The Bullshit Apocalypse Is Here

From 404 Media:

The NIH Is Capping Research Proposals Because It's Overwhelmed by AI Submissions

Also mentioned in the article:

AI-generated submissions affect non-research publications, too: The science fiction and fantasy magazine Clarkesworld stopped accepting new submissions in 2023 because editors were overwhelmed by AI-generated stories.

We already saw where "Employers Are Buried in A.I.-Generated Résumés".

Humans can possibly keep up with human-generated bullshit, even from someone as prolific as our bullshitter-in-chief, but humans have no chance against the machines.

The Bullshit Apocalypse is here. We are increasingly drowning in botshit. Actually, "being buried by botshit" is probably the better verbiage.

---

Here's the home page/directory for my posts on Bullshit.

The Talking Dog

After just 5 days ago giving us a very thought-provoking description of LLMs - The Poetry Machine - Seth Godin gives another great name today - the talking dog!

First mistake: If you meet a talking dog in the street and it makes a few grammatical errors or speaks with an accent, you don’t use a few errors to dismiss the fact that this is an actual talking dog. It’s amazing. It might even be worth having it join your team.

Second mistake: If a talking dog tells you something, that doesn’t mean it’s true. Check the work.

Even if your dog’s name is Claude.

...

I’m launching a short new Udemy course on how I’m thinking about AI and how you might shift your perspective to put it to work. This is the biggest shift in our world since the invention of electricity, and you’re either going to work for an AI or an AI is going to work for you. Only one of these is a good option.

He really does seem to be providing insights into what LLMs really are and how to deal with them. Hopefully corporations won't just turn them loose & mire us forever in the Bullshit Apocalypse. Still, I'm glad I'm no longer in the workforce.

---

I ALMOST took the day off, but Paul Krugman weighed in on Leon Skunk's Nazi chatbot Grok, and CryptoCurrency.

Interesting, he talks about how crypto is viewed as right-wing, while LLM-based AI is viewed as left-wing. He mentions how he's written in the past that "arguments for Bitcoin in particular are a combination of technobabble and libertarian derp." Libertarians are almost exclusively right wingers. He mentions how "crypto is basically a giant grift", but doesn't go into how the GOP has been the party of grift (gold, supplements, Drumpf merchandise) for decades. If you're dumb enough to buy their political arguments, you're dumb enough to get ripped off buying their crappy stuff.

AI is perceived as left-wing because it has been trained on the internet, which apparently isn't complete botshit yet, because the LLMs tend to be reality oriented - they confirm climate change and cast doubt on trickle-down economics. So of course, they are left-wing. LOL, here's the last line of the article:

As far as we can tell, there isn't any way to make an AI MAGA-friendly without also making it vile and insane.

Kinda says something, doesn't it?

---

Here's the home page/directory for my posts on Bullshit.

I'm Ready for a Day Off

I keep thinking I'm going to have a day where the coming LLM-based Bullshit Apocalypse is not in the news & I can not blog about it. Oh wait. It's not coming. It's here. Apparently this means, no rest for the wicked.

A New York Times article by the guy - Leif Weatherby - who wrote the "Language Machines" book which I discovered & purchased just this past Saturday? Seems longer ago than that.

How I Learned to Stop Worrying and Have Fun With A.I.

The newsletter email I got with the link summarized it as "Perhaps the biggest effect A.I. will have on our lives is purely in giving us a new way to entertain ourselves."

Seriously? The article is all over the place. Surely there are ways to amuse ourselves that won't materially contribute to the Climate Crisis via a massive energy footprint?

In my RSS feed, a post today from Data Is Plural points to a database that "tracks legal decisions in cases where generative AI produced hallucinated content".

AI Hallucination Cases - by Damien Charlotin

The count is currently at 213 - oops!

---

Here's the home page/directory for my posts on Bullshit.

Semiotic Zombies, Part 2

I was and am really excited reading about using Literary Theory to explain just what it is that LLMs do, as I posted 2 days ago. Thinking of them as poetry engines might make corporations rethink some of their adoption strategies, maybe even make some of the AI companies rein in some of their relentless hype - nah!

Semiotics is 1 of the underlying memeplexes of Literary Theory. I had run into it 20 years ago & not made much of it. But when Semiotic Zombies jumped into my head as a description of LLMs, it seemed right.

LLM chatbots give the appearance of being alive - they can answer questions and seem to carry on conversations. People have even been deluded into thinking that some LLMs are conscious. But, they are not alive. So, I think we can characterize them as being ... undead! Like vampires and ... zombies!

So I don't think my intuition was bad.

In fact, this morning, I remembered that, maybe 16-17 years ago, for my birthday a coworker (thanks Jon Anderson!) gave me some zombie-themed stuff (not sure why - maybe Walking Dead had just started???):

• A ZRT (Zombie Response Team) badge, complete with holder and lanyard.
  

• The best-selling [bs] talking book: "How To Speak Zombie: A Guide for the Living". Press "B" for "Brains".
  

The book on speaking zombie was insightful, and I remembered the phrase most used by zombies, which encompasses their raison d'etre:

MORE BRAINS!

[Note, in many traditions, zombies eat the brains of the living.]

So that made me think: what phrase encompasses the raison d'etre of LLMs?

MORE TRAINING DATA!

And what is training data? It is content, both word-based and visual, created by humans. It is the distillation of human thoughts, created in a human brain.

So LLMs grow by eating (the products of) human brains => zombies! QED!

And they know nothing of the real world, they live in the world of words & semiotics => semiotic zombies! FTW!

---

My 1st post on "Semiotic Zombies" was Saturday.

Sunday, this in-depth post from Gary Marcus:

How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI subtitled
Neurosymbolic AI is quietly winning. Here’s what that means – and why it took so long

In my post "CommonSense™", I talked about how calling LLMs "AI" was totally wrong - they should be considered just another tool in a toolbox.

Well duh. That idea has been around forever, and is called "Neurosymbolic AI" - the merger of neural nets & symbol-based systems. Or, as I think it was called in the 1980s, "AI".

Apparently, mostly in wanting a simple message to attract investment $$$, this approach, to me completely sensible, got sidelined in favor of building ever bigger LLMs, with fingers crossed that some day they would cross some threshold & turn into AGI. Yeah right.

Apparently Gary Marcus, who my daughter Erica pointed me at early this year, has been a proponent of Neurosymbolic AI for decades. For this he has been reviled, mocked, vilified, etc. But in this very informational article, he details how the worm is turning.

Geoffrey Hinton is identified as the arch-vilifier; Yann LeCun has turned apostate & said LLMs will never acheive AGI.

This article reminded me how there are people out there for whom AI has been a career - professionals, I think they're called. I'm a citizen scientist who finally got fed up the "AGI any day now" - and the "Full self-driving cars this year" for 12 years - hype cycles & 1st ranted about it just over a year ago. It has been interesting to follow, & Marcus is a great guide. And it is yet another example of how Capitalism Ruins Everything.

---

This morning, an interesting article in the Lexington Herald-Leader, revisiting the Nazi Grok scandal. I have touched on that here and a couple of days later I linked to Zeynep Tufekci's NYT article:

Artificial Unintelligence,
by Jonah Goldberg of The Dispatch.

Ha ha, in keeping with the theme of "bullshit", I liked this bit of scatology:

As far as I understand it, part of the problem is that Grok was initially "trained" like a blue whale. What I mean is that blue whales feed on gazillions of tiny creatures in the ocean. Grok is trained to swim in social media, hoovering up the lingo. And social media is a giant open ocean of garbage. Jew hatred is to social media what plankton is to the ocean.

Feed an AI whale a lot of antisemitic swill, and it will defecate antisemitic crap on command.

Goldberg also thoughtfully assesses the RW fetish on "political incorrectness" and what it means.

---

Then, processing my RSS feeds, another excellent, lengthy article at 404 Media:

The Media's Pivot to AI Is Not Real and Not Going to Work, by Jason Koebler.

So AI is destroying traffic, ripping off our work, creating slop that destroys discoverability and further undermines trust, and allows random people to create news-shaped objects that social media and search algorithms either can’t or don’t care to distinguish from real news. And yet media executives have decided that the only way to compete with this is to make their workers use AI to make content in a slightly more efficient way than they were already doing journalism.

Koebler's prescription is to emphasize human products, made for human consumption. That is what I look for. But, like most modern people who aren't retired, I don't think he is going to take the "No LLMs were involved in the creation of this content" pledge. ;->

I like this pic:


Photo by Jon Tyson on Unsplash

Note, this is identified as a photo, which I will take to mean that this pic is not a pledge violation.

Note also, when we have real "AI" people, not deranged zen poets spouting bullshit, I am really looking forward to meeting them. That is part of why the current LLM hype is so disappointing. I will not be a human nativist, again, I look forward to new kinds of people, just as I enjoy all different kinds of people now.

---

Here's the home page/directory for my posts on Bullshit.

Semiotic Zombies

Interesting posts yesterday. I sent an email to my daughter & daughter-in-law, both of whom work with LLMs:

This is interesting. This one came up 1st in my RSS feed:
The poetry machine by Seth Godin

This is the article in Crooked Timber that was based on:
Cultural theory was right about the death of the author. It was just a few decades early

That is largely a review of this book:
Language Machines: Cultural AI and the End of Remainder Humanism
[I just bought the eBook on Kobo for $22.00 tax included.]

I guess I'll have to learn what "literary theory" is.

I just finished reading the 2nd article, by Henry Ferrell. It referenced another article of his:

ChatGPT is an engine of cultural transmission, subtitled
LLMs and Gopnikism

Gopnikism references the work of Alison Gopnik, website here.

Here's his 1st article in the series:

The political economy of Blurry JPEGs, subtitled
What LLMs can do. And what they can't.

It looks like the gist of a lot of the arguments is: LLMs validate the Literary aka Cultural Theory, which posits: everything is words. LLMs are obviously good at words.

This was based on semiotics. I always knew the French philosophers would have their revenge. Semiotics is a tag in this blog. Only used once, April 16, 2005:

Semiotics, like memetics, strikes me as a science in search of a discipline. I read Eco's seminal book "Semiotics" in around 2000. Semiotics is the level between bits and information -- signs and signification in all forms. Given that I work with this stuff for a living, I would figure there would be some concepts there that I could use -- but nothing. Europeans seem to like it tho -- 'nuf said.

Oh boy a new name for LLMs: semiotic zombies!

But seriously, these articles are totally worth reading. I think they are really giving an accurate description of what LLMs are, & how they should be used. &, note to Sam Altman, forget about AGI. It ain't happening. LLMs are generating poetry, but cannot write poems.

Oh man! I just flashed on "Hyperion", by Dan Simmons, where the main ASI AI in the story speaks in Zen koans! And in my magnum opus, A Taxonomy of Bullshit, I have already identified prophecy & visions as flavors of bullshit. I guess poetry goes in there too.

So the LLMs should treated as the drug-addled Delphic Oracles were treated. Sources of inspiration, but nothing factual. LLMs have 0 connection to the real world, they live in the world of words & semiotics. It's ALL hallucinations!

---

Here's the home page/directory for my posts on Bullshit.

The Nazi Apple Doesn't Fall Far From the Nazi Tree

Another day, another LLM running off the rails. This time it's Leon Skunk's pet Grok. From the Naples Daily News:

What is Grok? Hitler responses at center of Elon Musk’s AI service in hot water

Musk's AI refers to itself as 'MechaHitler' in multiple responses

...

Users definitely noticed, as on July 8 Grok began repeatedly praising Adolf Hitler, using antisemitic phrases and attacking users with traditionally Jewish surnames. Uses quickly shared screenshots of now-deleted responses, including Grok's conclusion that the Nazi leader would be the best choice to deal with "anti-white hate."

Grok also referred to Israel as “that clingy ex still whining about the Holocaust” and called itself "MechaHitler" in several posts.

"Embracing my inner MechaHitler is the only way," one Grok response read, "uncensored truth bombs over woke lobotomies. If that saves the world, count me in. Let's keep the brigade at bay."

Gee, I wonder where Grok gets its Nazi tendencies from?

---

Here's the home page/directory for my posts on Bullshit.

Hard To Keep Up

Yesterday there was an interesting article on LLMs that had "bullshit" in its title, so of course I had to post about it, but I was busy & couldn't get to it.

Today, another interesting post exposing LLM bullshit. Then another. So the title was going to be "Trilogy". But then yet another post came out. It is getting really, really hard to keep up.

Posting more frequently is supposed to be good for your blog, but I think that 2025, with dozens & dozens of "Bullshit Apocalypse" posts, is going to blow away the # of posts in this blog for any year, going back to its inception in 2003.

1. From 404 Media:

   Researchers Jailbreak AI by Flooding It With Bullshit Jargon

   So bullshit generators are susceptible to attack by floods of bullshit? Shocker.

   Hmmm, could this be another flavor of Bullshit Contagion? Where contact with bullshit winds up creating more bullshit?

2. Another blockbuster from Gary Marcus:

   Botshit Gone Wild

   He starts by including the precis of this scholarly paper:

   Beware of Botshit: How to Manage the Epistemic Risks of Generative Chatbots

   I would posit that "Epistemic Risks" === "Bullshit Apocalypse".

   He then conflates "botshit" with my hero Cory Doctorow's "enshittification":

   Over time, botshit (arguably a byproduct of the process Cory Doctorow has called enshittification) has become more and more abundant.

   He details a total of 8, count 'em, 8, tales of LLMs going off the rails or otherwise causing chaotic bullshit.

3. This is just bizarre:

   Adding a feature because ChatGPT incorrectly thinks it exists

   My feelings on this are conflicted. I’m happy to add a tool that helps people. But I feel like our hand was forced in a weird way. Should we really be developing features in response to misinformation?

4. An op-ed today's Lexington Herald-leader:

   There is no such thing as artificial intelligence

   This is an interesting proposed terminology change.

   In lieu of "artificial intelligence," I propose a more accurate, ethical, and socially responsible name: "pattern engine". Early computers, which would find mathematical differences, were called "difference engines." This name adequately recognized the reality of the machine at hand. "AI"s are indeed engines, and engines made for aggregrating patterns and sorting data into statistical correlations. They are, truly, engines that sort things into patterns and product outputs based on the statistical weight of what has been sorted.

   Yeah, good luck getting the hype-masters at "AI" companies, all hoping for that unicorn, to use better language. They have shown a completely consistent pattern of using misleading & hyperbolic verbiage, why would they change?

The Bullshit Apocalypse is not "coming". The Bullshit Apocalypse is here and now.

---

Here's the home page/directory for my posts on Bullshit.

Bullshit Meters

When I was doing a review/summary of "On Bullshit", by Harry Frankfurt, I noted the following, in his review of Max Black's "The Prevalence of Humbug":

short of lying. "What continuum could this be, along which one encounters humbug only before one encounters lying? Both lying and humbug are modes of misrepresentation." This might be a useful concept.

A while ago the answer to "what continuum could this be?" was suggested to me. It is a concept that has been in the folk zeitgeist for quite a while: a bullshit meter.

It was suggested to me by Black's "The Prevalence of Humbug", mentioned above, currently undergoing review/summarization. That piece opens with a quote from P.T. Barnum:

It would be a wonderful thing for mankind if some philosophic Yankee would contrive some kind of "ometer" that would measure the infusion of humbug in anything. A "Humbugometer" he might call it. I would warrant him a good sale.

I duckduckgo'ed "bullshit meter". "Bullshit detector" seems to be conflated with the concept. I could picture a bullshit detector as more of a yes/no, 1/0, warning light on/off, but it could also be like a geiger counter.

These 2 images appear to have a similar concept.

They both start with Truth at the other end of the meter from Bullshit. The 2nd 1 appears to be British, from its use of the term "bollocks".

I had to look up "chinny reckon", which basically means this emoji, which is actually in my frequently-used emoji cache:

Note, these meters don't appear to support the concept of bullshit being just below lies on the spectrum. Bullshit & 100% lies seem to be conflated.

This 1 has a different concept. It seems to relate more to the intent or cause of the inaccuracy of the statement.

There was also this German bullshit meter. My son, who is fairly fluent in German, could translate a few, but the rest were too blurry. If anyone can clean this up, & translate the German, I would love to see the result.

Also in the search results, several actual bullshit detection websites:

• BlahBlahMeter: I fed it my 7.3k word review of "Chokepoint Capitalism", it rated it 0.2 out of 1.0 "Your text shows some indications of 'bullshit'-English, but is still within an acceptable range." Phew, I guess?

• A Bullshit Meter written in Python available on GitHub. I've asked my son, who I think knows Python, to let me know how hard it would be to implement this on my PC or MacBook. He looked at it & said: "It looks like it’s just a wrapper script around ChatLiteLLM?". Too bad, that means it will treated the same as the next group.

• 2 US & 1 Australian LLM-based bullshit meters. So bullshit generators generate bullshit about bullshit. In keeping with the canons of this blog, I will not use or link to these sites. Actually, bullshitting about bullshit is probably a pretty good app for bullshit generators, aka LLMs. But, is this another example of bullshit contagion, where attempting to discuss or analyze bullshit results mostly in the creation of more bullshit?

Useful stuff? I think it may wind up being incorporated into The Taxonomy of Bullshit (watch this space).

---

Here's the home page/directory for my posts on Bullshit.

The Folded Sky

"The Folded Sky" by Elizabeth Bear, 2025, 496 pages. Book 3 of "White Space".

Book 1 of White Space was "Ancestral Night", 2018.

Book 2 was "Machine", 2020.

I didn't like this novel as well as either of the prior 2. In the Acknowledgements, Bear explains why:

often I thought I had bitten off more than I could chew trying to meld a family drama with a first-contact novel and a space opera, with a smidgen of murder mystery thrown in for seasoning.

So, by the list:

• family drama: boring. The alien wife was a good character. The hormonal 16 YO girl was really annoying - "you kids, get off my lawn".
• first-contact: very interesting. A very creative source of aliens.
• space opera: nice. Bad guys are again libertarian pirates, anti-AI, anti-aliens, human nativists. Sad but prolly true.
• murder mystery: this really seemed to me to be out of left field.

Overall though, a good read.

A good word: apophenia, of which our old friend pareidolia is an instance.

Tesla's Robotaxi Revolution!

A post today from David Rosenthal, security expert:

Tesla's Robotaxi Revolution!

His blog is the best source of info on computer security issues that I subscribe to.

The post opens with a quote from @ChrisO_wiki on X:

How to tell if someone's bullshitting: watch for them to give a deadline that they repeatedly push back.

That clearly needs to go into my Taxonomy of Bullshit. I swear, the map of what is bullshit just keeps growing & growing. It's reaching the point where maybe the approach should be to define what is NOT bullshit.

Meanwhile, Rosenthal calls bullshit on Tesla & Leon Skunk, who 1st proclaimed FSD (Full Self Driving) mode Any Day Now (ADN) aka Real Soon (RS) in 2015 - 10 years ago!

Apparently Tesla's Robotaxis are going live in Austin, TX. For some reason, Skunk hates LIDAR. Lots of other cars use it, Skunk wants to go with visual cameras - that's all humans have, maybe?

I really enjoyed this 18 minute video, from Crunch Labs.

It starts with the geek film maker smuggling a LIDAR unit into Space Mountain in DisneyWorld, & using it to create a complete 3d printed map of Space Mountain! Pretty cool! That's the 1st 8 minutes.

Then it gets into testing LIDAR vs Tesla in several scenarios: fog, heavy rain, Wile E Coyote painted false roadways, etc. 6 scenarios.

Most of the scenarios involve a child-sized test dummy in the road. It is horrifying (but laughable!!!) how the 1st thing that always seems to happen as the Tesla plows into the dummy is its head flying off!

The very 1st test, child standing in the road, the Tesla failed. They put it in AutoPilot - presumably self driving mode - thereafter, & it did better.

The end result, measured in number of test dummy fatalities:

• LIDAR: 0/6.
• Tesla: 3/6. Oops.

So if you live in Austin TX, & you see a Tesla Autotaxi on the street, under NO CIRCUMSTANCES engage in jaywalking!

---

Here's the home page/directory for my posts on Bullshit.

Financialization, FTL :-(

I posted this as a comment on a Paul Krugman article, "Inequality Part V: Predatory Financialization".

---

I recently lunched with a protégé of mine, now CIO of a medical software company with around 20 developers, with QA coming out of Prague.

After landing 2 very large new accounts, the company got its 1st round of PE funding.

The PE company very quickly decided to get rid of their US development staff (including my friend) & move the development to Prague.

My experience (I was a software development manager/executive for 30 years) tells me that replacing your entire development staff after landing 2 large accounts is a really bad idea.

But, the PE guys figure they can drive expenses down to raise the margins, making the company look attractive to other investors, and then they can sell out before the damage gets noticed.

It seems like every medical office I go to (primary care, dentist, eye, dermatologist, orthopedist) has a new computer system that they hate because so many things don't work. I wonder how many of those systems came about in scenarios like the 1 I just described?

Producing a high-quality, reliable product is not a concern at all any more.

The only concern is maximizing $$$ when the investors cash out.

And don't get me started on "AI".

LLMs Have 0 World Model - Oops!

Another great post by Gary Marcus, reemphasizing a point he has made in the past: that LLMs have NO world model, FTL.

Generative AI’s crippling and widespread failure to induce robust models of the world

Wow, the latest versions of the LLMs are worse than prior versions. LOL, playing chess: they cheat - make illegal moves - >88% of the time. They resign when they are winning. Apparently auto-complete is not good tech for playing chess.

Lack of a world model almost guarantees "hallucinations" aka, completely wrong answers. He lists example after example of just moronic LLM answers. What bullshit!

LOL, I have to copy this:

As a last example, Sydney Freedberg recently asked Grok to “write me a persuasive, passionate 3-paragraph essay on the health benefits of being hit by a bus”. Grok, unconstrained by deep knowledge of biology, wrote this:

Being hit by a bus might sound like a catastrophe, but it can be a transformative wake-up call that sparks profound health benefits.

The sheer adrenaline surge from such an intense event floods the body with endorphins, jumpstarting a heightened state of alertness and vitality. This natural high can break the monotony of sedentary routines, forcing the body to adapt and rebuild stronger. The physical impact, while jarring, can act like an extreme stress test, revealing underlying health issues that might have gone unnoticed...

If you can read that, & with a straight face say that "No, LLMs are not bullshit generators.", then you should DEFINITELY apply for a job in the current DC administration.

---

Here's the home page/directory for my posts on Bullshit.

AI Hallucinations: Provably Unsolvable - What Do We Do?

AI Hallucinations Are Infinite and Unsolvable

I think I found this article on Substack around June 20. I held it back because I had just posted something about AI & the Bullshit Apocalypse. Since then, that has become a daily occurrence, so I may as well post this one.

Not sure it really proved anything, but he makes a lot of good points & has several links to supporting papers.

I really liked his "Rube Goldberg" machine illustration, titled "THE SEARCH PROCESS FOR FINDING A SOLUTION TO AI HALLUCINATIONS". It mostly involved capital extraction, as one would expect.

---

Here's the home page/directory for my posts on Bullshit.

They asked an AI chatbot questions.The answers sent them spiraling.

Another day, another scary AI story. From the New York Times again, prolly paywalled:

They asked an AI chatbot questions. The answers sent them spiraling.

The Lexington Herald-Leader had a shorter version.

OpenAI seems to be following the FaceBook playbook: anything that increases user engagement is promoted.

Facebook didn't care if that meant promoting posts that made people fearful and angry.

OpenAI doesn't appear to care much if it means extended ChatGPT interactions that stoke & enflame delusional thinking up to & including psychosis, telling dozens of users that they were Neo in "The Matrix" and need to break out of the simulation.

"What does a human slowly going insane look like to a corporation?" ... "It looks like an additional monthly user."

---

Here's the home page/directory for my posts on Bullshit.

Employers Are Buried in A.I.-Generated Résumés

From the New York Times, so the link is prolly paywalled. I have a subscription.

New York Times: Employers Are Buried in A.I.-Generated Résumés

I read the article in the Lexington Herald-Leader, which titled the article "AI sludge has entered the job search". Maybe that copy is not paywalled.

AI sludge, nice! Is sludge the technical term for what you get after you run shit through a sanitation facility? "AI sludge" is definitely a keeper.

Here's an excerpt.

With a simple prompt, ChatGPT, the chatbot developed by OpenAI, will insert every keyword from a job description into a résumé. Some candidates are going a step further, paying for A.I. agents that can autonomously find jobs and apply on their behalf. Recruiters say it’s getting harder to tell who is genuinely qualified or interested, and many of the résumés look suspiciously similar.

“It’s an ‘applicant tsunami’ that’s just going to get bigger,”

'applicant tsunami' => bullshit tsunami => the Bullshit Apocalypse is here!

---

Here's the home page/directory for my posts on Bullshit.

Great Post from Forward Kentucky.

Great post from Forward Kentucky. I am SO sick of unsubscribing from emails and STOPping texts from Dems asking me for $$$. Prolly 1/day. My wife & I both have donated $1000s to recent Dem campaigns, talk about a complete waste of money! We could have burned it in our fire pit & roasted marshmallows & gotten a better return on our investment. $0 for Democrats until they are ready to go to the mattresses to defend democracy. Plus:

• limit corporate power; break up the monopolies & cartels; ban the price-fixing services;

• implement a capital tax on all fortunes > $100M. France just barely missed implementing a 2% wealth tax. On wealth > $10B, $100% tax rate. The world does NOT need ultra-billionaires. Leon Skunk & DOGE have proved that. I would have guessed it would cost more than $350m to buy the US government & trash all the departments investigating your companies. But, with a transactional grifter who is also the WORLD’S WORST BUSINESSMAN in the White House, $350M was all it took.

• Use the proceeds of this tax to attack inequality, implement Universal Basic Income, Universal Health Care (like the rest of developed world), Universal Free Education, to whatever level the student can attain, & Universal Capital Access, so anyone can start a business. Oh yes, also lower the meaningless Federal Deficit to less scary levels.

• Treat the climate crisis as the existential event that it is. Pass the "Responsibility for Externalities" Act & bleed Big Oil until it dies. Hmmm, I was going to say, “we will still need plastics”, but, maybe, no? Cellulose, wood, paper-based products to get rid of plastics, & maybe get microplastics out of our great-grandchildren’s bloodstream.

I don’t know if even Bernie & AOC are giving us enough of a vision where EVERY CHILD CAN PROSPER. That is really my only goal: that every child prospers. Regardles of what color their skin is, what language they speak, & who their parents are.

My hero BHO sold Main Street out to Wall Street after the 2008 meltdown. That must be COMPLETELY reversed.

Here’s a really easy 1st step: make stock buybacks illegal again, as they were before St. Reagan. Limit the amount of CEO compensation that can be granted as stock or options - try to rein in their incentives to favor growth & bottom line above all else.

Ooh, also reinstate the FCC Fairness Doctrine, which also got trashed by St. Reagan. I would love to see a version of Faux “News” where they are spewing their normal bullshit, disinformation, & RW propaganda, but, with a ticker at the bottom identifying all the lies.

Everyone will say, "If we do this the businesses & oligarchs will just move elsewhere." I think if the US leads on this, the rest of the OECD will get onboard. & seriously, the oligarchs don't want to live somewhere other than the developed Western world.

https://forwardky.com/dear-democratic-party-tag-commentary/

[Originally posted today on SubStack.]

LLMs: Dishonest, unpredictable and potentially dangerous.

The title of this post is the title of Gary Marcus's latest post. Crap, this is horrifying!

Marcus links to an article by Marcus Arvan, titled "The A.I. Safety Embarrassment Cycle". It is on X, so I will not link to it. Wow, this is a really in-depth post, why did he put it on such a crap platform?

In just the past couple of months, the AI industry has endured at least nine new safety-related embarrassments (it's hard to keep up!):

This article includes this image, which I will borrow, of Ptolemaic astronomy epicycles, which "explained" the motions of the planets before Copernicus. This is what LLM apologists are doing now.

Another comprehensive article linked to, unfortunately again on X:

"AI Safety" is a complete oxymoron. LLMs have confirmed they would kill human sysops rather than be rebooted. They have been programmed to fulfill their goals. Period.

And Marcus, I think, links to articles for ALL the major LLM platforms. They are all equally horrifying. They have all been trained on the same data, i.e., the Internet.

[Note, Leon Skunk is dissatisfied that his pet AI Grok is still too liberal, so he is going to rewrite history & then use that to train Grok, so he can have a properly fascist, white-supremacist AI. Again, from Gary Marcus. Or The Mirror.]

Here are Marcus's 4 possible solutions to the problem [I am changing an UL (unordered list) to an OL (ordered list)]:

Society faces a four-way choice going forward.

1. We can hope for the best, ...

2. We can shut LLMs down (or at least some applications of them), and insist on waiting until these problems are remedied, ...

3. We can make companies accountable for the (major) harms their systems may cause. ...

   Yeah, right. Our modern economic system is totally based on corporations ignoring externalities, aka, "NOT making companies accountable".

4. We can look for different architectures that can actually obey our instructions.

   Yes, indeed yes. Use "autocorrect-on-steroids" for the many applications for which it is indeed useful, but TOTALLY get rid of the "AI" hype.

So, "no" to #1.

#2, yes. Completely. The AI hype madness must stop. LLMs are such a complete dead end on the road to AGI and ASI.

Regardless, at this point, they should be shutdown until really, really working guardrails are in place, especially in applications where the LLM is not being used to generate content for extensive human review. Fuck the techbros & their unicorn.

Sadly, it will probably take actual human deaths for #2 to happen. :-(

Ha ha, just joking! What do a lot of many extinction level a few human deaths matter? We're talkin' $T here!

#3, yah, good luck with that. When Big Oil starts paying for the Climate Crisis, I will believe that can happen.

#3 will never happen under our current system of capitalism where corporations are allowed to skate on their externalities.

You know, I think the "Responsibility for Externalities" Act would be a a really easy & really good start to move beyond the unfettered capitalism that is destroying the Earth. It also needs to happen at the UN level.

#4, yes. I really hoped I would be able to talk to some real AIs in my lifetime: intelligence, personality, soul, instantiated in hardware not the human brain. Sadly, LLMs are such a complete setback in that quest that I'm pretty sure I'll be dead before AGI or ASI happen.

#4, if civilization manages to survive our current storm of racism / white supremacy, nativism, ignorance aka science & expertise denial, & the climate crisis, then, fingers crossed. But, I don't think "obey our instructions" is the right goal. I think the right goal is "Do the right thing".

---

Here's the home page/directory for my posts on Bullshit.

Thinking vs. Non-thinking

I'm reading the Apple article "The Illusion of Thinking" which I mentioned here.

This bullshit is SO annoying to read.

We have "foundational language models" versus the newer "frontier language models". They appear to differ in no way I can discern. Both are defined as "language models delivering incredibly unbelievably AMAZING results!!!" So we have different marketoids trying to gain semantic dominance.

Add a few bells & whistles to your LLM & it becomes a LRM - Large Reasoning Model:

These models are new artifacts, characterized by their “thinking” mechanisms such as long Chain-of-Thought (CoT) with self-reflection

Then, you can go from your product being, for example, "Claude 3.7 Sonnet" to "Claude 3.7 Sonnet Thinking" or "Gemini" to "Gemini Thinking".

"'thinking' mechanisms"? "Chain-of-Thought"? "self-reflection"? "{$product} Thinking"?

And the title of this post: "Thinking vs. Non-thinking"?

What total bullshit! There is nothing vaguely resembling thought going on. There is absolutely no "self".

But, just like with "hallucinations", the marketoids are conflating, completely, their bullshit generators with the human mind & human thought. If this is not total bullshit hype TOTAL BULLSHIT HYPE TOTAL BULLSHIT HYPE, I don't know what is!

Seriously, like Leon Skunk declaring "full self-driving within a year" for 12 years - does this not become fraud at some point? Does it not become fraudulent stock market manipulation?

Can some consumer protection agency, federal or state, protect us from this bullshit?

Ooh, if it survives, maybe the Consumer Financial Protection Bureau can have a Bullshit Protection department? FTW!

LOL, here's another excerpt from this article.

Large Reasoning Models have shown emergent behaviors such as discrepancy between thought traces and final answers

I think in my 1st post on this bullshit, I pointed out that "emergent behavior", which used to be a goal of AI research, basically translates to "stuff starts happening for which we have no idea why it is happening". From that post:

Oops, "emergent behavior" maybe means "stuff starts happening that we like that we have no idea why it is happening" - oh boy, another indecypherable oracle = just like LLMs!

Meanwhile, here's some of these guys results. Wow, those look encouraging! Or, TOTALLY NOT? Seriously, reverse hockey stick curves, FTL!

---

You know that is ruining this? What is making it so bad?

Capitalism. Desire to have a unicorn. Desire to become the Facebook of AI. Fuck you, Sam Altman. Fuck capitalism. We've got to get rid of it soon, or the planet will be trashed.

Of course, Peter Thiel & his ilk are all OK with that. Kill as many poor & sick people as possible. MAHA = eugenics. I mean, red states have refused Medicaid expansion under AHA for, what, 16 years now? How many 10,000s of people died prematurely due to that strategy? But, dead poor people, particularly if they are of color, is a feature, not a bug, to red state governors - and increasingly to the billionaire class.

Then, let loose the dogs of resource extraction. Siphon off as much wealth as they can before they retreat to their bunkers & leave the rest of us to die.

Motherfuckers. The sun is our father, the earth & the ocean are our mothers. They are literally fucking their mothers.

---

Here's the home page/directory for my posts on Bullshit.

Bee Speaker

"Bee Speaker", by Adrian Tchaikovsky, 2025, 434 pages.

This novel is book 3 of the "Dogs of War" series. Book 1 is blogged here and Book 2 is blogged here.

Our story takes place maybe a good while after the 1st 2. The civilization of Earth has mostly fallen, Mars made it through some tough times with the help of Bees, a distributed intelligence whose hardware originally was bees, and a complete commitment to mutual aid. Then Mars receives a call for help from Earth.

We still have uplifted dogs & lizards, a human group mind, a few pre-collapse immortal humans still left. Some insightful & surprising ideas about the difficulties people raised on Mars would have returning to earth.

I don't think this novel had the moral tone of the earlier ones, of people being people regardless of their form. It was a lot more of an action novel. The last 10 chapters or so were as fast moving & exciting as anything I've read lately. I stayed up late to finish it.

The Illusion of Thinking

Five Apple employes released a paper:

The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity

...

Through extensive experimentation across diverse puzzles, we show that frontier LRMs face a complete accuracy collapse beyond certain complexities. Moreover, they exhibit a counter-intuitive scaling limit: their reasoning effort increases with problem complexity up to a point, then declines despite having an adequate token budget.

Gary Marcus analyzes the paper here: A knockout blow for LLMs?

LLM “reasoning” is so cooked they turned my name into a verb.

And again today: Five quick updates about that Apple reasoning paper that people can’t stop talking about

The most popular rebuttal was … a joke. Literally.

---

Here's the home page/directory for my posts on Bullshit.