The Histrionics Will Continue Until We Get Our $T IPO

My wife & I had a very nice, mostly Canadian, vacation, 8/22-9/4. So doing a bit of catch-up here.

Seriously, you read the headline, you say "Oh shit", & then you remember, this is the 20th to 30th time in the last year that OpenAI or Antropic - and now Nvidia - has made some overblown and/or alarming post about the state of GenAI. Maybe claiming we are at AGI - which we totally aren't.

LLMs. aka Autocomplete On Steroids, are still, from the last I heard, producing at a minimum 9% error rates.

Agents build on that to get to a 15% MINIMUM error rate.

YOU CANNOT BUILD RELIABLE SYSTEMS WITH ERROR RATES THAT HIGH.

Yes, the stuff that GenAI produces can be impressive, but, still, it is by definition BULLSHIT: it looks like it might be credible, but, maybe it is completely & utterly false - completely made up out of statistics, with 0 real knowledge of anything. Anyone producing AI slop output without serious human fact-checking should be completely liable for criminal charges.

---

LLMs are sucn an interesting technology. It's incredible that 1 algorithm, the Transformer, discovered by the Google Translate team, could create such an upheaval.

It's so tragic that psychopathic TechBros decided that this was their ticket to generational wealth.

LLMs are a normal technology. They are a great tool for experts who can readily check their output. For everyone else, they are Acme Slop Generators (going into my table of names).

Wow, Small Language Models running on PCs or phones are 98.6% as accurate as LLMs running on $B of hardware??? Oops.

From Joachim Klement on SubStack:

If this is true, the hyperscalers are toast

As you can see, in every domain, the best SLM is able to find the same or better answers than an LLM in 90% or more of the cases, with an average across all domains of 98.6%.

---

From the Bard of the Revolution, Cory Doctorow:

The actual epistemic crisis:

AI's assault on reality is an opportunistic infection.

AI is alarming for many reasons: it's a dangerous financial bubble, an environmental catastrophe, and a tool for eroding wages and labor power. But in addition to all that, AI is an epistemic disaster.

We've had photoshopped images, voice impersonators and visual effects for years, of course, but with deepfakes, we've democratized access to reality-bending images, sounds and videos that appear real but are not. It's harder than ever to know what's true. Politicians and celebrities and activists show up in our feeds, declaring their fealty to this cause or product, or their fury at some turn in the world's events. Battlefields mound high with bodies and influencers marvel at impossible, sumptuous meals. It all seems plausible, and some of it is real, but not all of it, and because we know some of it is fake, we can't be sure if any of it isn't.

---

Again from Doctorow, somewhat more abstract. When you were born determines what doors are open for you?!?!?

Born on technology's third base:

Material forces shape life-chances.

...

By the same token, continuing to give trillions to AI companies because they experienced early success with theory-free inference at scale tells us nothing about how to solve the vast range of problems that theory-free inference at scale sucks at. Doubling down on AI to overcome its increasingly obvious limitations is like doubling down on building post-war suburbs to fix today's housing market.

It's possible to achieve impressive feats because you're smart and hard working and also because you were in the right place at the right time. Historical contingency produced the AI bubble, and it is producing the conditions for that bubble to pop.

---

OK, that clears the links from my Lexington PC, from just before we went on vacation. I will next switch over to the links on my MacBook, from while we were gone.

Here's the home page/directory for my posts on Bullshit. This is post #143.

A Trip to the Great North

Following our summer habit of the last 2-3 decades, my wife & I recently vacationed Up North. We used to frequently visit our middle daughter in Ithaca NY, my older brother in Portland ME, old college friends in the Boston area & Connecticut, & our oldest daughter in Brooklyn.

This year, we 1st visited our Canadian FL snowbird friends in Windsor, Ontario. We drove there Saturday 8/22, had dinner, on Sunday visited the Aquarium & Botanical Garden on Belle Isle back in Detroit, before a fabulous dinner at Take Five in Windsor.

We went over the new Gordie Howe bridge 2x! What a beautiful piece of architecture! The pic is from my Canadian friend Owen.

Monday morning 8/24 we picked up our middle daughter in Ann Arbor & drove to the northeast shore of Lake Michigan, Frankfort MI. A very nice little tourist town. Wow, Lake Michigan water temperature was 37 degrees Farenheit - brrr!.

Wednesday morning, 8/26, we dropped our daughter back in Ann Arbor, parked our car in Detroit airport, & flew to Calgary, AB, Canada. Canada's 5th largest city, at 1.481 million just after Ottowa at 1.488 million. We rented a car.

We did local Calgary sights: Calgary Tower, Music Museum (Studio Bell). We lunched at King Eddie's, affiliated with the music museum. Sadly, we didn't ride the public transit, everyplace we visited was in walking distance.

We drove to the world's greatest dinosaur museum: the Royal Tyrrell Museum in Drummheller, ~100 km NE of Calgary.

It was my 1st time in a Badlands: land along a river which has eroded itself to 100-400' below the surrounding prairie. Amazing!.

Saturday 7/29, we drove west ~150 km to Banff, at the center of Banff National Park, in the Canadian Rockies. We stayed there til Thursday, 9/3. The mountains were so beautiful! So much variety! Sharp points, repeating sawblade edges, houses & temples at the top of peaks! Snow & no snow!

We took the Banff gondola to get a birds-eye view of Banff. The small Tunnel Mountain that sits down near town is so small compared to the surrounding peaks.

The botanical garden (Cascades of Time) in Banff was very pretty.

Lots of good meals, poutine 2x.

1 day drove west across the Continental Divide on Canadian Highway 1 into Yoho National Park in British Columbia, to visit Takakkaw Falls - 4x as high as Niagara! Beautiful!

On our way back, we turned north on the Icefields Parkway, Rt 93. Around 150 km to the Columbia Icefields & the Glacier Discovery Centre. 6 glaciers coming in from the west terminating just before the Centre.

Hmmm, I uploaded this short video to YouTube, hoping I could then embed it in the blog, as I do for music videos. No. Apparently not an option. If anyone can point me at instructions on how to make this work, hep me, please!

https://youtube.com/shorts/-aki_dC_TdY?si=grAwECMHmEdPs5tw

Also stopped at Mistaya River Gorge. Incredible tunnels (?) worn through the rock by the glacial water.

Thursday, 9/3, we drove back to Calgary, stopping & spending the morning at the tourist town Canmore on the way. I had my 1st breakfast poutine: hash browns, 2 eggs, bacon, sausage, bison, cheese curds fried, hollandaise sauce. Yum!

We flew back to Detroit Friday 9/4. The 330 mile drive back to Lexington took just over 5 hours, we made good time, got home right around 10 pm.

Wow, since we've been home, I sleep so much better in my own bed!

A very pleasant vacation, & yesterday I miraculously weighed in at under 200#! Air travel was fairly painless. Only drama was when I almost left my passport in the room safe in Banff. But, we only got yards away from the room before I clicked on it.

Comments on the US Trade War With Canada

I've been trying to push Cory Doctorow's agenda on several sites, pretty much 0 engagement. I'm not sure why I get so little net engagement - except for the AI training bots, who LUV this blog.

Too old? Too odd? Who knows. Well, I mostly write to populate my exocortex, which does not require external readers.

This was mostly from a couple of weeks ago. We just got back from a 2 week vacation, mostly in Canada.

Posted to Paul Krueman's substack:

July 1 Carney backed off on the digital services tax on Big Tech under threat of tariffs. That threat is now moot, Canada should move head with those taxes.

It should also listen to Canadian Cory Doctorow and revoke Canada's anti-circumvention law (Bill C-11, the Copyright Modernization Act of 2012, Canada's version of the DMCA section 1201). This was also passed under threat of tariffs, now moot.

Canada would become the hero of the world, including US consumers, allowing app stores that don't collect a 30% vig on all transactions, enabling jailbroken smartphones, John Deere tractors, auto analysis computers, and inkjet printers.

Posted to Dean Blundell's substack earlier:

Again: under threat of tariffs, Carney cancelled the digital services tax on US Big Tech companies July 1. Canada got tariffed anyway, the digital services tax should be revived.

And, do the right thing, follow Canadian Cory Doctorow's plan to rescind Canada's laws supporting the US DMCA provisions making it felony to bypass digital IP locks. Those laws were only passed because of the threat by the US of tariffs, now a moot point.

Let Canada become the country that jailbreaks all smart phones, creates open app stores that don't take a 30% vig of all commerce, jailbreaks John Deere tractors, jailbreaks auto repair diagnostics. They will be a hero to the entire world, including the US. & gawd knows, US Big Tech could sure use some comeuppance.

Posted to Dean Blundell's subStack chat:

2 things that were done because of the threat of tariffs, now moot since the US has imposed completely unjust tariffs on Canada, which should now be undone:

1) July 1 Carney backed off on the Digital Services Tax which would collect tax revenue from US Big Tech firms. This tax should be revived immediately.

2) Follow Canadian Cory Doctorow's plan to revoke Canada's anti-circumvention law (Bill C-11, the Copyright Modernization Act of 2012, Canada's version of DMCA section 1201). This was also passed under threat of tariffs, now moot.

Canada can become the jailbreaking center of the world, allowing app stores that don't collect a 30% vig on all transactions, enabling jailbroken smartphones, John Deere tractors, automobile analysis computers, and inkjet printers. This will benefit consumers worldwide, including in the US. Plus make $B for Canadian companies.

Doctorow's detailed prescription: Pluralistic: How Canada can help Americans and defeat America (23 Aug 2026) https://pluralistic.net/2026/08/24/elbows-really-up/

Despite Carney's doing some highly Elbows Up work, maybe Doctorow's ideas are still too radical to establishment politicians? Not sure why. Canada needs every advantage it can get.

Earthquake Weather

"Earthquake Weather", by Tim Powers, 1997, 744 pages, 202k words. Book 3 of the Fault Lines trilogy.

We get some new characters, & a lot of the characters of the 1st 2 books also make an appearance. As usual, Powers plays fast & loose with mysticism, ghosts, gods, you name it.

I enjoyed my reread of this series.

Watermarking, My Ass

1st from the New York Times:

This Is One of the Most Important Policy Decisions of Our Lifetime

These A.I. Policies Will Hurt Our Business. We Should Do Them Anyway.

Written by Bridgewater Investments, which has apparently invested $B in GenAI companies. This strikes me as yet more hype: AI is going to take over the world, so here's how we should tax it. Regardless of their motives, I'm going to say, token tax, hell yeah! Labor is taxed, LLMs should be also.

---

From the ever reliable Gary Marcus, more crappy hype from the King/Queen of Hype, Anthropic CEO, Dario Amodei :

No, Dario Amodei, we will not be curing cancer and “most human disease” in five to ten years

What Amodei either ignores or doesn’t know about medicine

---

Wow! Anthropic was surging ahead of OpenAI. So happy to see them self-immolate.

1st, from Doc Searls:

You can hear the squeak of sphincters closing

How Claude marks AI-generated content is the most self-destructive utterance I’ve ever heard coming out of an AI company. Gary Isaac Wolfe (who contributes a comment below): “Anthropic watermark scheme appears as a comic policeman blowing his whistle while his pants fall down and Mark Pauline loads a cannon with rotting fish.”

But maybe that’s a good thing. Let’s start with that police whistle:

When a supported Claude model generates text, it weaves an imperceptible watermark directly into the text itself. You won’t see it, and it doesn’t change the meaning, quality, or readability of Claude’s response.

Because the watermark is part of the text, it will travel with the text when it’s copied and pasted elsewhere, and may persist through some editing. Watermarking will be applied at the model level, which means it will be present no matter which Claude product or surface the text comes from.

Will any writer, of anything, ever want to use any Claude text anywhere now? I’ve been a satisfied Claude customer, and I’m ready to bail completely. And I’m not alone. Because watermarks in text are invisible time bombs ready to scream, “The author didn’t write this! I did!” when some reader of the writer’s text uses the right probe.

All the times I have read how writers go over & over their text to pick the 1 perfect word, the 1 perfect phrase. Clearly not a concern for Claude when it processes words as nothing but statistics, to be altered however is desired, as long as the statistics are still OK. Semiotic zombies, FTL.

---

404Media also weighs in on this unbelievably ... arrogant ... lame ... delusional??? idea of Anthropic:

Anthropic’s Text Watermarking Proves AI Companies Do Not Care at All About Writing

AI companies see words as interchangeable and have zero clue how to judge the “quality” of writing.

...

“Nothing is added to the text and there are no hidden characters,” Anthropic wrote in that company blog post. “The difference between watermarked and un-watermarked text will not be distinguishable to readers.” The way it will work, the post explained, is that Anthropic will subtly alter the word choices in AI-generated text in a way that is only known to Anthropic and its algorithms. Anthropic will know the watermarking algorithm, which will change “the source of the randomness used to pick among words” and thus can write a tool to detect whether something has been AI-generated.

So Anthropic is encoding secrets in every piece of text it produces for you??? This is semiotics on steroids, Claude is indeed a semiotic zombie. The arrogance of this is to me incomprehensible. Just say no.

---

Here's the home page/directory for my posts on Bullshit. This is post #142.

Expiration Date

"Expiration Date", by Tim Powers, 1996, 568 pages, 176k words. Book 2 of the Fault Lines series.

I was going to power through & binge this series after I reread the 1st book, "Last Call", for maybe the 3rd or 4th time. Technical difficulties at Kobo prevented that. Those difficulties finally got resolved a few weeks ago. So I got to reread this book.

Quite a bit of new mythos is added to the 1st volume of the series. Ghosts exist, & can be captured & smoked. Smoking ghosts is, of course, addictive, but it is apparently a real rush; gives you new memories & experiences; & apparently increases lifespan.

An 11 YO boy whose cultist parents were grooming him to be a Holy Man winds up ingesting the powerful ghost of Thomas Alva Edison. But as he is pre-pubescent, he cannot digest Edison. They wind up cohabiting the body.

All of LAs most powerful ghost addicts are totally lusting for the Edison ghost. Drama & plot ensue.

I'm 2/3 of the way through Book 3, all these are longer than I remember. But, still good reads, a good series.

The Dike Is Crumbling! Need More Fingers!

---

From 404Media:

The Roboguard Revolution is Short-Circuiting

Knightscope and other robotics companies are rethinking automated security following canceled contracts. One pivot? Human guards.

You'd think "security guard", which must mostly be an incredibly boring job, would be ripe for replacement by AI-powered robots, neh?

Nope, absolutely wrong. The outliers, the anomalies, these are what security guards have to handle.

& these are exactly what GenAI sucks at. Oops!

---

From Malwarebytes.com:

The AI Act kicks into action, forces companies to be clear about AI chatbots

The European Union (EU) has started enforcing key parts of the AI Act, with immediate, visible consequences for chatbots, deepfakes and other consumer‑facing Artificial Intelligence (AI) systems.

Hooray for the EU!

Our 20 YO 52" Sharp Aquos LCD TV died. I'm trying to replace it with a 65" LG OLED. lg.com says, "ChatBot not available." & connects me to a "human" sales agent.

If this is a human being, it needs psychiatric intervention ASAP. It is all over the place. It asks me 3x, after I say "no", if I want extended warranty. I ask it flat out:

Ben, you really seem to be a chatbot. Are you really human?

It's response:

Haha, nope! Just Ben here, your LG product expert. Though I'll take it as a compliment if I was answering that fast!

Really? Ben's a human being who is flattered to be compared to a fucking chatbot?!?!?

If this were a human being, it totally needs its meds checked.

Presumably with the new EU law chatbots are not allowed to lie like this. Goddamm, I hate chatbots! Does anybody like them?

---

From the New York Times:

A.I. Will Affect Your Social Life. Yes, Yours.

But in conversations with other Gen Z-ers, I’ve found that young people are relying emotionally on A.I. more than we realize. That subtly changes our relationships, even with people who have never touched A.I. When a flood of synthetic support enters the worlds of those around us, it distorts our mechanisms for intimacy and makes everyone a little lonelier.

---

From Robert Reich on SubStack:

Stop AI Before it's Too Late

Its costs and risks are gigantic. Its benefits, unproven.

Its benefit: to keep the bubble inflated, to maintain the outrageous P/E ratios of the Big Tech companies.

Of course, that's only a benefit of you are a TechBro or otherwise heavily invested in this bubble.

Its error rate: 9% for LLMs & chatbots, minimum 10-15% for agents. You CANNOT build infrastructure on that kind of error rate.

---

Again, from Malwarebytes.com:

Love/hate relationship: The AI affair. Young people love AI, but it’s breaking their trust

Young people use AI for everything. From schoolwork to interview prep, relationship advice to shopping decisions, the technology has become part of how young people live. For the most digitally fluent generation ever, AI is a competitive edge, a creative partner, and an always-on assistant.

But the same technology making young people’s lives easier is also making the internet harder to navigate. AI is making scams more convincing, identities easier to manipulate, and online content harder to trust. Seven in ten (70%) 18-to-22-year-olds have experienced an AI-related scam in the last year, compared to half of the general population. And nearly every young person worries AI will be used against them.

---

From 404Media:

Company Offering ‘100% Human-Written, Never AI’ Medical Research Is Entirely AI

Research Gold's team of human methodologists are either AI generated or using the identity of real people without their permission

Yawn. I need to compile a list of words that rhyme with "fraud".

---

Fuckety, fuckety, fuck the people who do this. I've made 100s of contributions to eBird. I've taken many pictures of green herons. This is the 1st green heron I've seen with 3 legs. ARRGGGHHHH!

Oops, I was going to include the pic of a green heron with 3 legs shown in this article. But, that would violate the terms of this blog, which specify that it is AI-free. Phew, dodged a bullet!

From eBird.org, the world's premier birding site, run by the Cornell Ornithology Lab & the Audubon Society:

From Truth to Fake: The Threat of AI to Media Archives

Community-driven science projects such as eBird, the Macaulay Library, and iNaturalist are seeing a rapid uptick in AI-manipulated and AI-generated photos. A new publication in Nature has recently highlighted the threat of AI photos to these projects, leading to some popular news articles. We want to raise awareness and inform the community of how AI reduces the scientific value of photos.

GenAI, FTL!!!

---

From Lauren’s data Substack:

Temperature Zero for Culture: Why Everything Is Starting to Look the Same

What 640 London shopping streets, 4,000 Billboard hits, a million films and synthetic AI personas tell us about why everything is flattening into sameness and how to regulate the algorithms behind it.

This is slightly afield, but still completely apropos. We were already experiencing stylistic & semantic ablation before ChatGPT was released - modern pop music was even then a maximum 1 or 2 levels above AI slop.

But AI slop will make it even worse.

---

From Yves Smith at naked capitalism:

Layoffs Tied to AI Hurt Worker Productivity – and the Reason May Surprise Managers

Yves here. BWAHAHA. The way employers are implementing AI, to reduce headcount, is scaring workers so much that they are loath to use AI. So this job-protective reaction amounts to sabotage without consciously being so. Well played!

---

Again, from 404Media:

AI Generated 3D Models Flood Market, But Almost No One Is Buying Them

The number of AI generated uploads to CGTrader would suggest AI is taking over the platform, but buyers are refusing to pay for AI generated models.

...

CGTrader, an online marketplace for 3D assets, found that one in six models uploaded to the site these days is AI generated, but that AI generated assets account for only $1 out of $90 in revenue on the site. These numbers show that AI generated assets are quickly flooding the marketplace, but that most people are not interested in paying for them.

Why aren't people buying them? Because they aren't as good as the human-created models. Close, but no cigar.

Meanwhile, they create more slop that people have to wade through. I mentioned that a few posts back, we all now have to waste increasingly more of our limited brainpower to filter out AI slop.

Maybe we could get this classified as a form of spam? There are laws against spam.

---

From a software developer of 2 decades experience:

I'm done using AI

I've been programming for 20 years. It started as a hobby as a teen, making fansites and little video games. Then it turned into a career. And it's something I've largely enjoyed doing. I like learning, I like problem solving, and I like making things, which is a lot of what programming is. Over the past 18 months, my career, my hobby, and my craft have been drastically changed by the introduction of LLM-powered coding harnesses (e.g., Claude Code) and what's often referred to as "agentic programming." Because of its negative impacts on the environment, society, and my mental health, I've decided to stop using it.

So his management encouraged him to use Claude Code, etc. He got it to where it mostly worked. He did it for 1 year.

But, it killed his spirit, his joy of creation, his zen. He rejected it.

I'd guess he'll be let go soon :-(

---

From Human Ready, the SubStack of John Kwarsick. John's company I think is a "AI Works, You're Just Not Doing It Right" company. They have a whole package to deal with the idiocy of making AutoComplete-On-Steriods your AI architecture.

How to Catch AI Lying Without Knowing It Is Lying | Thought Engineering

The discipline: The Over-claim Watch / K-I-E Tagging

So the K-I-E stands for Known, Inferred, Extrapolated. So, OK, maybe OK, bullshit. Great, you've successfully wasted a bunch of your life concluding that this particular LLM output was ... bullshit. Or possibly bullshit. Now what?

I can't imagine any other state change in the world of technology that would have succeeded with the level of error that LLMs & GenAI exhibit. I mean, come on, we all use autocomplete on our phones every day. (I periodically get so sick of it that I turn it off, but I have it on now.) How often is it wrong? Horribly, badly, wrong? & you are going to fire people & replace it with this?!?!? Barf-o-rama!

---

Here's the home page/directory for my posts on Bullshit. This is post #141.

AutoTune for Writing

1st up, the always reliable Gary Marcus:

Two critical updates re: Astra and mathematics

So OpenAI announces a new model has solved 10 hard math problems. Then within a day an Antropic researcher (Alpöge) is able to reproduce 1/2 the results using an older public model.

Alpöge’s fast partial replication made me wonder whether OpenAI’s real advance here was finding (through AI) which open problems were amenable to a certain kind of search and verify technique.

So now in addition to continuously trying to 1-up each other with miraculous claims, the 2 leading AI companies are now starting to play defense as well. Yawn.

---

From Doc Searls' Weblog, we get the title of my post, which will go in my table of names for LLMs.

Throbsday - On AI wronging and writing

I hate AI writing and can spot it in a second. It’s not just the excess of one-line paragraphs, the contrasty “not that, but this” argument constructions, the excessive use of em dashes. It’s a tone that sounds like the average of a million well-put hunks of text published by competent writers of opinion columns and advertising copy. In other words, a lot like my own writing.

AI writing is AutoTune for the human kind.

Doc Searls has written many times about his use of GenAI, so I was kind of surprised at the negative attitude here. But he makes excellent points: it can be a tool, but please no slop.

Note, lots of musicians totally hate AutoTune.

---

From the King of Snark, sci-fi author John Scalzi. He joins the group of authors becoming ever more outspoken about AI slop. Hunh, I guess this goes in the table of names as well.

Generative “AI”: The Guitar Hero of Creativity

Yah, you can generate stuff that almost seems decent. But you aren't developing any real skill or, I think more tellingly, any real style.

I remember when my kids would play Guitar Hero & I'm totally like "Why not actually learn to play guitar?" (I think all of my kids do play instruments.)

---

From Programmable Mutter by Henry Farrell:

The downside of robot solutionism

Mapping the Abelard Snazz feedback loop

Apparently a comic series by Alan Moore in the early 1980s is eerily prescient re the many arms races that seem to define GenAI - using GenAI to detect GenAI slop, cheating, plagiarism, etc. Spoiler alert - the human race winds up being forced to flee the planet - oops!

---

Here's the home page/directory for my posts on Bullshit. This is post #140.