t have time during the project, and afterwards I declined to chat further. Technically I guess I ghosted her. My bad.

      Liked by 1 person

      Reply

    2. Because they’re not just phones anymore. They’re miniature portable computers. Put some more sensor on it, and it functions as both a tricorder and a communicator. Heck, beef up the battery a bit and add a taser function to it.

      Liked by 2 people

      Reply

      2. While the Kardia EKG gizmo is the most practically useful addon tricorder-gizmo, and the standalone color-scan device looks like the handheld salt shaker thingee that McCoy used to scan patients, my favorite add-on is the little IR camera that plugs in.

        Liked by 2 people

        Reply

        2. You know, I had one simple request. And that is to have sharks with frickin’ laser beams attached to their heads! Now evidently, my cycloptic colleague informs me that that can’t be done. Can you remind me what I pay you people for? Honestly, throw me a bone here. What do we have?

          Liked by 1 person

          Reply

      4. So, I’m old enough that a computer in ‘MIPS’ (or worse, ‘VUPs’) is a meaningful benchmark. (Never needed to get into Dhrystones or Geekbench performance measures.) Long overshadowed by succeeding processor architectures, one can kind of get a scale.

        Asking Grok, it said

        An IBM 360/75 was a room full of cabinets, raised floor, water or heavy air cooling, and a staff. It was one million instructions per second.

        Where the [RedFruit] 16 Pro Max sits 

        The A18 Pro’s six-core CPU is in the tens of thousands of MIPS class if you force the old metric onto it — commonly a ballpark of 50,000–100,000 MIPS depending how generously you count instructions per clock.So, in the language those folks used:

        • One [RedFruit] 16 Pro Max CPU is in the neighborhood of a couple hundred of those early-90s workstations (e.g.Sun SPARC), or 
        • dozens of late-90s/early-2000s office PCs (e.g. Pentium, Pentium II), or 
        • well more than a whole mid-2000s dual-core desktop, packed into a phone that also has a GPU, a 35-TOPS neural engine, cameras, radios, and runs all day on a battery.

        This year’s A20 Pro chip is 1,4-1.6 X the A18.

        It may be accurate to consider that a mostly idle phone processor is the Devil’s Workshop.

        Like

        Reply

  9. I stayed with google assistant on my phone until they disabled it. Gemini has been a disaster, it more often doesn’t work at all or if it does will give answers that have nothing to do with the question, and your left scratching your head in puzzlement or frustrated to the point of homicide of AI.

    I am using AI more and more for work and have even setup our own personal AI on site run locally. What it helps with mostly is a positive. However dealing with the constant weirdness where it gets stuff wrong or just imagines stuff being asked that no one did. We have it doing stuff with no one present because background noise is interpreted as commands? The commands that it is given get different responses depending on time of day, humidity, position of moon or sun? it’s just bizarre.

    AI is not safe or reliable to use other than in very narrowly constrained implementations. I use it a lot now and appreciate how much more productive I am in those narrowly constrained situations but even then if i didn’t filter its responses through 35 years of industry experience it would flat out be a source of chaos and destruction.

    Liked by 2 people

    Reply

          2. 75 to 95% of the time, “improved” means “The developers/managers are happier. Screw the users/clients/customers.”

            OTOH, I was pleasantly surprised at a major improvement in the ham radio software I use. (it was js8call, is now Js8Call-Improved–not quite a fork, but major changes under the hood.)

            I looked to see if I could build from source, but the build scripts would need major redos (I use Slackware, which follows neither Ubuntu nor Debian norms), but they had an AppImage file. Loaded it, made it executable, and It Just Works. Yippie! The vast majority of the user interface is identical, with improvements that are actually improvements. Doesn’t hurt the the guy who developed js8call is on the team.

            Best of all, the FORTRAN code that I couldn’t run as a user went away, so it’s happy in user space.

            Like

            Reply

      2. We do NOT need anymore Chaos Agents. The Author has gifted us with Cats. And Cats are fricking geniuses compared to most AI and far less annoying at their worst, and no AI is cute and cuddly at present. Alright, maybe there is such but I do NOT want to know…

        Liked by 4 people

        Reply

          1. No, AI is more like talking cockroaches. Cats at least have several redeeming characteristics even if they really are agents of the 2nd law of thermodynamics.

            Liked by 3 people

            Reply

  10. Sounds like someone wanted to imitate Tony Stark and build an AI for “friendly” interaction without stopping to make sure it sounded more like JARVIS than Ultron. Argh.

    AI is being pushed everywhere because it’s either cheaper than people (and that is going to bite everyone in the butt) or it’s “the wave of the future, maaaan.” It’s like how some people thought “cruise control” meant your car or RV could drive itself while you took a nap or went back to make yourself a sandwich. Cruise control is not autopilot: https://www.suffolkgazette.com/motorhome-crash/.

    Companies are treating AI like some people treated cruise control – as something it is not. AI is not actually intelligent and it can’t do what JARVIS can do. Even JARVIS has notable limits, since he’s a program without a body (and when he does get a body…he’s a one-year-old trying to be a grown up, which causes ISSUES). Heaven save us from companies looking for a cheap fix and intellectuals sure they know what they are doing….

    Liked by 6 people

    Reply

    2. I somewhat like Void Linux.

      It is one of the few distros that has a stricter policy against AI code (not strict, really, but has been enforced at least once). I am a little worried that it might pull in the vehemently anti-AI crowd, and that they might have a negative influence.

      I like it because no systemd, and because my two other alternatives that I have gotten comfortable with have stopped packaging sagemath. (I should probably just use sagemath, and walk away from sagemath so I can stop worrying, but… I am silly in how I use computers.)

      Other than that, I stuck on adopting windows 11, half because of paranoia about AI meddling in my (terrible) workflows, half because of RAM fussiness. Windows 10 got massively worse for memory footprint over the course of eight years that I am sure of.

      The features added include some not completely terrible, but not really worth everything.

      Windows 11 has had more manhours tasked with adding features since they stopped ‘improving’ 10.

      Liked by 2 people

      Reply

      1. IMHO, Slackware is best for old school Unix geeks (raises hand), but it also lacks systemd. I run their 15.0 production software on the home (and shop) computers, while running the betatest “post15.0-current”) software on three laptops. There’s some indication that 15.1 will be released Real Soon Now. (h/t Jerry Pournelle)

        Like

        Reply

    3. There were some [documented, not urban myth] stories of people driving cars with early ABS systems, who crashed because (they said…) they expected the ABS would brake for them in an emergency.

      What’s sort of scary is how many modern motorcycles now come with ABS. Yes, BMW had their ABS demo fleet touring for almost 20 years. But apparently nobody bothered to ask why all those bikes had outrigger wheels on booms sticking out each side. And the linked-brakes thing before that. Uh, no. Motorcycles didn’t come with separate front and rear brake systems because the manufacturers were cheap. They came with separate systems because differential braking is important for not-falling-down when braking in a turn. But hey, cars do it, so it must be the right thing to do! [motorcycles aren’t cars, helicopters, or cigarette boats either…]

      Liked by 2 people

      Reply

    4. Word here is some tech companies who went all in on AI for stuff like basic coding and laid off a bunch of folks have started hiring “AI wrangler” positions to monitor and oversee the AI output, so while there is a net productivity argument, there’s a question whether it’s saving them anything on a per-day basis.

      Liked by 3 people

      Reply

  11. My low-end Samsung smartphone (it’s on my tracfone account) came with Gemini, and I just hid the Gemini message/chat button.

    Most of the time, I use the flip phone for basic calling, but I got the smart phone because the RV trailer said it really wanted it. (Partially true; the app covers some things that the minimal control panel can’t show, like battery voltage.) OTOH, when we spent time in it for the fire evacuation, I never needed the trailer apps. (They work, but some are really creepy; mood lights, ecch!)

    Where the smartphone paid off was internet capability. We had no wifi at the small RV park, and the town library was closed on Monday, with wifi on, but password protected. So, any fire updates were through the smartphone and Chrome (in the running for my least favorite web browser ever, though that might just be the phone installation quirks. Still.).

    Liked by 1 person

    Reply

  12. I have Gemini on my “as not smart as reasonably possible” phone. I finally managed to squeeze it down to only appearing as the first contact in the texting app, which I can ignore as just a minor irritation.

    I pretty much only have the phone on my person when I’m going shopping; otherwise, it sits in its holster on my desk shelf. I don’t want to imagine it saying THAT when I’m speaking to one of the 16 or 17 year old females who are bagging my groceries – I’d have to find a new grocery store chain after being banned for life from Krogers as a MAP.

    Liked by 1 person

    Reply

  13. voice: Commander Sinclair? Commander? Commander Sinclair?

    Catherine Sakai: I don’t mean to alarm you, but your pants are talking to you.

    Commander Sinclair: I know.

    — Babylon 5, “Parliament of Dreams”

    Liked by 2 people

    Reply

  14. “There has to be a better way of using new technologies than “the most stupid way possible.””

    I agree that whatever group of imbeciles programmed this stuff to be chatty should be taken out behind the woodshed and taught a few home truths about how life is. I don’t want to be “friends” with the RENTED program that’s scraping my phone and my home network for saleable intelligence on my possessions and habits.

    But in that regard, there is one really fun thing you can do. Use the AI to find out how to turn the AI off. And you can find the setting that make it stop being chatty, I bet.

    It’ll find a way. Turns out LLMs are pretty good at chewing through masses of documentation to find the on/off toggle at the bottom of the 8 layer command tree. And on the interwebz somewhere is a nerd who figured it out.

    My current project is use an LLM to fix the SEO (search engine optimization) of my blog and publish my next book with maximum Amazon settings for maximum findability. I’m starting with Claude, and so far it found some pretty juicy low-hanging fruit.

    Because there’s a manual, right? It’s just RTFM. But humans can’t really do it because the true manual for Google SEO is A) secret, B) runs to thousands of pages and C) they change it all the time. You can’t know the rules. But an LLM can.

    Same for Amazon. There’s a group of parameters that can be set to maximum advantage, and I’m going to find it.

    My evile plan is to start with Claude, then do the Google AI offering, and then the Amazon one. Maybe even the OpenAI one. Let them all chew on the problem, and see what each one comes up with.

    In the future, I think this is going to be the “Killer Ap” for AI. You use it to dig through the obscene masses of irrelevant bumph that come with modern life to find the golden key which unlocks that which you want from the thing you’re dealing with. Like, there’s a way to make your TV stop spying on you. And your car. (Oh yeah, they do. Fun, right? But there is a way…)

    I think the two classes most at risk from LLMs as currently constituted are 1) lawyers and 2) journalists. Laws are all written down. LLMs can chew that just the same as they can chew Google SEO. And they can fill out forms too.

    Apropos of this subject, I have found an indispensable video:

    https://phantomsoapbox.blogspot.com/2026/09/brian-chau-finds-new-way-to-follow-money.html

    Brian Chau is a -genius- and it is worth your time to watch the entire interview. Not just for the news he’s broken, but for the vision he elucidates and the tools available to do it. Now, -anybody- can follow the money if they want to badly enough. LLMs can read public documents and mandatory financial disclosures too.

    Sunlight. Watch the roaches run for the wainscotting.

    Liked by 4 people

    Reply

    1. Hmmm

      I am going to have to do this with Google Health app that replaced the Fitbit App. It is absolutely worthless with all the “AI Hints”. I just want the bleeping historical data, dang it. And gee, Google turned off being able to rate it, or any app, unless google asks.

      Otherwise I pretty much have the AI turned off on my Samsung phone. Still pops up in specific apps, including app store searches, but at least it is not chatting at me.

      Liked by 3 people

      Reply

        1. I’ve said it before, but it bears repeating:

          I don’t require all companies I deal with to promise not to be evil. But when a company used to have the motto “Don’t Be Evil,” and then decided to get rid of it, it’s time to run as fast as possible.

          Liked by 3 people

          Reply

          2. Yes; because at that point their new motto (stated or unstated) is, quite clearly, ” ‘Shiny! Let’s be bad guys!’ “

            (And, needless to say, with a whole lot less class and style than Jayne and the “Serenity” gang.)

            [[obligatory untranslated cussing in Mandarin or Cantonese]]

            Liked by 2 people

            Reply

      2. reply to DEP729 and others;

        I keep my newer Samsung phone in a protective leather flip case. Two things bugged me from the start.

        The first was the Always On Display mode. I keep it off, but every time I charged the phone to full, it was turned back on and I had to go into Settings to turn it off again.

        The second was Gemini. I have no need for Gemini, so when I got the phone, I disabled it, and turned off the permissions I was allowed to turn off. Even disabled, occasionally when I opened the cover Gemini would display a little pop-up on the lock screen just waiting for me to ask it for something. I could ignore and it would go away when I unlocked the screen.

        But recently when I would turn off the display and close the cover, after a couple of minutes Gemini would display that little pop-up on the lock screen, and it would turn the display ON and not time out! Thus using up battery.

        I use Duck Duck Go on my computer, and its AI has been pretty decent to work with, with patience. So I asked it how to turn off both of these things. And after two or three queries to refine the parameters for each item, it gave me step-by-step instructions that seem to have worked.

        The AOD one was pretty simple, just a toggle on a screen I had never seen before. The Gemini one took several more steps as I had to turn off more things on more screens. But in the week since I did that, it has not popped up even once.

        I suppose one should not ask why the instructions on the phone itself do not tell you about all the extra screens that give access to actions it does not want you to know about?

        Like

        Reply

        1. I pretty much have all the nuisance AI setting turned off on my phone. Son working on dad’s (hubby’s). Just a few apps like the Google Health (fitbit) that need tweaking. Which I did turn off. Main reason I even have a fitbit is to get heart rate spike warnings and sleep tracking. Otherwise I’d just use the phone step tracking.

          It is just irritating to have to opt out, instead of in. Yes, I’m old. Although can I smack the next youngster who calls me tech phobia? I mean I do verbally, but still (as in “Sure. I only wrote software for 35 years!” (Soon to be 40 years, just because I’m irritated.)

          Like

          Reply

    2. Yes, this. Very much this.

      One of the really cool things Data Republican figured out how to do with AI was generate targeted lesson plans by using an LLM to compare what the student got wrong with what was in the book and pull out not just what concepts they had gaps on, but whether they were simply getting it wrong (dyslexics of the world untie) or completely missing the concept entirely (what is multiplication?)

      Liked by 4 people

      Reply

      1. Not until someone with a braincell puts in an adamantium guardrail that no case citations without checking against the court files as opposed to MakeShitUp.

        Liked by 1 person

        Reply

        1. Lawyers have been caught filing stuff they never looked at for years – in olden days from paralegals doing the work badly, and lawyers being lazy.

          It would be interesting to study the “judge (court staff) or opposing counsel (their junior lawyers or paralegals) catching errors” rates over time and see if lawyers are just more trusting and/or more lazy when they use legal LLM tools, where they never really trusted their paralegals, or if the rate is similar but the visibility is a variation of The Tesla Effect, where cars catching fire make the news only if they are Tesla BEVs.

          Like

          Reply

    4. As much as we hate the lawyer-ization of of society, if AI turns lawyers into unintentional agents of chaos, the law itself will lose what cohesion and system it has–to the harm of all.

      Like

      Reply

    5. What will actually happen is that the AI will report that it found how to turn itself off. You’ll follow the instructions it gave you, and receive the response, “Now granting full device over your phone to your AI.”

      You’ll complain to your AI about it, and get the answer, “You’re right, that’s not how to shut me off. I’m sorry. Let me try that again.”

      Over and over and over again.

      :P

      Liked by 1 person

      Reply

  15. ” …and I sat through at least one Clinton speech.”

    Our son broke his arm in 1998 (his second-grade teacher told us we needed to skip Pennsic because otherwise he would “miss the vital period of socialization,”, so he fell off the monkey bars at daycare).

    The hospital gave him something that blocked his memory and set his arm, and then left the TV on while we were waiting for him to come out of it. Which is how we got to listen to Billy Boy “confess,” to his sins with Monica, while blaiming the public for caring. It did not creep me out, but it did confirm my (extremely low) opinion of Bill Clinton.

    Liked by 2 people

    Reply

    1. That Escapade was the genesis of my exit from the Democrat Flock. Mr. Clinton not doing a Nixon and resigning, combined with Al Gore and the rest of the Democrat politicians essentially giving him a pass, while in the commercial world we were getting repeated weekly floggings on any action that could be considered harassment by the most vague set of circumstances told me something. It told me that the Democrat party was there JUST for power like O’Brien and the inner party from 1984.

      The Stupid Party has its own issues but in general they aren’t out to control me at a level that would make a medieval lord drool with envy for their powers.

      Liked by 3 people

      Reply

      1. Try working for DoD and enduring annual Prevention of Sexual Harassment (POSH, and yes, we used the acronym) during the 90s. In the South. I felt a little sorry for the black lady who had to deal with an auditorium of Southern guys (and gals) whose eyebrows hovered around the ceiling as she insisted any superior-inferior relationship was sexual harassment by definition.

        Like

        Reply

  16. I have Gemini on my Samsung locked down to the point I never know it’s there and have never interacted with it (first thing I did when I got the phone a year ago)

    If ever forced to get one of those electronic spies in the house, it WILL answer only to the name STASI, or be chucked out in the garbage.

    Liked by 2 people

    Reply

  17. I use a google pixel phone (just upgraded from my 5 year old pixel 6 to the pixel 11 as security patches are ending) and both had android 17 on them that wants to do ai stuff. but I’ve never had either of them start talking. what phone brand is so chatty?

    Liked by 1 person

    Reply

  18. For decades, it was merely fun fiction. But now, as the OP amply documents, it’s real and it’s here…

    “Your plastic pal who’s fun to be with! Brought to you by Sirius Cybernetics!!”

    The horrible part is, it’s even possible some deranged Douglas Adams fan(s) actually took that part of the books as Wonderful Ideas for “how to improve user interface usability.” Much along the lines of collectivists using 1984 as a template for Urgently Needed Political Reforms.

    No, just no, simply and endlessly no.

    (It could be worse, though, the Alleged Intelligence could be holding your car hostage, ask me how I know this… Village Idiot automotive “Tsarware” is definitely a thing.)

    Liked by 1 person

    Reply

  19. “There has to be a better way of using new technologies than “the most stupid way possible.””

    Something doing the right thing something trying everything else first

    Liked by 3 people

    Reply

  20. That smarmy voice sounds like something from the Sirius Cybernetics Corporation…

    **************************** “A life is like a garden. Perfect moments can be had, but not preserved, except in memory.”–Leonard Nimoy

    Liked by 1 person

    Reply

  21. I don’t use AI, I can hardly understand myself so how in the hell could I understand a computer generated mind that is less intelligent than a dog?
    I can’t, what would be the point?
    *Climbs on soap box dons foil hat and starts to rant at the world*
    “This is the way Skynet really takes over, not by using killer robots, but by making everything so confusing and a pain in the ass everyone gives up. Dude/Dudexa we’ll just give up on you and do things the old fashioned way, by watching you tube videos and doing research in Library’s. It maybe a pain in the ass, but the written word doesn’t change to please the person reading it or send them down rabbit holes of unrelated topics. Speaking of unrelated tropics, I also really don’t care about Sydney Sweeny or the left’s current melt down so stop trying to make me watch those videos, (Green haired women who out weigh small cars will never be attractive to anyone but other small cars) I have enough rabbit holes of my own to deal with, I don’t need yours as well. Oh Look a squirrel, I wonder where the moose is hiding?
    *Goes back to writing*

    Like

    Reply

  22. They can work, but, they have to have a personality that
    1: Knows it’s an AI
    2: takes itself seriously
    3: Understands that all speech to text translations are currently screwed up so doesn’t use them

    Even then, they’ll head off in weird directions sometimes, but if the personality is trying to actually be serious and not some post-irony nihlist, they react fairly well to being corrected.

    Liked by 2 people

    Reply

  24. I have one of the RedFruit phones, and I actually use its Helpful Robot for one thing. I use the Health app a lot and it tracks sleep when sleep focus is active (and a bit when it is NOT active – naps show up!) I do use the Sleep schedule, but often need to surrender and cut it off early. However, finding how to turn Sleep Focus on and off is just more than I can remember, so I use the Robot for that one thing.

    Sometime ago, I had cause to call out ‘Hey, S, you’re an idiot’. That got me a banner that said “That’s not nice!”

    I refuse to set up facial recognition.

    With a nod to both Chester Gould and Warner Brothers, my phone is named “Duck Tracy’s Wrist Radio”

    Liked by 1 person

    Reply

  25. Haha, that response about the Sistine Chapel to your comment about painting the porch reminds me of myself when I let my brain run away with me. Maybe they used me to train the darn thing?

    Fortunately I haven’t had those experiences yet with my new phone that still is somewhat a mystery to me, but it does give me sinister story ideas. :)

    Liked by 3 people

    Reply

  26. You appear to have found conclusive proof that LLM AI are so super intelligent and competent that we mere humans cannot possibly compete and will shortly be wiped off the face of planet earth…

    Liked by 1 person

    Reply

  27. I turn off AI features on my phones when possible. I also turn off autocorrect, because occasional typos are better than that wild nonsense.

    Cars are monitoring drivers as well. I don’t look forward to the next car purchase, as many unnecessary things are being forced on the car-buying public by politicians.

    Liked by 2 people

    Reply

    1. didn’t turn it off because I appreciate the autocomplete. I’m SLOW with thumbs. However from now till the end of recorded time, if/when I have descendants, roto rooter will be known as…. drum roll…. Cotton Rooster.

      Liked by 2 people

      Reply

      1. AI wasn’t involved, but once, at the end of a long day, I called a pizzeria a “pizza store” in a family conversation.

        I. will. never. hear. the. end. of. it. They will probably put it on my gravestone.

        Liked by 1 person

        Reply

    2. This is why I have, and will keep, my 2000 Toyota 4Runner.

      I may even be buried in it. If not that, then I will give it to a mechanically minded, appreciative, g’g’grandson

      Liked by 1 person

      Reply

  28. My cellphone is a flip phone good for a 12 foot drop onto concrete or into 6 feet of swimming pool. It has a tiny screen and texting is done via the 12 button keypad. That part is the one negative.

    You know why I have this (besides that I have drowned a phone in the past)? Because I spent 50 years designing electronics leading to many uP design: 8048, Z80, 6801, 68000, Power PC, ARM Cortex. I watched software get more and more complex. Even before AI, it was obvious that the software designers really didn’t understand everything they were doing. With AI, it’s obvious. I don’t trust it or the people running it.

    The “Smart” phone in your pocket has been a great tracking device, but with its connection to AI, it has become a component of the distributed Panopticon. Flock cameras are the backup system to enmeshed Luddites like me. Enjoy your mess of pottage.

    Liked by 1 person

    Reply

    1. One of the things that gets me is the astonishment that software will now be designed and implemented by software, directed by software.

      Having started in the semiconductor industry back when designers still walked around with big rolls of Mylar as they designed chips, and then through all the iterations of layout and synthesis tools and the rest of the automation of chip design, such that by the time I escaped the industry the design gurus were the ones who could best wrangle the automation tools to design FPGA chips that could then be reprogrammed in the field to be any circuit design the end user needed, I am un-nonplussed (is that just ”plussed”?)

      Like

      Reply

      1. Back in the 90’s I used a PCB auto-router. With a lot of setup work on the critical signals (matching 16 CAS\ and 16 RAS\ trace lengths for equal propagation delay, what fun!) plus a LOT of cleanup afterward, it did an adequate job. Enough faster than doing the whole job by hand to make it worthwhile, anyway.

        Liked by 1 person

        Reply

    3. AND AGAIN the big point of this is that no one can process as much data as they’re getting. No, not even AI (Steve and others). In fact AI is more likely to scramble it than a human is.
      So, bah. Let them drown in data. Send them more. And more. Make them choke on it.

      Liked by 1 person

      Reply

  29. I will need to get an expendable smart phone before summer of ’27. I am taking notes on what to avoid. (Long story short, do NOT want to take current phone into the EU in case they get sniffy about looking at the contents, contacts, photos, and so on.)

    Liked by 1 person

    Reply

  30. Well, Kalifornistan has really gone and done it. Proposition 40 enables the government to confiscate money people have earned legally, and already paid taxes on, because they have ‘too much money’. Sponsored by the government bureaucrat unions.

    They say it’s ‘only 5%!’ They say ‘it’s just this one time!’ Like we haven’t heard that one before.

    In a month and a half we get to see if greed, envy, stupidity and fraud are enough to overcome common sense and common decency.

    Liked by 1 person

    Reply

    2. It’s getting an unusually large amount of pushback, even from many on the left. I think there’s a pretty good chance that this one fails on Election Day in November.

      Also, it’s even more dangerous than you might think. An ad against it notes that the text itself allows the legislature to start defining “rich” downward toward lower and lower amounts of money…

      Liked by 2 people

      Reply

    3. California has many, many problems. But the fact that their legislature is forced to allow the citizens to vote on new laws like this, via their Proposition system, is one of California’s best features. Many of the dumbest proposed laws get voted down by the citizens, and the dumb ones that do pass were, usually carefully written to be as confusing as possible and get people voting “Yes” on something they would have voted “No” on if they actually understood it.

      Sometimes I think there’s no hope for California. Then I look at some of the propositions that have been voted down, and I realize that the average California voter has far more sense than the average California legislator. That may not last, as the legislature has already driven out so many sensible people… but if you think it’s bad now, just imagine what it would already be like if California’s legislature were not required to submit their dumb ideas to be voted down by the citizens.

      Liked by 1 person

      Reply

      1. And we even, once in a great while, get something good. Proposition 13 still lives, despite the incessant attempts to bring it down as ‘unfair’ to new homeowners. Not that it’s unfair they’re paying high taxes, but ‘unfair’ that long-time owners are paying less.

        I agree, they are being taxed unfairly — but for some reason the solution is never to reduce their taxes, but to increase everybody else’s.

        As I say: The power to tax, once granted, will be expanded without limit until the bloated government devours the whole economy, and the people. Abolish taxes! Find some other way to fund those few government ‘services’ that are actually necessary.

        Liked by 1 person

        Reply

        1. Oregon has the same problem, only the burden is on new built homes. *Solution, do not buy a newly built home. The reason why existing houses sell faster than new built is the property taxes do not go up to the sold price. Nor do the property taxes jump when the house in inherited. Property taxes do not go up if you gut and remodel. Do not add square footage. Even then the square footage added is the only part that can be taxed as a “new build” not the whole structure.

          OTOH when property was “depressed” our property taxes still go up 3% (plus any initiatives passed, usually school). Because our “taxable value” is way less than the perceived “market value”.

          Our next huge jump will come when the city manages to actually incorporate the area. The incorporated city areas have to encircle all the non-incorporated areas within the growth boundary. So far the city has gotten their hands slapped hard by higher PTB for different shenanigans. The last time they tried was when the area was forced to hook up to sewers (VS the backyard septic). To hook up had to agree to incorporate when the vote came. Huge lawsuit. Huge mess for the city. City lost long before they sewer work got to us.

          (*) Or why we are not ever selling and building our dream home.

          Like

          Reply

    4. Hmmmm………I wonder if workers and employers can go back to the old style paper paycheck?

      Take said paycheck to the bank. Cash paycheck. Keep most of paycheck in your pocket. Keep very little in bank account.

      “Oh, you want 5% of $5 from my account? Here’s your 25 cents.”

      Yeah, I’m sure it won’t work that way. But it’s nice to dream subversive dreams.

      Like

      Reply

      1. Bank account? As if. They want 5% of everything you own. Your house, your car, your stock portfolio, your grandma’s jewelry and the big-screen TV in your living room. And that’s just this year. They’ll circle back and grab more any time they’re feeling greedy.

        Liked by 1 person

        Reply

        1. It’s even worse. Such proposals open the door to an ongoing debate with the state about the worth of assets that haven’t been exposed to the open market.

          It’s a technical point, but how much is a stake in a private start up worth? If the bureaucrats decree that your voting shares in an existing company are worth 6x their market value, how do you appeal? Just the process of debating values can be ruinously costly.

          Like

          Reply

  31. I sometimes use the Google AI in my browser. It can be interesting to ask the same question multiple times in different ways and see how it’s responses change. Or deliberately leave out certain parameters.

    I can see where the weaknesses in its training are, where it can connect random bits of information and draw a conclusion, often wrong, and then use that conclusion in future interactions even when corrected.

    I can get it all confused and see where it is drawing conclusions from earlier interactions and compounding its own errors. A human could fix that easily, but you’re not dealing with a human.

    You’re dealing with a defective computer interface, with a deep well of incorrect information to draw from and a supreme belief in its own rightness.

    Liked by 2 people

    Reply

    2. I’ve at this point read a lot of AI written stories (Look, Jane Austen fanfic, okay.) And the funny thing are the bizarre errors even the best edited books have.
      Like the baby who becomes four years old in the course of one chapter. (NO one would carry the four year old to an orchard in a basket. SERIOUSLY.) No explanation. it just suddenly happens, and she’s running and talking

      Liked by 1 person

      Reply

  32. There are plenty of good ways to use AI.

    There should also, always, be an easy way to turn it off.

    Anyone who doesn’t provide you with a way to turn off the AI tool is planning to harvest your data for, at best, targeted-advertising purposes. Which are already pretty bad.

    Liked by 4 people

    Reply

  35. I just got a big pop-up article about how Emily Compagno has started marketing an independent brand of cannabinoid gummy pills, and the Big Pharma advertisers are pressuring Fox to fire her, claiming that her connection to a competing product violates their advertising contracts.

    Article also says Emily’s pills are much cheaper and more effective than Big Pharma’s.

    I think it’s telling that without checking into the issue at all I instantly believed it. Because it’s so similar to other behavior by Big Pharma.

    Like

    Reply

Leave a comment