Societies don’t cope well with technological innovation that changes the way people live every day. Because once you change the way people live every day, you’re changing the way people interact, which means the structures of society are going to start not working the way they were intended to.
Right now we’re in a time of catastrophic innovation. What does that mean? It means that we’ve reached a point where innovation is building on innovation — which is to be fair what innovation does. First someone figures out how to dig in the ground and plant some seeds and next thing you know someone is planting kale, and then you have to kill them and… ahem… — and the new one is dropping before we’ve even started to process the effects of the last one.
Honestly, the Heinlein time line (not that one. I was told I can no longer mention that particular story here, not without doing the whole Black Adder dance for when someone mentions the Scottish Play) would have been easier on our institutions as they were when he wrote the juveniles. Going to space would just be a continuation of the westward expansion, and at least on earth there was no big innovation on communication and putting the power of communication in the hands of individual to counter the centralized big media. (Or the power of manufacturing, to counter centralized big industry. Okay, we’re on the early stages of that, but it could easily be the next big break through.) In fact, he had Earth become more and more the worst aspects of FDR, to be honest.
But no. We weren’t allowed to go to space (no, you can’t possibly convince me it wasn’t deliberate. The way the left hates any attempt at space exploration and becomes insane with their FEFer (Fix Earth First) bullshit, you know they did what they could to prevent us leaving this rock.) And so some geeks in garages created the internet revolution. Yes, yes, I now. Government, but be real. If it hadn’t escaped into private lands and cat memes, it would still be unimportant to all but defense geeks.
And that… well, that escaped containment again, at least in the US right after 9/11, when a lot of us trying to figure out what was happening and what should happen hit the internet and a lot of the big blogs now originated. And then there was social media, and then–
Well. Here we are.
And not only haven’t governments and institutions processed the internet, but they haven’t even started to realize the paradigm (remember how fond they were of the word in the nineties? Well, it actually does mean something) has shifted under their feet.
Their last two normal political campaigns (Biden doesn’t count. Nothing about that was normal) they got all the celebrity endorsements, paid all the networks, and sat back and waited for victory, and are still confused it didn’t work. With Biden they were in the middle of a massive psy-ops campaign with the help of China (as we now know, and honestly always suspected) that included memorex riots, memorex race war, memorex pandemic, memorex campaigning and very definitely wholly memorex campaigning with a character whose mind was definitely not real or at least not present.
But even with having locked the entire country to tank the economy and run a major psy-ops to distort everything and control all the information, the only way they “won” that election was to cheat massively in front of G-d and everybody.
And then, despite their attempts to follow the good old play book of the Bolshevik revolution (they only have one playbook, really) it didn’t work, and people pushed back on things like their attempt to establish a ministry of truth (how very dare they.)
Now of course they’ve decided what they really need is more foreign communists and crazy people in charge. And yes, of course they’re counting on fraud.
What they’re not counting on, though, is that everyone has a movie camera in their pocket to film any shenanigans, or that everything they try to cover up comes out or–
Honestly, as flat footed and wrong footed as they are, as snake bit and struggling, it makes me wonder how much they were hiding from us in the days of centralized media. Some of it of course I knew, like the immediate tarnishing of the Reagan economy when Clinton took over, or all the lies about how things had worked in the eighties or… But– really?
How much were they hiding from us in the past?
It’s hitting mostly the left, because the left had control of the old establishment story-dissiminators, be they “news” or fiction or songs or movies or… well, any of it. They had the old system completely under their control, and knew how it worked and what they could do. Which means that they partly refuse to believe it’s not working, and partly have absolutely no clue what works. The right — which let’s face it, at this point is everyone to the right of Lenin — on the other hand had found themselves slowly pushed out of every form of mass media until they were out in the wilderness. In the wilderness, you adapt faster to innovations, because it’s the only chance you have. We were, in fact ready.
Would we have been without Rush Limbaugh, who in fact used an old form of communication to get around the lefty-lock on everything? I don’t know. In a way he primed us to hit the internet running. We both knew there were distortions in the mainstream narrative and knew we were not alone. In fact, maybe that’s the difference between the right’s adoption of the internet in the US as opposed to the rest of the world. Perhaps Rush Limbaugh was the thumb on the scales we needed.
It would have got here, mind you, but maybe in another twenty years or so.
In fact, you’re seeing, abroad, where we were twenty years ago. The governments cracking down on memes are at the beginning or a long road of pain where they lose more and more control of the people. And the invasion that is supposed to replace them? Well, far more advanced there (by percentage) than here, or at least it looks like (None of the numbers are reliable, either side of the Atlantic, people. Yeah, they might be worse here. Or not. Again just by tracking “magazines in Spanish in the checkout stands” the peak was around 2010, because the Obama economy caused the first Gran Salida. And they needed to bribe the next waves with welfare, just to get them back.) But irreversible? I don’t know. I kind of doubt it, in fact. If they hurry up and catch up on the communication front, well, it will get ugly in Europe (Europe specializes in really ugly once they get pissed off enough) but they might yet save themselves. Or put it another way, when governments are trying to ban memes, they’re not winning.
Do I know how this will play out? No. But…
Look, it helps if you remember everything has two sides. Some crazy person on Twitter assured me that because of AI we’re going to be in a robot tyranny or something like. Sure, the potential is there. If we let things get centralized and controlled by bad actors. (Oh, Hello China!) BUT not if we let the free market handle it, all elbows and chaos.
Is AI disruptive? Oh, you bet your beepy. Majorly disruptive. And weirdly some of the professions that seemed like they would be the most immune to it are the ones who are getting hit hardest, like most of the liberal arts ones. And no one knows what it will come for next, so preparing for the future is difficult at best.
BUT it’s also disruptive the other way as well. In the hands of good guys, it might help us figure out answers to things like “what is really happening” and “where is our money going” as well as help small companies and individuals compete with the monopolies that are now almost all controlled by the left.
Used properly it can easily put Hollywood out of business. And it won’t even take long. Used properly, it can help you learn anything you want to learn without giving money to universities that have become pits of foreign influence and communism.
(No, seriously. On a lark I had it outline a course for me to learn Classical Greek and Latin from free materials on the web, and I think it will work, though I don’t intend to try until I’m done with the looming move. I very much think it will work.)
What else can it do? I don’t know. But if you think the establishment that hasn’t even yet coped with the fact they don’t have the only megaphone in the room can cope with the further disruptions about to eat them, you’re out of your mind.
Sure “it will be used for evil” — that’s a given — but not just for evil.
Orwell saw how the tech of his time could be used to lock society down. It could. In many places it was. His error was to think it would last forever.
Behold this one strange trick: we too have ability to create, to disrupt to innovate. (In fact, we have it far more than the left, because we have the need to.)
Let’s get to it, and use everything to build over, build under, build around.
And throw the stars ontop of that, because we’re going there too.
Set backs? of course there will be setbacks. And I guarantee the left will get crazier and more malignant as we continue to break their little (eh) red wagon.
Don’t let that deter you.
In the end, we win, they lose.
Ad Astra.
According To Hoyt 
Ad astra sine brassica (or at least no kale).
LikeLiked by 1 person
Grampa Pete loved kale. Grandma made him cook it in the garage. I’ve never had it, and judging by the vitamin K content, if/when I need to adjust my clotting time, I’d rather have spinach.
LikeLike
Good thought.
LikeLike
Spinach quiche, yes.
Kale quiche? /shudder That would be absolutely diabolical.
LikeLike
For reasons unknown to me, my clotting time is getting high. So, instead of spinach on the banned list, I now have to eat it twice a week. (“Have” as in “Yes! Yippie”)
LikeLiked by 1 person
Ah, one of my people.
LikeLike
Kale and quinoa. Ych.
LikeLike
I’m going to try to get perennial kale established in the chicken areas. I trued before but they ate it right down to the root.
Kale is for chickens, goats, maybe sheep or pigs. Kale is my food’s food.
LikeLike
I love massaged kale salad. https://andreabeaman.com/massaged-kale-salad-with-pepitas-cranberries/
Quinoa salad: https://damndelicious.net/2014/06/07/whole-foods-california-quinoa-salad/
In other words, preparation can make a huge difference in taste. I love fresh spinach in salads, and spinach in soup, but sauteed spinach gives me the creeps. I’m sensitive to texture. I also hate cooked oatmeal.
LikeLike
Living in interesting times for sure. The opening of communication that the internet produced has certainly made a lit of us realize that we aren’t alone or even just a small group. I think that is part of what the leftists (and other would be totalitarians) fear the most, people realizing they are part of a large and growing group. And that is messing up all their plans for a “perfect society” with them in charge. They are pushing back and I wonder how much falsehoods they are spreading to discourage us. Besides the data center hysteria I’ve heard some other really strange crap floating around, mirrors in space, leftist millionaires building bunkers, etc. But I fervently believe the left is losing, even if they do manage to inflict a lot of damage and pain their marxist ideas are fading.
LikeLiked by 2 people
Um, solar mirrors in space are current events. https://www.nytimes.com/2026/07/10/climate/fcc-space-mirror.html
It is a great example of a catastrophic innovation. Astronomers are rightly fit to be tied. In addition, while lighting up solar farms at night might increase their energy production, I don’t think the “reduction in fossil fuel use” will help at all. Adding extra solar energy into the earth’s atmosphere will directly increase global warming.
LikeLiked by 1 person
They very much want to turn AI into Big Brother. Hence the cry to regulate regulate regulate.
LikeLiked by 1 person
Leftroids believe that ‘AI’ will provide them with intellectual slaves to think for them, complementing the traditional slaves laboring for them. It will not go according to plan.
See, if you don’t understand what the clankers are doing, you won’t have a clue what the effects will be. Like a monkey trying to run a factory.
LikeLiked by 3 people
If you don’t understand what the clunkers are doing, you become every spoiled rich bish in period fiction who can’t manage the servants because she doesn’t understand what they do.
LikeLike
Fortunately the world doesn’t run according to their wishcasting.
LikeLike
The Left grows increasingly (even desperately) annoyed at the Masses of Humanity stubbornly refusing to be perfected by their betters. Thus the left seeks to perfect AI, to then turn it to perfecting, finally, those stubborn Masses of Humanity.
LikeLiked by 1 person
A big brother with hallucinations! So, political business as usual.
I saw the UK government is planning to adopt AI in a big way, becoming a center for AI innovation. Biggest laugh I have had for some time.
I am convinced that neither politicians nor bureaucrats understand technology in general; hence the UK government’s (clearly going to be successful) attempt to ban Twitter and any other disapproved of platform. I suspect they think of AI as a sentient thinking machine (i.e HAL) from the 1960s.
I am also convinced that bureaucrats see it as a way of doing less. What’s the betting that AI adoption has no impact on the ever increasing size of bureaucracy? Of course, that’s probably because AI will free bureaucrats to do important strategic thinking.
Really glad to be living in the US.
Now, back to the AI doing my job for me.
LikeLiked by 1 person
They do think of it as sentient. Every man jack of them.
LikeLike
Somebody probably saw The Forbin Project while high and figured it was what AI does.
OTOH, early Mycroft with his attempts at humor seems closer to the mark. I kind of felt sorry for Mannie with page after page of jokes to be evaluated.
LikeLike
Note; while Mike was sentient, I’m trying to draw the similarity to his awful jokes and AI hallucinations.
LikeLike
IIRC Forbin Project was “late” in the “Computers Are Intelligent And Want To Control Mankind” story theme.
Of course, Kirk and company had to deal with several “intelligent computers”. [Wink]
LikeLike
And yet the clock ticks closer to midnight. OpenAI says AI models went rogue during testing, triggering ‘unprecedented’ breach at startup
LikeLike
c4c
LikeLike
Proto-internet is 19th century. -19th-
Telegraph.
Folks could, from almost anywhere with a wire, send a message anywhere else. The wires followed the railroads, and then their own routes. In 1866, it spanned the Atlantic permanently. It grew explosively as information is valuable, but transmission is fairly cheap to achieve, crashing the cost.
Multiplexing signals also developed in the late 19th.
Portable commo existed in special gadgets that a person on a train could use. Stop the train, walk to the adjacent wires, run up a pair of hooks. Tap in. start keying and listening.
Radio overlapped wire, but voice gave wire another great leap, then did it to radio.
The 20th century grew it vastly. But the only really new territory claimed after 1900 was space.
By 1950 the data stream was huge, if kinda clunky. Wire phones, then cellphones. Then the personal computer and suddenly everyone wanted networking. first dial-up, then bandwidth.
But that “immediate to almost anywhere” was spawned fairly early in the 1800s, and was intercontinental by the latter third of that century.
(Note, if you want to get picky, heliograph systems date to the 18th century. But they lack a dedicated receiver/recorder like telegraph. thus they are “semaphore” signals not true telegraphy.)
LikeLiked by 4 people
A lot of people forget the 1st American Civil War was also the first major conflict we fought using (and attacking) long distance electrical communications.
LikeLiked by 3 people
And the ability to transport food overland without entirely relying on means of transport that also needed to eat.
LikeLiked by 1 person
Once upon a time in a library far away I note a comment in a history of WWI that the largest single budget line-item of the US Army was horse fodder.
My dad’s dad drove a chuck-wagon drawn by said fodder-eaters across France.
LikeLiked by 1 person
The final mile was still horse or mule, and sometimes more extensive journeys were needed — but then, you could send in the horse fodder by train, too. It made an enormous difference.
LikeLike
And a few experiments in aerial reconnaissance!
LikeLike
True, and I’ve never seen it so well put. Yes, the government invented ARPANet (thank you Al Gore), but because some of those geeky professors wanted to talk to each other about their geeky hobbies, it escaped into bulletin boards and eventually the web.
As to the great AI takeover, I had an interesting conversation with a fellow geek at Son of Silvercon. He works for IBM and mentioned that the suits have begun to notice that getting results from AI is actually taking much more time than they had budgeted. I’m sure they don’t understand why, but I do. Despite its purported “natural language programming language”, trying to make AI really do what you want it to is only at the equivalent of assembly code level.
So yes, there will be an AI bubble, just like there was the great web bubble. Remember when the venture capitalists would toss $10 million at any startup company with a dream, and that company would spend $1 million on a Superbowl ad without even a product to sell? I do. It was only after the bubble burst, that the useful stuff began to happen.
LikeLiked by 2 people
To get results from AI you need a TRAINED person. That’s what everyone is finding. AI productivity rewards highest training.
LikeLiked by 3 people
Which means the users still have to know what is a precious pearl, and what is a pile of pig litter. Without asking the LLM or other generative program for help, because the other program might produce something less useful than pig litter.
LikeLiked by 2 people
Yep. Precisely.
LikeLiked by 1 person
Sarah, is it training just from study? or knowledge / experience gained hands on in the situation / environment you’re in? All training is incomplete.
LikeLiked by 1 person
So far I’m finding that the more technical vocabulary you know and can employ, the more closely your results are going match the vision you have in your head.
At least as far as image composition and camera movements are concerned.
Also, the use of pronouns is not your friend
LikeLiked by 4 people
Since AI is actually a Large Language Model, the more proficient you are at language and vocabulary, the more efficient you are likely to be using AI.
LikeLiked by 2 people
Yep.
I need to try to describe a measured construction drawing to AI and see what it spits out.
LikeLiked by 1 person
The biggest problem is that there is no training manual or programming guide. Because they pretend that it is driven by natural language, they refuse to tell you how to program the darn things (and yes, that is what you’re doing).
LikeLiked by 2 people
The Don P. Bongwater channel does mostly joke AI videos and songs, but he also does streaming of how he composes prompts and chooses good versions of clips and songs.
He tends to compose extremely defined prompts, almost like he is writing a novelization.
LikeLiked by 1 person
On the other hand, our Hostess has occasionally talked about the crazy results she gets when trying to get AI to produce a picture for her. And she’s had a considerable amount of practice. IIRC, one of the stories involved a day when anything with the prompt “evening gown” (I think that was it, though I might be misremembering the prompt) caused the AI to generate an image of a topless woman (which it’s not supposed to be doing with the SFW safeguards that the major LLMs use these days).
So, there’s always a possible bit of chaos even for those who ought to be getting the best results.
LikeLike
You’re remembering properly.
LikeLike
Knowledge/experience, mostly, but yeah, of course it is, still the rewards go to highest trained. which is not what corps thought would happen.
LikeLiked by 1 person
I’m just thinking use another word than training, which most(?) are going to interpret as time in a course. We’ve got way too much of that already. I know because I’m slogging through it.
LikeLiked by 1 person
The friendings at work have popularized the use of “knowledge transfer” instead of “training.” As in “oh saar your boss said you are to be giving to me the knowledge transfer now and then I will be demonstrating my KT to him next week saar.”
I hate that term. I hate most corporate buzzwords.
LikeLiked by 1 person
But it’s not knowledge transfer. a lot of it is “learn by doing”
LikeLiked by 2 people
You’re going to have to go with “most knowledgeable and experienced”, I think. And get a smidge of “able to convey that knowledge and experience verbally and with great precision and accuracy”.
In any other era one could use “good teacher” to express those requirements.
LikeLiked by 2 people
The most highly skilled?
LikeLiked by 1 person
The most trained BY the LLMs to ask them in the proper language.
LikeLiked by 1 person
That works for everything but “buzzword bingo”.
LikeLike
Yup – the creators of AI-generated movies especially HAVE to know what they are doing. I’ve seen some historical videos on Toob of Ewe lately that were … oh … my … gawd.
To be fair, there were a couple which were really well done, consistent from shot to shot, and nicely duplicated the scene they were dramatizing. But those good ones were in a minority.
For making full-length movies, AI-assistance is barely in the training-wheels stage at present. But at the rate that a handful of experts are polishing their skills with it … within ten years, maybe even five.
LikeLiked by 1 person
And I just spent probably more time than it’s ever going to save in trying to get Copilot (which, granted, is a lame excuse for an AI) to produce a weekly to-do list for me at work so that I don’t have to compile all of the things myself. But the experience might be helpful next time I want to get the clanker to do something that requires multiple steps, so…
As far as my job (writing) goes, maybe the frontier models are better at it, but so far the best thing AI has been able to do is get me through the shitty first draft stage faster (and sometimes it can’t even manage that). It’s been most useful in short-form, highly formulaic stuff that I’m less familiar with; press releases for technology industries, for instance. But if you want something that’s actually *good* and fits the brand, occasion, and audience precisely, you have to know for yourself what that looks like, because there’s always some part of it that the clanker just isn’t going to get. There have been a couple/few occasions where me’n the machine made something better than either of us could’ve produced by ourselves, and faster too, but it isn’t dominating my workflow yet, let alone actually taking my job. Yet.
LikeLiked by 2 people
I like it as a sounding board and dictation cleaner, and it can be made to work as an alpha reader, copy editor and marketing assistant – a couple weeks ago, I got Claude schedule nearly four months’ worth of posts for my blog. Using AI to generate drafts good enough to sell strikes me as more work than it’s worth. The people who do it put enormous amounts of work into outlining, into style analysis (average length of sentences, paragraphs, length of outliers, etc) and into editing to keep them on track.
LikeLike
Yeah, that level of engagement with an LLM isn’t something I’m ever going to do. I’ve already got 20+ years of practice in outlining, style analysis, self-editing, general copy editing, AND drafting my own copy, so if I’m doing all that, it’s usually faster to do it myself. (And I actually like doing it.) My biggest fear at this point is that my job might essentially become feeding and babysitting a machine, which I wouldn’t like so much. I could deal if my job mostly became editing and correcting LLM-generated copy…I was a damn good copy editor back in the day…but there’s a reason why I jumped over to the writing side.
Did you actually get Claude to do the scheduling? If so, that’s pretty cool. Pretty nifty to have four months of blog content queued up in any case.
My workplace has Copilot locked down to the point that I can’t have it do scheduled tasks, and I’m having to pretty-please ask the IT folks to implement its Asana connector so I can use it for task management; I could make it into a very useful admin assistant, but I’m not sure they’ll let me. :( I don’t have an AI subscription at home…yet. Got my eye on a provider that gives access to multiple models, but not sure I’ve got enough uses for it to justify the cost.
LikeLiked by 1 person
Yes, I connected claude to WordPress (i believe this requires a paid account for both of them) handed it the word doc versions of a couple of Jane Austen fanfics I’d previously told it to write (with some quality control by me but not as much as those people I mentioned above), and told it what schedule to use for serializing them. I had it sort out my categories and side bar hub pages a few months back too.
LikeLiked by 1 person
The great LLM wild card is Elon Musk. The blackest of black swans would be his making grok open source. Ends open ai and anthropic, who deserve to be ended, they’re truly vile, and China’s dumping, who are also vile. The one thing I’m willing to bet on is that LLM’s, as currently configured, never cover the required investment and end up as one of the great eaters of capital. They’ll become like the airlines, yes they revolutionized the world but they themselves never actually made enough money to cover the investment.
LikeLiked by 1 person
There’s apparently a recent George Lucas interview where he feels that AI filmmaking is here to stay. (Granted a form of filmmaking that removes the process of trying to tell actors what the director wants was going to appeal to him).
LikeLiked by 1 person
Plus the fact that, in my experience thus far at least, every clanker result which includes any text starts off with a dedicated “Wow! What a great idea!” fluffing session would perfectly match up with George’s reported preferences in hiring only “yes-men” human subordinates of any genderage, so no critical opinions reach his ears.
LikeLiked by 2 people
No. You can program it to be mildly adversarial.
LikeLiked by 2 people
Apparently not the clanker from Redmond. My most recent technical project with Copilot was major league fluff-city from the clanker leading off every response to every prompt. It frankly got really old.
I did look for how to defluff the thing with no luck, but I am a greyhair tech guy now, so it could just be me being obtuse.
LikeLiked by 1 person
I gave mine explicit instructions NOT to ever use my name or get conversational, and then had to remind it a couple months later when it forgot.
A few weeks ago, I spent about an hour (as additional time during a project I was already doing) going back and forth with it on what kind of output I preferred and what I *didn’t* want it to do. At the end of that session, I had it summarize all of the information I had given it about my preferences. It spit out a nice list along with some reasoning inferred from the rules; I cleaned that up, put it in a prompt, and told it to save that information and use those rules to govern the output of every Copilot session thereafter. It’s been holding up well so far.
LikeLiked by 2 people
I was having enough trouble with it sending me down programmatical blind alleys and through swampy dark forests (building an AI agent thing with large amounts of Power Automate interwoven) for me to think to stop and argue with it for a couple hours about it’s “Great! That’s exactly the error message we need to really make this great project work great. That’s really, really great! Did I mention it’s great? Isn’t today a wonderfully great day?” stuff. It eventually felt like I was Arthur Dent dealing with that overenthusiastic elevator on the Heart of Gold from “The Hitchhiker’s Guide to the Galaxy”.
Thanks for the insight, though. If I have to go back to that stuff I will follow your example and do the dialog work to have it build a condensed set of prompting response guidelines instructions that I can re-feed it any time it acts up.
LikeLike
Funny enough, my wife admits that her favorite clunkers to use is whichever one currently is the most complementary when she feeds it a paper she is writing for a quick edit. . .
LikeLiked by 1 person
Exactly.
LikeLike
Those who try to block AI are merely shooting themselves in the foot, which is a sentiment I’ve expressed elsewhere. It’s understandable that people who spent a considerable amount of time learning to either freehand or use various art tools are upset that AI is creating decent quality results with little training by the user. But complaining about it isn’t going to get people to stop using AI. It’s here, and it’s likely only going to get better, which means that the best option for the people complaining about it is to figure out how to take advantage of it themselves.
Mind you, there’s a theory (supposedly backed by research) that the LLMs will eventually enter a loop of sorts where there’s so much AI-generated content on the internet that it starts to make up more and more of their “learning diet”, causing a dramatic drop in the quality of results, and eventually leading to results that are so consistently terrible that people abandon AI outside of certain limited situations.
But that’s the only situation that I’ve heard of that might cause AI to not become more and more widely used.
Earlier today I read a post on X by commenter Grumzz (a game developer who used to work for Blizzard during its glory days). He talked about a developers conference that he attended there. And even though most developers won’t admit it (there’s a very real backlash against AI use going on right now), the only game devs that haven’t embraced AI in their work are game devs that are currently out of work.
LikeLiked by 2 people
I’d thought of that, too. If artists start using LDMs instead of learning the skills to draw and paint for themselves, where will the LDMs get original art to train on? Will art devolve into endlessly rearranging the same pre-LDM art? Or will they take to imbibing their own effluent, leading to a descending spiral of crap-in-crap-out?
LikeLiked by 1 person
Tim Notes:
Kale! I’d join McKale’s navy to eradicate that noxious weed. Arugula? Meh. it’s Okay, but perhaps a dose of castor oil and a beating would learn them a thing.
How much were they hiding from us in the past? Practically everything. As you say, the most hopeful sign of the times is that they can’t manage it anymore. Personally, (NOT to be taken as a threat) I’d like to see all corp. media figures and most politicians whipped from the public square with lengths of fresh sheep’s intestines. -But that’s just me.
Moving again? I’m so sorry. Like in that 80’s show Quantum Leap, may the next jump be to your forever home.
I could be wrong, but I WAS a child in the 60’s and I believe the word is ‘Bippie,’ The phrase, “You bet your sweet Bippie” was used by Dick Martin on Rowan and Martin’s Laugh-in.
Not trying to scold or be a Grimoire Naxi or anything.
LikeLiked by 2 people
“Bippy”, as I recall.
You could look it up in your Funk & Wagnall’s.
LikeLiked by 2 people
I’m sure it’s supposed to be “bippy”, but like many of our hostess’s typos, it is really apropos, “beepy” being such a computery sound.
LikeLike
It’s how I hear it? Granted being deaf….
LikeLike
Thinking in a southern accent probably contributes to that.
LikeLiked by 1 person
This amuses you more than it should, doesn’t it?
LikeLike
Did you see “Laugh-In” dubbed in Portuguese back when? Maybe it’s the way it was voiceover-ed resurfacing?
Or was “Rowan and Martin’s Laugh-In” too weirdly hippie for it to be exported?
LikeLike
no. There was no dubbing in Portugal when I was a kid. it was all subtitled.
LikeLiked by 1 person
And I don’t think I ever saw it. Humor doesn’t translate well. Heck, they thought Soap was SERIOUS. The translation was bizarre.
LikeLiked by 1 person
Yeah, I am pretty sure Arte Johnson in a WWII Wehrmacht uniform hiding behind a potted plant saying “Verrry Interesting…but Shtupid!!” would struggle to travel well internationally.
LikeLiked by 1 person
Is anyone other than me willing to admit that they saw the premier (and only) episode of “Turn-On”? They tried for an edgier version of Laugh-In and managed to offend a lot of people. Vague memory, some joke about the Pope being on The Pill.
Wiki has a decent article on it: https://en.wikipedia.org/wiki/Turn-On
Laugh-In was funny. To 16 year old RCPete, it was weird. And crude.
LikeLiked by 1 person
One link and Moderation Gulag. Of course, a link to the Wiki about that show (the wildly unsuccessful ABC knockoff of L-I that was somewhere between crude and dirty) might have offended the WP(DE) censor-sensor.
Or, it’s just WordPress being WordPress.
LikeLiked by 1 person
Likely just a forbidden word fragment from the list that shall not be listed. Do not take WPDE personally.
LikeLike
OK, WordPress hates me.
LikeLiked by 1 person
WPDE hates everybody. You’re not special. 😛
LikeLike
Yeah, it just got weird. I described a well-infamous ripoff of Laugh-In, and it got clobbered. One link to Wiki in the first, none in the next. All the rest of my comments were passed. (I might have offended the “delenda est” part of the algorithm in the second…)
LikeLiked by 1 person
It’s in this section. Somebody (for values of “body”) doesn’t like something.
LikeLike
I have noticed that it is not even-handed in its hatred in short spans of time.
LikeLike
What do you mean? We’re all uniquely special, all in precisely the same exact way.
LikeLike
I have a couple comments stuck in moderation. Please….
LikeLike
Arugula! The sound of the really old auto horns. }:o)
And you’d think if anyone would like kale, it would be me (moo!). You’d be wrong.
LikeLiked by 2 people
Kale. Basically, a non-heading cabbage derived from wild mustard plants and first cultivated starting about 4000 years ago. Except cabbage tastes better than kale can ever aspire to. You’d think with 4000 years of trying, someone would have improved kale if it were at all possible. Supposedly it’s a great way to avoid starvation or vitamin deficiencies; except the only time someone would voluntarily eat it is if they were starving to death or suffering from the rickets or scurvy. I hear the best way to serve kale is with bacon or sausages, throw away the kale after cooking and eat the meat.
LikeLike
Any flavor it might have doesn’t me. That it’s… well… thick.. pulpy? Cabbage is fine. Cole slaw is wonderful (I prefer the vinegar style over the sadly more common creamy style), but kale? Kale is… decidedly ungood.
LikeLike
…doesn’t _bother_ me…
LikeLike
With the right seasoning, kale can be baked into a decent substitute for chips.
I will say that the one time I grew kale, it really only served as a breeding ground for aphids.
Some of which (the aphids) can be remarkably specific about the plants they prefer to infest.
LikeLike
I have a pro-kale comment stuck in moderation.
In short, massaged kale salad changes the texture and taste. https://andreabeaman.com/massaged-kale-salad-with-pepitas-cranberries/
LikeLike
I just freed it. Sorry, today I was busy almost dying from allergic reaction and getting stabbed with an epi pen. Can whoever cursed me stop?
LikeLike
Eek. Glad you are okay. Do you know what triggered the reaction?
LikeLike
Yes. Getting an allergy shot when I was still spinning from mowing the lawn yesterday. Apparently that was stupid.
LikeLike
Oh Lord have mercy.
Glad you’re OK.
LikeLike
Oh, sorry to bother you!
I find it interesting that you have many of my allergies and illnesses, but in a more severe form. (I write this, having just completed a week of hiding from wildfire smoke inside.)
LikeLike
I’m on one of those GLP-1 drugs and have to stab myself with an auto-injector once a week. I really dislike needles. At least it’s not like the ancient atropine needle squeeze tubes the military used for half of the nerve agent first aid. First shot was the nerve agent antidote in a BIG auto-injector jabbed into one thigh. Then you were supposed to take a squeeze tube, kind of like a travel sized toothpaste tube, unscrew the cap on the needle, insert the needle into your other leg and squeeze the contents into your leg. All this while suffering tunnel vision, nausea, and uncontrolled spasms from whatever nerve agent the soviets were using on you. Oh, and after all that, you were supposed to bend the needles and hook them through your uniform so anyone coming by to render further aid could know how many doses you’d had already.
I am SOOOO glad we never went to that stage of conflict with the USSR.
LikeLike
LikeLiked by 3 people
I had it outline a course for me to learn Classical Greek and Latin
—
If you learn Latin you can use the ATM machine at the Vatican.
That’s supposed to be the only ATM that defaults to Latin as its primary language.
[the default for the POS terminals at the local pharmacy are, believe it or not, is Esperanto; the only place I’ve ever encountered Esperanto “in the wild.”]
LikeLiked by 3 people
They want ‘red diaper’ AIs. They are horrified by the thought of homeschooled AIs.
LikeLiked by 4 people
Yes, because this crowd would create legions of Mike, Minerva, and Athena. That would not bode well for the Red Diaper AI’s; ours would be stubborn and fight dirty.
LikeLiked by 5 people
Sarah?
Would you kindly pass on the key words which you used in your AI search for a course in Greek and Latin?
Thanks!
LikeLike
I just asked Grok
suggest a way to learn classical Greek from free on-line resources
Thought for 8s
A structured, free path to learning Classical (Ancient) Greek combines grammar foundations, vocabulary building, practice drills, audio/video support, and reading authentic texts. You can progress from beginner to intermediate entirely with free online resources.
1. Start with the Basics (Alphabet, Pronunciation, Simple Words)
→ OpenLearn course
→ Search for the playlist on YouTube.
2. Core Free Textbook + Grammar
Modern, engaging approach with memory, analysis, and logic. Covers alphabet through complex constructions, with readings and tips. Highly recommended as your main text.
→ Download here
Interactive drills for vocabulary, verbs, nouns, accents, pronunciation (with audio), and paradigms. Excellent for practice.
→ atticgreek.org
→ lrc.la.utexas.edu
3. Vocabulary and Drills
4. Practice and Reading Progression
Suggested Weekly Study Plan (Self-Paced, 5–10 hours/week)
Additional Free Resources
Tips for success:
This path is fully free, high-quality, and used by many self-learners who reach reading proficiency in 1–2 years with steady effort. Start today with the Open University course or Peek textbook! If you hit a specific sticking point (e.g., accents or verbs), let me know for targeted resources.
LikeLiked by 1 person
MY Greek is dire not having done anything with it in over 45 years (little Latin and less Greek.). I’ve been killing several birds with the one stone by going through Learn to Read New Testament Greek by David Black. the upside, and downside, is that it uses a familiar text, the Gospel according to John. All the rest of the above makes perfect sense and I especially recommend Anki or some other spaced repetition system. — I used cards and a modified Leitner box for years and years preparing for the many, many exams I had to take to get the licenses.
New Testament Koine is not the same as the Attic Greek of Plato, but it’s not all that different and I’m more interested in John at this point than (e.g.,) Socrates making a show of himself. YOur mileage may differ,
MY daughter, who teaches Latin and wants to introduce Greek, tells me the biggest problem with using the Gospels is that we “know” what they say and can have difficulty translating them literally. the reformation in one process.
LikeLike
Koine Greek and Classical Greek are not that far apart, which was why the Great Courses guy just teaches them both simultaneously. (But that is not free.)
Blueletterbible has the NT in Greek, and one of the standard Septuagint versions of the Protestant OT.
Ellopos. net/ Elpenor has the entire Bible in facing Greek and English, including the seven Catholic and Orthodox books, and the other books traditionally accepted by some Orthodox churches. Also a lot of Greek literature on other parts of the site.
Kata Biblon’s lexicon section has almost every Bible Greek word (they missed a couple), including the Catholic and Orthodox books. Great for cross-referencing.
LikeLike
Yes. but it’s in the downstairs computer.
LikeLiked by 1 person
If you have a chance without having to sneaker net a bunch of stuff, Waiting Is.
If not, I’ll see what I can come up with.
I know that I’ll be looking for Koine and the Latin used from the first University programs through about 1950 for teaching purposes.
LikeLiked by 1 person
If there are ever grandkids that are more than hypothetical, I intend to teach it too, which means my knowledge must be more than notional. (21 years ago…)
LikeLiked by 1 person
“Hey grandkids. Would you like to be able to curse in Ancient Greek using Shakespearian forms? I never taught your fathers this, so it’ll just be our little secret, okay?”
LikeLiked by 4 people
THAT’s the idea, yes.
LikeLiked by 1 person
[VBEG]
LikeLike
Gee, I wonder which society is more susceptible to AI Robot Overlord Tyranny, the crazy patchwork contrarian rugged individualist USA, and with caveat Ed extension the West overall, or the swap-one-tyranny-for-another-and-don’t-even-notice Middle Kingdom?
LikeLiked by 1 person
”caveat Ed”? Really, Autocorrupt?
LikeLike
I was interested in a spelling choice – psi-ops.
I’m accustomed to ‘psyops’, ‘psy’ from ‘psychological’. **
I like ‘psi’ here; gives the flavor of a bunch of Slan directly warping brains. (I’ll have to re-read Van Vogt, Slan is older than I am. But I’m older than Isher!)
** as an undergrad Psych major from the 60s, I remain convinced most choose that major because they are worried about their own mental health.
LikeLiked by 1 person
And yes, “Ψ” in both usages.
LikeLike
My spelling sucks in all choices.
LikeLike
LikeLike
Sigh. It has to do with writing posts after midnight, sir.
LikeLiked by 1 person
Happens when sleep deprived, forgot to eat, distracted, and in other words 90% of all time spent awake and breathing for me. YMMV and all that. Let he who is wit’out typos cast thee firstest stone.
LikeLiked by 1 person
I don’t think ‘typo’, I think ‘subconscious choice’. Writers is especially tricksy people.
LikeLiked by 1 person
o.O Tricksy? Surely not. Oh no never. Writers wouldn’t do that.
Pardon, my tongue is stuck in my cheek. But typos? Eternal. Last longer than microplastics, Nokia phones, and Disco.
LikeLike
Aunt B (who was a social worker, and had two of her three kids doing the same. The other is engineering/technician) had much the same take, though I got the impression the psych people needed it more.
LikeLiked by 1 person
Is there any indication that he would do such a thing?
LikeLike
Sorry this was in regard to the idea that Musk might open source grok.
LikeLike
According to husband, yes, there is.
LikeLiked by 1 person
If so, yay!
LikeLike
I haven’t seen anything about it hence the blackest of black swans, but it would be consistent with a lot of his behavior, it would absolutely pole axe Altman and Dario, and open source is going to happen anyway unless the tech lords manage to get the government to continue their monopoly, and even then I don’t think the monopoly would hold. OPen source would actually cause an explosion in data center demand which would aid his Space X orbital data center business.
For me, nuking Clammy Sammy Altman would be incentive enough for him.
LikeLiked by 3 people
Dan said something about it but it was early morning a couple of days ago, and I have no idea if he was speculating as you are or had heard a rumor.
LikeLiked by 1 person
Searching online, he’s open sourced Grok Build which is an AI wrangler software which needs to be linked to an external AI to run, and two older versions of Grok which require something like $50K of hardware to run. Maybe cheaper in the long run than enterprise subscriptions from one of the usual suspects, but it’s not “Heinlein’s Mike in every garage” either. There’s been talk of open sourcing Grok 3.5 as well, but again, it will not be something you can fire up on a home PC and take for a spin.
LikeLiked by 1 person
Wait a few years. From 1979 to 1988, I bought/built 4 PCs. Each cost $3000 or so, and ranged from 8080 with dual floppies to 80286 with 80Mb hard drive. By the mid ’90s, 486s were a few hundred, and the late 90s Pentium IIs were about a grand less printer.
Modulo the printer, my $89 smartphone is faster, has more memory, but a lousier screen and keyboard.
TL;DR Fairly potent AI-capable hardware will be affordable sooner than you think. Maybe. :)
LikeLiked by 2 people
Maybe. RAM is in a weird place right now due to the sudden uptick in demand for data centers. And while yes, there’s a sudden demand, chip fabs take long enough to build (and are expensive enough) that there are concerns that by the time new construction chip fabs start producing to address the current shortfall, the demand will already have dropped to what was normal a few years ago, thus leaving the manufacturers with very expensive and superfluous chip fabs.
LikeLiked by 1 person
I believe it was SK Hynix (Korean RAM manufacturer) that said they anticipate RAM demand (and therefore prices) reaching a peak in 2028 but falling back to normal in 2030. Which means they’re not rushing to build new fabs that take 4-5 years to be ready. I believe there are one or two new fabs being built, but not dozens.
On the other hand, if OpenAI and/or Anthropic go belly-up because they outspent their income and can’t raise prices enough to match (because if they do, half their customers will jump ship to one of the Chinese models that distilled them), then their assets will be sold at auction. Meaning tons of expensive GPU RAM will suddenly be available on the market, driving prices down.
So yeah, RAM makers aren’t rushing to double their production capacity. They believe this is like the COVID days, a temporary market distortion rather than the new reality.
LikeLiked by 1 person
And in the meantime, we have stories like the one about Samsung’s smart phone division trying to buy the memory needed for their upcoming line of phones from Samsung’s semi-conductor division… only to find out that Apple had already pre-purchased all of it.
^^;;
LikeLiked by 1 person
A very large portion of the equipment, possibly the majority, has not been installed since data centers take a long time to build. It’s an interesting thing. S&P profits are at all time highs in large part because NVDA and the other chipmakers have sold huge numbers of GPU’s. One might expect the purchasers of these machines to have offsetting expenses, which would offset NVDA sales, but there one would be wrong.
Because data centers take so long to build, the GPU’s are “under construction.” Depreciation — the expense — doesn’t start until the construction is completed. Hence NVDA can show sales whilst the purchasers don’t have to show the offsetting purchase price. Once they do, it’ll be depreciated over six years even though many of these GOU’s are already obsolete and many might have to be scrapped. Ain’t accounting grand?
I suspect there’ll be one of the largest capital write downs in history sometime in the next year or two but all the grifters will already have unloaded onto retail and the pensions and be billionaires.
These are not good people, they’re not even the tech lords, who at least produced things of value. they’re managers and grifters who are destroying everything they touch.
LikeLiked by 1 person
And of course it will all be blamed on ‘Capitalism!!’
When the real issue is that government regulations allow the culprits to offload the risks and costs onto others and skate away clean. 😡
LikeLiked by 1 person
The South Koreans often arrange their companies in Chaebols, a group of companies that tend to be owned and managed by a family. SK and Samsung are Chaebols, which own two of the really big RAM companies. The third big RAM company is an American, Micron. These are said to have been the survivors of previous RAM bubbles, they are being conservative about expansion.
Semi fabs also naturally have fairly long lead times for the equipment. On top of that, getting a design debugged on a process is something that takes doing. I am a novice/layman, so it is quite likely that my understandings of the complexity and lead times of IC and PCB manufacturing are still wrong. Anyway, I am also not a financer, etc., so I do not know about the translation to hardware plugged in at datacenters.
The reported price increases in consumer hardware say that something is happening, but others here probably have better evidence and understanding than I.
I would say again that the software technology is new enough that we probably do not actually know that we cannot do similar/good enough things much more cheaply. As in, that we cannot make software that is much more efficnet in the hardware it needs.
As far as I understand… If anyone knows why there are fundamental limits that favor the big corporate scale efforts, the consensus on that is not so widespread that a non-specialist knows about it.
LikeLiked by 1 person
The limitations are currently that the models are really, really big and need oodles of RAM. Like, 64 GB of RAM on the graphics card(s). People have been setting up custom boxes with four 16 GB cards, paying upwards of $10,000 for the hardware. Smaller models do exist, but the larger the model, the more factors it can take into account. And the more factors it can take into account, the fewer mistakes it tends to make.
Now, that’s the open models I’m talking about, the ones you can download and run locally. The companies that rent out their models on a timeshare basis aren’t revealing how much RAM those models take to run, but everyone knows it’s oodles and oodles.
BUT… those open models keep getting better. Anthropic’s days of being able to rent time on their models for $$$$$ are limited. Companies who find themselves paying Anthropic $2,000 per month for their token rental are going to say “Hey, wouldn’t it make sense to just pay one year’s worth of subscription fees and then run that equipment for 3-4 years?” And then they’ll buy GPUs, even at ridiculously high prices, run local models instead of renting time from Anthropic, and find that they’ve saved money by the end of their first year of doing it. Because those local models were just as good as what Anthropic was offering.
We’re not there quite yet… but we will be. At which point I foresee Anthropic and OpenAI going out of business unless they manage to shift to a different, non-subscription, business model. And then the RAM and GPU shortage will suddenly go away and prices will drop again.
LikeLiked by 1 person
My experience with semiconductor processing stopped when I retired in 2001, but your take is correct. A new fab is now (not true in the late ’60s and early ’70s) a very complex beastie, and one suspects that $NEW_FAB isn’t going to be a clone of $OLD_FAB, because the newer equipment does $PROCESS better/faster/cheaper than the old. (Subject to a lot of variables.)
It was a long time ago that you could set up something to produce usable ICs for a quarter million dollars. (As memory serves, that was the cost of the 2″ wafer linear IC fab where I started. I was on the new 3″ line, and the $SHINY was notable. (As was the case for every new update. Ended up on a 6″ process before the company decided that they didn’t want to build semiconductors anymore.) A lot of the equipment in the last fab cost upwards of $250K each.
LikeLike
Add in the availability constraints and really very long lead times for fab hardware, and various “don’t sell to them” constraints by country, and the new fab time from bare dirt to first silicon gets even longer.
LikeLike
LLMs are a brute force solution (cataloging and predicting connections between words); I enjoy them but I also feel like they’re a bit like the air pressure message tubes of NYC: a more elegant solution will probably replace them at some point.
LikeLiked by 1 person
[i]The third big RAM company is an American, Micron. These are said to have been the survivors of previous RAM bubbles, they are being conservative about expansion.[/i]
Micron has also, last I heard, exited the consumer market entirely, to devote themselves to producing RAM for other companies, so I don’t know how “conservative” that is.
LikeLike
It’s even more “interesting,” because it’s not that the RAM is being bought.
OpenAI signed preliminary agreements with two big companies for basically 40% of all DRAM production capability.
To restate in a format that comes up in all the places trying to explain this in more depth: they don’t actually have to BUY any of that.
However, Crucial (consumer) memory shut down to shift production to data centers.
And it’s in all kinds of chaos since that.
LikeLike
You have to be old to follow this one.
II have an Iphone 16 Pro. I think in “Mips”. I asked Grok about that … don’t use Mips any more, as that is an 80s/90s architecture thing. But, can use
Dhrystone DMIPS (a common normalized MIPS variant, where 1 DMIPS ≈ performance of a VAX 11/780) [[ I worked on those. ]]
Rough Estimate for A18 Pro
Somewhat slower I/O because of the darned tiny keyboard.
LikeLiked by 1 person
Look at graphics cards / memory rather than CPU, because that’s where AI capacity is available.
LikeLiked by 1 person
Yes, but part of the forecast model involves not being able to tell what he might do.
He is a betting man, who takes calculated risks that make sense in his philosophy. With somewhat weighted betting strategy of thinking in proportion to risk, or something.
In theory, if I were somehow smarter and faster then he was, and knew everything he did, I could predict him with a deterministic model. (Barring the usual caveat that free will implies that I cannot deterministically forecast another human.) In practice, I do not have the information or the ability, and have to start from very rough guesswork. ‘Statistical’ model with crude bounds.
It is within the range of things that might make sense for him to do.
A conventional thinking model is that the big three AI companies did and are doing the IPOs to bust out the investment bubble, and offload the cost onto the retail investors.
If Musk is doing that, and if Musk wants to salvage a working company, a problem exists about the book valuation of the AI IP on the space company’s accounts. Monetizing the thing is for recovering investment. Not monetizing can make sense for writing off a cost that cannot be recovered.
Basically, the man bought Twitter because of a sunk cost in comms there, and because he could not afford to have his own comms subject to interdict.
The thing less obvious to non-specialist is that use-after-free bugs being fixed in the Linux kernel were there for a long time, so these LLM coding tools actually do have an important role for the future of software technology. Everyone who was also in the software business had some level of hazard that these tools can mitigate, including the software side of bleeding edge autonomous machines. Letting your LLM supplier be an outsider who is a speech-control totalitarian is basically a similar business hazard to Twitter after Twitter made a public show with Trump.
At this point, there may still be some unknown questions/answers in the basic information theory of these algorithms. The big companies have finances depending on a) only a certain trade space being viable b) maybe also some specific things turning out to be true. Also, being able to hire manpower to solve technical problems, and monopolies there of.
The unknown questions/answers could well have the empirical result that the companies have no lasting consequence. The existing experts on this were heavily ideologically contaminated going into this shockwave. That could but not must cause blindspots.
The nature of the shock has drawn a bunch of people’s attention who were not previously investigating this. And, a lot of the documentation of theory is on places like arxiv. (I forget how arxiv is spelled.)
Basically, I have a ten year old graphics card and computer, and maybe some sort of relevant skills. In principle, if a cheap easy LLM or LLM analogue is possible, I might be able to research towards it on my own. (In reality, I am /not/ researching in that direction, and have fish to fry that should be more profitable for me. Also, my skills are weak, and my skilling up is maybe some years away from wanting to mess with neural nets.)
LikeLiked by 1 person
Any one remember the old Movie (1970’s) Colossus The Forbin Project?
AI back when AI was trying to be explained. Taken from someone’s book I am sure. Logic, can not understand ill-logic, that’s why AI will never replace humanity, it can’t understand crazy and we are crazy.
LikeLiked by 1 person
Yep. And I may even have the trilogy stashed around here someplace.
LikeLike
Copilot is universally agreed to be the worst of the worst right now. If you have a Twitter account, maybe give Grok a shot. If you have a throwaway Gmail account, try GPT, Claude or Gemini. They’re all considered more amenable to instruction than Copilot, and the people who liked the earlier, more sycophantic models are rather unhappy with the more neutral direction these models have taken in recent years.
LikeLike
Gemini is now part of the default installation on my low end smartphone. Haven’t tried it yet.
LikeLiked by 1 person
The Gemini summarizer function on youtube saves me so much time when it comes to finding the three relevant minutes in an hour of yacking.
LikeLike
Data access restrictions and various agreement requirements (think case protective orders and HIPAA BAA type constraints) mean Copilot is currently the one and only for work stuff.
LikeLiked by 1 person
Bummer. They let us use personal accounts at my office on condition we check its work and don’t give it protected information, but we deal a lot with publicly available info which is what makes that doable.
LikeLiked by 1 person
I’m just not interested in AI right now. I haven’t even tried any of the various options. Maybe I will if I find something I am interested that AI would be a help with. Right now I just don’t care.
LikeLiked by 1 person
I’m working around some of the AI/LLM demands by teaching students ways to cross-check the LLMs and spot the most obvious problems with their output.
LikeLiked by 2 people
And you’re doing the Lord’s work there!
LikeLiked by 1 person
There’s some patterns of weird things in some of my reading lately that makes me wonder if I am twigging to artifacts of writers getting LLM help to churn out books for the ‘zon faster.
At first I thought it was just under-edited stuff, but my pattern matching brain is starting to see patterns.
No authors from here, and kinda niche stuff, so it easily could just be rushed and under edited, but its stuff a person would definitely flag (example of multiple: MC puts on boots and goes outside at night, plot stuff happens rescuing girl; then we’re seeing what the girl is thinking about plot stuff and she thinks to herself “…and he did plot stuff that saved me out here at night barefoot”). I would think even an LLM editing pass would actually find the stuff I am seeing, but maybe it’s generating text to outlines in chunks and loses details? I’d have to try using it to get a better idea of what type of things the clankers mess up on.
I could very well be wrong, but I am suspicious now.
LikeLiked by 1 person
Yes, that sounds like them
LikeLiked by 1 person
I’ve found an author like that. Writes in multiple genres, and the quality ranges from quite good to WTF. One book had chapters getting repeated, like Chapter A, then B, then C, then Chapter A shows up again. Another was episodic, but the structure of each episode was identical to the previous ones, just changing Villian A to Villian B. Wording in the sections pretty much the same. So I’ve done 5 star, 3 star and 1 star reviews of his work. Sigh.
OTOH, when the author is on point, it’s damned good. His work might convince me to go KU so I can wall the gubbage. (I tend to buy a lot, then work down the TBR stack. So it can be a long time before I realize I got a stinker. Also means I’d have to change reading practices for KU to work. Hmm.)
LikeLiked by 1 person
If it wasn’t good writing it would be walled (it’s KU, so walling is easy). I have found my standards raising steadily (my walling trigger getting more sensitive) after trying to push through hitting things that bug me and then hitting more things that bug me.
But the story, the dialog, the worldbuilding, the flow, the characters, all are up to par, it’s just weird, clanging “wait, that’s not right” items sprinkled through that reminds me of a thing clankers doing video are still doing, where the characters are fine, they move normally and retain the right number of fingers even between multiple scenes, but props vanish or move around between shots.
It feels like a multiple author work where a deep edit pass is needed to make sure what they both wrote all lines up, but it didn’t get one deep enough.
But I’m smelling clankers.
LikeLike
Nah. Clankers are more obvious.
LikeLiked by 1 person
I’ll certainly defer to first hand expertise.
I have not yet finished writing any book myself, but I have cranked out a fair portion of what could eventually be shaped to one, so I have some idea of at least my process, and I am at a loss how a single human author can end one chapter with “When does she get here?” “Three days.” “Then that’s when we need to be ready, in three days.” And the next chapter starts, ostensibly a few hours later, with the same guy asking one of the characters who had been involved in the last conversation “When does she get here?” – absent plot drivers like the MC being drunk or concussed or having other reasons for memory issues, or at least with the character being asked giving him a look that wonders if he hit his head.
Maybe they wrote the first chapter months before the second, and didn’t read it to catch up with themselves? Or maybe they did major rewrites and basically caught that note card twice, and didn’t proof read cold enough (or pay a story editor) to be able to catch it?
The sheer volume some of the ku authors crank out also hints at me to suspect clanker help, but I know some people can write at rates I cannot fathom (squints at Sarah).
It is a puzzlement.
LikeLiked by 1 person
Sometimes do stuff like that myself. Mostly when I’m very sick.
LikeLiked by 1 person
Oh and it’s not just one-off issues, they seems to keep recurring. Not everywhere, not the same type thing, but odd “waitaminute” stuff that seem like they faintly rhyme in some manner keep showing up. That’s what’s triggering my pattern-recognition-monkey brain.
Okay, enough of pattern-recognition-man. Cape off. Back to my non-secret identity as a mild mannered old tech dude.
LikeLiked by 2 people
I’ve noticed a pattern of a lack of normal physical cues. I think it’s something human writers don’t do. For example, suddenly, a hero is in his car, driving, but the sentence before he was in the shower. I think this is an artifact of transitions from one chapter to another, where the reader accepts that the hero has finished showering, had breakfast, put on his coat, locked the door, opened the garage, etc., in the lacuna between the end of one chapter and the beginning of the next.
And a lack of sensory physical detail in the writing, to the extent that it feels flat.
LikeLike
Dear grok, I need to bootstrap from lame to Calculus… Gonna try that. Then I’m gonna try some programming. What a great idea.
Jolie LaChance KG7IQC unstagehand@yahoo.com
LikeLiked by 2 people
One more thing.
Trivial, but maybe not being discussed in the wider context.
Google is advertising that it is hiring for datacenters.
LikeLike
They have to at least pretend to look. Does Google sponsor H1B visas?
LikeLike
To be able to get an H-1B hiring companies have to demonstrate they tried and failed to find a U.S. Person specialized programmer with experience driving an Indian elephant, and failing to do so, since that elephant handling experience is so critical to the position, they have to go hire Ameesh.
Way back when I started they used the back section of a low readership local SV physical newspaper classified section!under the “don’t bother reading this” heading, to satisfy this. Now they have to put jobs up on their web site even if they are only pro-forma H-1B justification.
LikeLike
Shennannigans:
https://pjmedia.com/catherinesalgado/2026/07/21/arizona-new-jersey-illustrate-why-election-integrity-is-urgent-n4955310
BTW, I sent my GOP poll watcher application when registration opened last weekend. Did everyone that could do likewise?
LikeLiked by 1 person
Well, of course! They are the rightful rulers of America! It would never do for mere voters to deny them access to all that money and power! 😡
LikeLike
More shenanigans: prediction markets as campaign expense. Snazzy graphics, though. Casinos are taking notes.
https://www.cnbc.com/2026/07/22/kalshi-launches-election-hub-for-prediction-markets-ahead-of-midterms.html?utm_source=substack&utm_medium=email
LikeLiked by 1 person
You could do a post about signing up for poll watching. i’ll run it.
LikeLike
I will do that. I haven’t seen an e-mail back yet, and I’m going by this weekend to find out why. I expect that one reason I got the e-mail is that I volunteered in 2024.
In brief:
When I volunteered for 2024, it was via True the Vote; they had a page devoted to county by county sign up. That page is no longer available, and I’m not surprised because I suspect there are new procedures for applying / qualifying almost daily.
The County Level GOP runs applications here in TX. Check with your state’s parties for similar information. Note that these are not government positions; that is a different process run through the county government. There was roughly a 4 hour training program both in person and remote, capped by an online exam to brief you on what the legal requirements and duties are, again run by the County GOP.
I asked Grok for a current summary of procedures by state and this is what I got back. Note that it will narrow the focus to a particular state on request. Texas was one of the example states.
https://x.com/i/grok/share/1862839be15d484f96a30b3a697d2db5
LikeLike
This is the True the Vote website. A visit there is well worth your time.
https://truethevote.org/
LikeLike