The Coyote Road

"The Coyote Road", subtitled "Trickster Tales", edited by Ellen Datlow & Terri Windling, 2023, 544 pages, 149k words. 26 pieces.

Grrr, the 2023 date is apparently the eBook. The original hardcopy collection was published July 19, 2007. I hate it when they do that shit. &, with AI, it will be happening more & more often. Because AI don't care if shit is correct, just that it's plausible. Here's the original hardcopy cover:

I think this was probably a $1.99 BookBub purchase. Of course, I love the trickster! My fav archetype!

Datlow & Windling edited several "Year's Best Fantasy & Horror" which I have copies of.

Overall, I'd say the collection was disappointing. Generally well written tales, but maybe not quite trickster enough for me. Not nasty enough. & not a single story featuring Loki!

There were 2-3 poems in the collection, I think I did not follow my "read aloud" prescription for poetry.

I think the 1st story that stood out was the Charles de Lint, "Crow Roads". I read some of his stuff 20-30 years ago. Any story involving corvids of course gets a boost.

I must have liked the Carol Emshwiller story, "God Clown", as I purchased her 2007 sci-fi novel "The Secret City". She died in 2019, age 97!

I really like Jedediah Berry's writing, his story "The Other Labyrinth" was good.

The head & shoulders above the rest, best story of the collection was the last, "The Evolution of Trickster Stories Among the Dogs of North Park after The Change", by Kij Johnson. "The Change" is most vertebrates acquiring the power of speech. As in Jablokov's "A Deeper Sea", where dolphins start talking, it doesn't go well. Way, way TMI. I bought a 2012 collection of Johnson's short stories, "At the Mouth of the River of Bees".

There still seems to be no sign of a "Year's Best" collection from Clarke. Reading a "Year's Best" collection every year has been part of my system for decades. Anybody got any suggestions or info on what's up with Clarke?

All Gary, All The Time!

I've been back in Lexington KY for 9 days now. I'm much less productive here, for various reasons. My latest Bullshit Apocalypse post was May 1. Since then, 5 posts have piled up to be shared - all by Gary Marcus!

As much as I appreciate the great work that Gary is doing, I think that this is a bad thing. It shows that the other people who post about the Bullshit Apocalypse, the horrible flaws of LLMs that are being completely ignored by the hype machine, are suffering from Bullshit Fatigue. They are worn out by the constant bullshit emitted by the AI hype machine. Bullshit Fatigue was in the title of my latest Bullshit Apocalypse post, I'm feeling it too.

---

Richard Dawkins and The Claude Delusion

The great skeptic gets taken in

This is one of the sadder essays I have ever had to write.

Wow, I so agree. This is heart-breaking.

I remember reading "The Selfish Gene" (1976) in the early 1980s. Dawkins invented the term "meme", & the science of memetics! So amazing!

Now, he spends 3 days talking to Claude, & "failed to convince myself that it is not conscious".

He's now 85 YO. Clearly his game is declining. Presumably it happens to us all.

From Gary's post:

The fundamental problem here is that Dawkins doesn’t reflect on how these outputs have been generated. Claude’s outputs are the product of a form of mimicry, rather than as a report of genuine internal states.

Consciousness is about internal states; the mimicry, no matter how rich, proves very little. Dawkins seems to imagine that since LLMs say things people do, they must be like people, and that simply does not follow.

---

Revisiting a theme we've touched on before.

Have LLMs improved patient outcomes?

A new review suggests otherwise

---

Breaking: Autonomous Agents are a Shitshow

Brace for chaos

I was somewhat disappointed that the article focuses on how easily hackable the agents are, via several avenues, and does not comment on their efficacy.

---

Breaking news: “they hadn’t figured out how OpenAI would pay for it”

Sign of things to come?

So OpenAI & Broadcom announce a Big Deal: 10GW of custom chips made together! But, who's paying? Oops!

Great diagrams of the "AI Circular Economy". As Doctorow says:

I'm worried that the seven companies that comprise 35% of the S&P 500 are headed for bankruptcy, as soon as someone makes them stop passing around the same $100b IOU while pretending it's in all their bank accounts at once.

---

Agents and ROI

Remember that MIT study that showed that the ROI for generative AI wasn’t really there for most businesses? Or any of the six or seven studies from other teams that followed, showing basically the same?

The situation for agents thus far looks to be similar: lots of hype; not so much ROI:

---

Here's the home page/directory for my posts on Bullshit. This is post #120.

Using a Looper

This post originally appeared in Jaz Dumoz Music Blog.

I bought my looper maybe 15 years ago. It is a Boss LoopStation RC-20XL Phrase Recorder. Here's what it looks like:

I 1st used it actively in Steve & Chris, 2017-2019. with me & most excellent harpist Steve Konopka (Fuzzy). I don't think you can buy them anymore.

I also used a Digitech LiveStation 3 harmony box. I haven't used that in years. It works great with Eagles, not at all w Louis Armstrong.

What Boss is offering now has buttons rather than pedals. I far prefer pedals, much easier to hit when you are playing standing up, in an unfamiliar configuration.

To start recording a loop, or to start playing it back, you press the left pedal. To stop the loop(s), you hit the right pedal. To erase the loop(s), you press & hold the right pedal.

Playing with a looper, over the years I learned 3 main principles. I added a less important 4th.

1. Per excellent looper user Jairaj Schwann: always start the looper on the 1.

   Jairaj was a serious musician, degree from UMass Amherst, ab fab bass player, moved to Atlanta. I had mostly found people using loopers to get pretty stale quickly. At a “guitarists” show, Jairaj was playing a 6 string bass. He would lay down a bass track on the looper on the bottom strings. Then add a rhythm track on the middle strings. Then play lead on the top strings. Incredible! & when he got tired of it, he'd wipe the whole thing & start over.

   Jairaj told me, “Always start the looper on the 1". I thought I knew better, I tried starting the looper on the 4 a few times. It never worked out. Always start the looper on the 1.

2. Given that the looper & the looper playback start on the 1, realize that sometimes you start a solo before you kick the looper in with the rhythm track. You may play 2-5 notes of the solo before you kick the background loop on. See example below.

3. This is a more minor point. Realize that, based on what you want to play a solo over, sometimes you will skip a verse or 2 before you kick the looper on to start recording. See example below.

4. When I 1st started the looper, I tried to use it as I would pedals on an electric guitar: I would start the looper replay, & then use the (awkward) volume knobs on my guitar to turn the volume of the solo up.

   At some point I realized, no! The looper has a knob (upper left of pic) that contols the output level of the recorded loop. I back it down from 12pm to 11pm. So the replayed loop is softer, the solo is at the original louder level. No playing with a volume knob, or overdrive pedal!

   I mentioned this to the sound man at an Open Mic I have been playing in Naples. I suggested him hitting the “Solo” button on the guitar channel, to put the rhythm guitar back to the volume it had been at & make the solo louder. He said he thought that the way it sounded now was perfect for the room we were in. He thought what I was doing was the proper thing to do, & wished others used the same technique - although I'm the only one using a looper there?!?!?

Normally I record a rhythm guitar part, then play lead guitar over it - so only 1 loop. You tap the left pedal to start recording, tap a 2nd time to replay the loop. You tap the right pedal to stop the loop.

I decided to try a Rickie Lee Jones song, “young blood” (1979), with 3 loops; a bass line, emphasis chords on the 1 & the 3, & a syncopated riff. I don't know if all loopers work this way, but on the Boss, you have to record the 1st loop, then play it back; record the 2nd loop, then play it back; record the 3rd loop, then play it back &  play a solo line over the top of all 3 of them. So you hit the left pedal a total of 6 times to get the 3 loops in place.

Note. this is my older musical alter ego, Jim Dumas. Jim plays music >~+ Elvis. Jaz is <~= Elvis. Note also, I use the harmony box at the end, on “Something happenin' there”.

Here's the song:

Here's a song demonstrating points 2 & 3 above: the Hoagy Carmichael classic “stardust”. I think this is pretty much Willie Nelson's version. You start the recording late. You start & stop the loop to solo over it 2x.

I (or rather Jim Dumas) have been thinking I'll go back & try some modern (> Elvis) songs with the looper. Some Hendrix maybe?

Reusing a Title? Bullshit Fatigue.

This bullshit is wearing me out. Hard to keep up.

From Project Syndicate:

AI Productivity Growth Won’t Match the Computer Revolution

Those promoting AI would be lucky to see the technology’s impact on output per hour match even the short-lived burst of the 1990s and 2000s. Productivity growth will underwhelm, not because the technology is weak, but because it creates a bottleneck that earlier digital tools largely avoided.

---

From Gary Marcus:

Dario Amodei, hype, AI safety, and the explosion of vibe-coded AI disasters

What the AI cheerleaders don’t tell you

LLMs are a great tool for professionals who know what they are doing. For the normal idiot, vibe coding is a coin toss, in a dark room.

---

I should probably have come up with a tag for this, it appears so often: LLM bullshit/slop overwhelms "{$domain}". From 404 media:

People Using AI to Represent Themselves in Court Are Clogging the System

More people having access to the courts is potentially good, but it’s not clear how the system can handle this increase in cases.

---

Introduced by Yves Smith at naked capitalism, a post by "Servaas Storm, a Dutch economist and author who works on macroeconomics, technological progress, income distribution & economic growth, finance, development and structural change, and climate change."

Russell’s Teapot: Dispatches From the Final Stage of the AI Bubble

Servaas systematically (and often wryly) takes apart the key claims of AI touts. One of my faves: “No, AI will create a lot of bullshit jobs.”

It is a good, thoughtful review of the current shape of things. The title references Bertrand Russell’s (1952) Tea Pot Analogy, which I had never heard of.

Russell's teapot is an analogy, formulated by the philosopher Bertrand Russell (1872–1970), to illustrate that the philosophic burden of proof lies upon a person making empirically unfalsifiable claims, as opposed to shifting the burden of disproof to others.

He lists 8 claims made by the genAI hype machine, & then explains why they are "teapots". The reference to David Graeber's "Bullshit Jobs" [the review/summary of which book I am still bogged down on] is in Claim #6.

---

Pointed at by Kottke, a very thoughtful article. The Bullshit Apocalypse is indeed apocalyptic - actually the author uses the word "catastrophic".

The Social Edge of Intelligence

AI doesn’t really “think.” Rather, it remembers how we thought together. And we’re about to stop giving it anything worth remembering.

The author mentions "degenerative AI", which we have seen before. AIs trained eating their own slop downward spiral into complete unintelligibility almost immediately. They need to be trained on tokens words produced by actually intelligent entities.

The more people use genAI, the less original training data there will be for our ever-hungry LLMs. The hundreds of daily hits I get on my blogs are certainly 99% AI training bots. So I am performing a service for the AI companies - for free, just like everyone else.

The author:

Bright Simons is a researcher, activist, and writer whose work sits at the intersection of global value chains, technology strategy, and institutional governance & design. He is the founder of the mPedigree Network, and affiliated with IMANI and ODI Global, both think tanks.

His theory of how to make genAI work is:

The Social Edge Framework outlined here is a direct counterpoint to Amodei, Aschenbrenner, and Altman. It is a program of action to counter the human redundancy fantasy. It challenges the self-fulfilling doom-spirals created by the premature reallocation of material resources to a vision of AI. I speak of the philosophy that underestimates the sheer amount of human priming needed to support the Great Recode of legacy infrastructure before our current civilization can even benefit substantially from AI advances.

---

Finally, today, another post from Gary Marcus:

“A model that produces code which compiles and passes the tests it was given is not the same as a model that produces correct, secure, maintainable, well-architected software”

A lot of code is being written by AI, but what does it mean?

...

Realism re AI coding is knowing that next-word prediction gets us a surprisingly long way in writing code, but less far in making sure that code is robust. Coders (especially vibe coders with little experience) beware.

And all you OpenClaw devotees, that goes 10x, if not 100x, for you.

---

That Gary Marcus article reminded me of some WOW (Words of Wisdom) I have printed & taped up on the bookshelves in my home office, for decades now. I have always thought these really conveyed the gist of coding & software development/architecture.

Computer Programming:

1. Every problem can be solved by breaking it up into a series of smaller problems.
2. The computer will always do exactly what you tell it to.

Software Engineering:

1. Writing the code is the easy part. Writing it so someone else can understand it later is the important part.
2. Make it work, then make it elegant, then make it fast.

---

Here's the home page/directory for my posts on Bullshit. This is post #119.

Music In, 2025, Q4

Following my Music In processing algorithm, I'm starting this post October 20, and will maintain it as the quarter progresses. I will be processing Q4 for a good bit longer. I'm back in KY & I don't get to listen to music near as much as I do in FL.

1. Ethel Lindsey, "Pretty Close", 2025, 8 tracks, BandCamp. Out of Paris. Who knew neo-Motown had moved from London to Paris? Very nice tunes. Ethel sure looks like a man in this video. There are references to she/her when you DuckDuckGo the name. Pictures have the same hairdo but no mustache & beard. Maybe somebody just goofing with "AI"? Regardless, good stuff, 4 stars. Here's "Together Again".

2. Werkha, "Unsung Irregular", 2025, 11 tracks, BandCamp. As usual, this is some good work by Tom Werkha, out of Manchester. 4 stars. This is definitely a catchy track, "Everyday (featuring Ríoghnach Connolly)".

3. Sudan Archive, "The BPM", 2025, 15 tracks, BandCamp. Out of LA. Violin. Some pretty compelling stuff. 4 stars. Here's the 1st track, "DEAD".

4. Magic Sam, "1957-1966 (West Side Guitar)", 1966, 21 tracks. Via my friend harpist Owen. Magic Sam died in 1969 of a sudden heart attack. OK blues. I will give "Roll Your Money Maker" 4 stars for an energetic call/answer with the band. 3 stars for the rest.

5. Magic Sam, "Black Magic", 1968, 18 tracks. OK blues. 3 stars for this as well.

6. Oscar Moore, "The 1945-1965 Years", 3 CDs, 28 + 28 + 25 tracks. Oscar Moore (1916-1981) was the main guitarist of the Nat King Cole Trio. He also played with several other people, and had his own trio. Here's the current overlaps with the Jaz Dumoz songbook:
   • A Foggy Day
   • Gee Baby, Ain't I Good To You
   • The Christmas Song
   • Fine and Dandy
   • Too Marvelous For Words
   • The Nearness Of You
   • April In Paris
   • I Can't Give You Anything But Love

   So I'll give those 4 stars, to hear his leads on them, & the rest 3 stars.

   Here's "April in Paris".

7. The Weather Station, "Airport & Only The Truth", 2025, 2 tracks, BandCamp. Out of Toronto. Very tasty female harmonies. Too bad it's only 2 tracks. 4 stars. Here's "Airport".

8. Lucius, "At the End of the Day", 2025, 3 tracks, BandCamp. Lucius has been confusing me lately. 2 tracks here, 3 tracks there. Maybe just an issue of $$$. They continue to reliably produce good pop tunes. 4 stars. Here's the title track.

9. The Paul deLay Band, "Heavy Rotation", 2001, 14 tracks, from Owen. This is a hard-working, modern blues band. A lot of good variety in the tunes. 4 stars. Here's "I'll Quit You Tomorrow", with a catchy signature lick & groove.

10. Juana Molina, "DOGA", 2025, 11 tracks, BandCamp. From Argentina. Why does it take Argentina to remind me what it was to be a hippy? Hmmm, I think I was erroneously conflating Juana Molina with Os Mutantes, a 1970s Brazilian psychedelic band. Molina started recording in the mid-90's. She's 64 YO now.

    Assimilating this new music has obviously been taking a while - it's 4 months after the end of 2025, Q4. But, this is the album of all of these I am still enjoying listening to the most. Go figure. 4 stars.

    Here's the 1st 2 tracks off of the album.

11. Victoria Port, "Barefoot in the Garden", 2025, 5 tracks, BandCamp. I continue to like these chill R&B tracks. 4 stars. Here's "Don't Let Me Fall".

12. the Mountain Goats, "Through This Fire Across From Peter Balkan", 2025, 12 tracks, BandCamp. This is like a rock opera. The opening track sounds overture-ish. It feels like a modern Broadway musical - mostly ugh. 3 stars.

13. Lucius, "New Old Tape", 2025, 2 tracks, BandCamp. More Lucius confusion. OK, the last 3 track EP had a song "Old Tape". This 2 track EP has a song "New Old Tape". Well, let's hear it. Hmmm, sounds the same as "Old Tape"????? 4 stars. There should be a deduction for confusion. Here's "New Old Tape".

14. Murder By Death, "The End (Egg & Dart Bonus Tracks)", 2025, 6 tracks, BandCamp. I have formed the impression that every song on this album is in a minor key. I mentioned before, I don't find slow songs sad, but minor key songs I sometimes do find sad. Not really so much here, but, still, it is a bunch of minor key songs. 3 stars.

15. Royce Wood Junior, "Penny Ballads", 2025, 5 tracks, BandCamp. How long has it been since Prince died? Oh fuck, April 21 was the 10th anniversary of his death. When he died, I was dumbfounded. He was clearly an immortal avatar of some kind, neh?

    It seems long overdue that somebody would try to fill that immense void. A few of these tracks definitely sound Prince-ish; others don't, but are also very good. Kind of sorry this is only 5 tracks. 4 stars. Here's the 1st track, "Clean Up".

16. David Byrne feat. Ghost Train Orchestra, "Who is The Sky?", 2025, 12 tracks, BandCamp. This is a very David Byrne album. Solid songs, quirky topics & lyrics. But, it seems like just more of the same of his recent albums. No standout track. Shockingly, I quickly got tired of listening to it. ?!?!? I had heard it, or its equivalent, before. ?!?!? 3 stars.

17. William Tyler, "Behold The Spirit", 2010, 9 tracks, BandCamp. The 3rd of his albums I've acquired. Very listenable guitar rock. Similar to the others, which were rated 3 stars. So ...

18. Redeye, "Stacks, 2025, 26 tracks, BandCamp. Redeye is a producer out of Toulouse, France. His last effort was very laid back, maybe too laid back. This one is "A collection of Beats, Loops and Edits" - most (well) under 2 minutes. Ahh, music for people with a short attention span. I think I'm going to like it. Since the tracks were so short, I did not pay $1/track, I went for $0.50/track: €11 for the 26 tracks.

    I enjoy listening to this, but, nothing stands out. So, 3 stars.

19. Various Artists, "Two Syllables Volume 22", 2025, 11 tracks. A free sampler from London record label First Word Records. From this label, I'm already following: Kaidi Tatham, Victoria Port, Allysha Joy, Children of Zeus. Amanda Whiting I don't care for.

    New artists:

    • Georgie Sweet. I think worth listening to. I'll watch for her releases. First Word will help. 4 stars. Here's "I Swear To You".

    • Myele Manzanza. Violin. Strong rhythms & melodic lines, nice! 4 stars. Here's "Absent Fade". Oops, pretty damn jazzy.

    • Ruby Wood. Surprisingly white. Nice tune. 4 stars. Here's "Let Him Go".

    So 4 stars for these new artists, 2 stars for the artists I'm already following (because I already have the tracks on an album), 3 stars for the rest.

Phew! Did a little push to get this published, I'm happy to be pointing people at such good music. But, getting harder & harder to keep up. Old age? Surely not!

Ooh, just checked, only 12 albums in 2026, Q1! Yay! I'll be slacking!

The Faith of Beasts

"The Faith of Beasts", James S. A. Corey, 2025, 448 pages, 128k words. Book 2 of "The Captive's War".

I blogged on the 1st book in this series here. In the Expanse series, which was so successful for these guys, they covered a lot of ground over the course of the series. Several different sci-fi tropes were pulled up.

So I'm curious if that will happen with this new series. The 2nd book in this series pretty much stays in the same lane as the 1st book. We learn more about the alien empire & their enemies. There are some interesting ideas here.

The ending of this book was somewhat anticlimactic. Well, 2nd books do tend to be transitionary. I enjoyed the read, & a 448 page book was good after 632 & 624 pagers. I'll definitely read the next book.

Incompetence & Thievery

Hah! I wrote the post title yesterday, reflecting the latest generative AI articles. I came back today & thought "Is this a post about the Orange Turd's administration?" So, correlation or causation? This is definitely the Golden Age of Grifting.

---

My wife used to watch "House" on TV. The ace diagnostician usually was able to get the right diagnosis in only 4-5 tries!

So maybe it's OK that genAI chatbots are wrong 1/2 the time on medical questions. From Gary Marcus:

Please don’t trust your chatbot for medical advice

Four separate studies all point in the same direction

...

All of this – the hallucinations, mistakes, and overconfidence – is entirely typical of LLMs, and entirely problematic in medicine. As the authors put it, in somewhat academic language, but entirely accurately, “continued deployment without public education and oversight risks amplifying misinformation.”

...

In a recurring theme, we see that LLMs don’t know what they don’t know; they work decently well with the information they’ve got but don’t know how to conduct clinical interviews, and in the hands of the lay public can easily give bad advice because the proper questions never get asked, either by the patient or the LLMs. (An expert doctor might use the LLM to better effect, by asking the right questions.)

Yves Smith of "naked capitalism" comments on the 1st study:

Half of AI Health Answers Are Wrong Even Though They Sound Convincing – New Study

It is telling that despite this article describing the abjectly awful performance of AI chatbots that it still gives a ritual defense of their use.

---

Now for the thievery. From 404 media:

This AI Tool Rips Off Open Source Software Without Violating Copyright

Malus, which is a piece of satire but also fully functional, performs a "clean room" clone of open source software, meaning users could then sell software without crediting the original developers.

This is really shitty. Open source software has been responsible for so much advancement. Most projects are kept afloat by a very small number of dedicated individuals, working for the common good & for the love of code.

But Doctorow weighed in on this, saying that it's really not that big a deal, because this "new" software, being AI-produced, cannot be copyrighted.

The (other) problem with automatic conversion of free software to proprietary software:

You can't add ANY license to a public domain work.

---

This last is just sad. From 404 media:

Startups Brag They Spend More Money on AI Than Human Employees

A new class of AI startups say they are taking money that would normally be used to hire people and are spending it on AI compute instead.

...

Startup CEOs who are “tokenmaxxing” are bragging that they are spending more money on AI compute than it would cost to hire human workers. Astronomical AI bills are now, in a certain corner of the tech world, a supposed marker of growth and success.

...

Companies like OpenAI and Anthropic are losing tons of cash on their products; even though artificial intelligence compute is expensive, it is underpriced for what it actually costs, and it’s not clear how long investors in frontier AI companies are going to be willing to subsidize those losses. Meanwhile, we have reported endlessly on “workslop” and the human cleanup that is often needed when AI-written code, AI-generated work, and customer-facing AI products go awry. There are also numerous horror stories of AI getting caught in a loop and burning thousands of dollars worth of tokens on what end up being completely useless tasks. Regardless, there’s an entirely new class of entrepreneur who seems hell-bent on “hiring” AI employees, not human ones.

Of course there's a word for it: "tokenmaxxing". This sure does remind me of other tech bubbles. I still cringe when I hear the word "pivot".

---

Here's the home page/directory for my posts on Bullshit. This is post #118.

Navola

"Navola", by Paolo Bacigalupi, 2024, 632 pages, 196k words.

Paolo writes very well, and heretofore was IMO one of our best climate fiction authors. I lost track of this novel somehow, it has been on my iPad a while, I'm not sure why I didn't get to it sooner.

This novel is definitely a change of pace. It is a fantasy set in a world reminiscent of di Medici Italy & Europe - the names are Italianish.

This is Machiavelli on steroids. It definitely follows Game of Thrones character handling guidelines, and has the expected body counts.

In the Acknowledgements, Paolo says this story came about when someone requested a story about a dragon. There is indeed a dragon in the story, but a very different & abstract type of dragon.

The book is definitely a page turner. I'm guessing there will be a sequel. We'll see, fingers crossed, I think.

Empire in Black and Gold

"Empire in Black and Gold", by Adrian Tchaikovsky, 2008, 624 pages, 194k words. Book 1 of 10 in the "Shadows of the Apt" series.

I have repeatedly mentioned how much I enjoy Tchaikovsky's work. I'm not sure why I bought this book, I'm guessing BookBub offered it to me for $1.99 or so.

This is quite a tome - 624 pages. Plot synopsis:

The world is populated by humans who are divided into races based on insect species they assimilated DNA from? Or? Clearly fantasy.

Each has different superpowers - Art that they can practice. Many species can conjure wings & fly. Insect species include:

• Ants - they are great warriors, & telepathic;
• Beetles - the engineers & merchants;
• Flies - the fastest fliers, very popular as messengers;
• Moths - they ruled everything until 500 years ago, when beetles & others became Apt: tech-savvy. Some races are Apt, others not, they cannot deal with tech at all;
• Mantises - the ultimate warriors, enforcers for the Moths;
• Dragonflies - non-Apt species living to the north;
• Wasps - the bad guys, imperialists knocking off other races' cities 1 by 1. All males are soldiers, enslaved races are tasked with the normal work of keeping their civilization running. Their superpower is shooting energy bolts from their palms - their "sting";
• Scorpions - desert dwellers;
• Various other species, some legendary.

So we follow a group of 4 college students, members of a fencing team put together by a beetle spymaster trying to wake cities up to the danger of the ever-encroaching wasps:

1. the spymaster's plain, nerdy niece ward, a beetle (Velma);
2. the spymaster's non-niece ward, a gorgeous ant hybrid (Daphne);
3. a dragonfly exchange student (Fred);
4. a beetle hybrid engineering geek (Shaggy).

OK, I'm sorry, but I couldn't help but get this totally mapped to Scooby-Doo! KMN.

The plotting is great, the characters are great (even the Scooby-Doo ones), the action is non-stop. The conclusion is satisfying, even though, yes, 9 sequels are looming. Tchaikovsky wrote 2 in 2009, 2010, 2011; 1 in 2012, 2013, 2014.

But 1 thing bothered me the whole time I was reading this. The various insect-friend subspecies of humans were referred to as "races" repeatedly. Is this not racist?

Level 1 racism is discriminating for or against people based on their race. This is now universally reviled, yes?

But, from what I have read - I think primarily David Graeber "Debt - the First 5000 Years" - race is not a scientific concept. It is a political concept, invented around the 17th century to justify the arbitrary enslavement of Africans.

In Rome, anyone could be a slave - it had nothing to do with "race". There were only a few ways one could become a slave: as a prisoner of war; being sold as a slave by one's parents; selling one's self into slavery.

So level 0 racism is believing that there is such a thing as race. It is not a concept in biology or genetics. There is only 1 race: the human race. All humans can interbreed & produce non-sterile offspring. It is mostly impossible to identify "race" analyzing human DNA. Science!

So in this book, by identifying different insect-aligned "races" - who can all appearently interbreed, meaning, by definition, that they are the same species - is racism being promoted?

If we go back to the wellspring of fantasy, Tolkien, he had "races": humans, elves, dwarves, wizards, orcs. They were the same species, yes, they could all interbreed? But, elves were immortal, humans not - halfbreeds not.

"Halfbreed" - what an offensive term. Humans breed with humans to produce more humans.

So I don't know what to make of this. We all want to oppose level 1 racism, but, somehow, we let level 0 racism slide? I'm not sure at all where I'm going with this? Maybe some bright young writer at tor.com or such will figure out what I am trying to say?

This book was very enjoyable, totally a page-turner. Tchaikovsky did not disappoint. But, I think I'm going to put the next 9 volumes on hold for the forseeable future. So I got that going for me now.

Are We Done?

I am so onboard with almost all of Cory Doctorow's analyses. This recent post of his touches on so many relevant points.

Last night I talked to a guy who was a huge AI cheerleader. I would rate him an intellectual lightweight - exactly the type that likes "AI". They can generate bullshit that makes them sound like a heavyweight!

Doctorow gives me moral support in opposing the Bullshit Apocalypse. Hmmm, I hadn't heard that Leon Skunk was proposing a Dyson sphere around our sun for the express purpose of powering "AI". Barf.

AGI isn't enough for the hardcore techbros, they want ASI - Artificial Super-Intelligence. The title of this post echoes Doctorow's question: when do you determine that you have wasted enough resources & energy on an unattainable goal?

I'm including a pretty big chunk here, it does not do justice to Doctorow's analysis. Go read it, please.

A Pascal's Wager for AI Doomers

Now, as I've stated (and as I said onstage) I am not worried about any of this. I am worried about AI, though. I'm worried a fast-talking AI salesman will convince your boss to fire you and replace you with an AI that can't do your job (the salesman will be pushing on an open door, since if there's one thing bosses hate, it's paying workers).

I'm worried that the seven companies that comprise 35% of the S&P 500 are headed for bankruptcy, as soon as someone makes them stop passing around the same $100b IOU while pretending it's in all their bank accounts at once. I'm worried that when that happens, the chatbots that badly do the jobs of the people who were fired because of the AI salesman will go away, and nothing and no one will do those jobs. I'm worried that the chaos caused by vaporizing a third of the stock market will lead to austerity and thence to fascism:

https://pluralistic.net/2026/04/13/always-great/#our-nhs

I worry that the workers who did those jobs will be scattered to the four winds, retrained or "discouraged" or retired, and that the priceless process knowledge they developed over generations will be wiped out and we will have to rebuild it amidst the economic and political chaos of the burst AI bubble:

https://pluralistic.net/2026/04/08/process-knowledge-vs-bosses/#wash-dishes-cut-wood

In short, I worry that AI is the asbestos we're shoveling into our civilization's walls, and our descendants will be digging it out for generations:

https://pluralistic.net/2026/01/06/1000x-liability/#graceful-failure-modes

---

From NYT News Service/Syndicate Stories, via the Lexington Herald-Leader:

How Accurate are Google's AI Overviews?

A recent analysis of AI Overviews found that they were accurate approximately 9 out of 10 times.

I don't think we can live with a 9% error rate. & I don't think that "autocomplete on steroids" is ever going to improve on that much.

Google search is now basically a front-end for Google's LLM, Gemini. All your queries are belong to us, to train our AI! I'm so glad I've switched to DuckDuckGo & Kagi!

---

From the people formerly known as AI Snake Oil, a framework for evaluating AIs. It includes a review of current capabilities, a very mixed report.

Open-world evaluations for measuring frontier AI capabilities

Introducing CRUX, a new project for evaluating AI on long, messy tasks

---

Here's the home page/directory for my posts on Bullshit. This is post #117.

Human Beings Are Weak

So, chatbots excel at blowing smoke up the ass of their users, &, guess what! Human beings love having smoke blown up their asses! Who would have guessed it? From Bruce Schneier:

AI Chatbots and Trust

All the leading AI chatbots are sycophantic, and that’s a problem:

Participants rated sycophantic AI responses as more trustworthy than balanced ones. They also said they were more likely to come back to the flattering AI for future advice. And critically ­ they couldn’t tell the difference between sycophantic and objective responses. Both felt equally “neutral” to them.

Abandon hope, all ye who enter here. Welcome to the Bullshit Apocalypse!

---

Here's the home page/directory for my posts on Bullshit. This is post #116.

We Have A Winner: "Just More Bullshit Hype"

Looking back at my last post "Mythos: Threat, Menace, or Just More Bullshit Hype?", we have a winner: "Just More Bullshit Hype". From Gary Marcus (as usual):

Three reasons to think that the Claude Mythos announcement from Anthropic was overblown

No need to panic just yet

---

Next, from me, an anecdote. I flew up last Thursday to visit my son & his family in Cary, NC. There didn't used to be any non-stop flights - there may be now, on Breeze Airways - my return flight was non-stop, RDU to RSW (Fort Myers FL). I usually flew on American, changing planes in Charlotte, NC, but the layover was only 40 minutes, &, in my increasing aversion to any drama, I decided to fly Delta through Atlanta, w a 2.5 hr layover. I had lunch at a nice piano bar, really good player. Nice spinach salad.

Waiting to board in Fort Myers, they called for carry-on bags to be gate checked through to your final destination. I'm like, sure, I'm happy to not have to fool with my bag in Atlanta airport.

In Atlanta, after my very nice lunch, I'm waiting to board. I'm in zone 6. They're boarding zone 4 & they announce, "All overhead bins are full, all carry-on bags must be gate checked." So I'm glad I did it in Fort Myers.

I get on the plane, the overhead bins are 2/3 empty!!! I ask a flight crew member, "Why did they say these were full?". They replied, "A computer algorithm determines when we make that announcement."

Seriously, WTF! Somebody using AI? For something involving physical things in the physical world? LLMs aren't good at anything but words.

We are going to see more & more of this, systems that are completely crappy, that don't work as often as they do, or more often than they do. LLMs, FTL! I feel that there is maybe a 50% probability that LLM use will eventually shut down large parts of the economy. Time frame, 2-3 years.

---

Another anecdote from me, on a more general bullshit topic. I made a little progress on my review/summary of David Graeber's "Bullshit Jobs". Spoiler alert: my current working conclusion is that "Bullshit Jobs" is bullshit - a monument to selection bias.

So I was chatting with an early 30 YO manager of a team of financial analysts working for a large insurance company. Graeber identified financial analysts as having a job that was usually bullshit. I asked them "Do you have a bullshit job?"

They replied (paraphasing):

No I don't think I have a bullshit job. I think we produce some useful results.

But, do I have to deal with tons of corporate bullshit? Of course I do! I work for a big corporation, the bigger the corporation, the more corporate bullshit you have to deal with.

So maybe "Bullshit Jobs" isn't a problem with jobs, it's a problem with corporations. Per Doctorow, corporations are "immortal colony organisms that treat human beings alternately as a source of nutrients or a form of inconvenient gut flora." Surprising that working in such an environment involves mass quantities of bullshit - NOT!

The young analyst also told me, they were under a lot of pressure to try & integrate LLMs into their workflows. Any pushback on this was treated as recidivism, & was severely career-limiting.

---

Finally, more from Gary Marcus. Claude Code may implement old school AI! FTW!

The biggest advance in AI since the LLM

Why Claude Code changes everything

...

The source code leak proves it. Tucked away at its center is a 3,167 line kernel called print.ts.

print.ts is a pattern matching. And pattern matching is supposed to be the *strength* of LLMs.

But Anthropic figured out that if you really need to get your patterns right, you can’t trust a pure LLM. They are too probabilistic. And too erratic.

Instead, the way Anthropic built that kernel is straight out of classical symbolic AI. For example, it’s in large part a big IF-THEN conditional, with 486 branch points and 12 levels of nesting — all inside a deterministic, symbolic loop that the real godfathers of AI, people like John McCarthy and Marvin Minsky and Herb Simon, would have instantly recognized.

Putting things differently, Anthropic, when push came to shove, went exactly where I have said for 25 years that the field needed to go: to Neurosymbolic AI.

That’s right, the biggest advance since the LLM is neurosymbolic. AlphaFold, AlphaEvolve, AlphaProof, and AlphaGeometry are all neurosymbolic, too; so is Code Interpreter; when you are calling code, you are asking symbolic AI to do an important part of the work.

So, there is more to getting real, working AI than the statistical "autocomplete-on-steriods" LLMs? Well, DUH!

---

Here's the home page/directory for my posts on Bullshit. This is post #115.

Mythos: Threat, Menace, or Just More Bullshit Hype?

So Antropic has a new model named Mythos, which apparently is good at finding defects in code. They are delaying its release, & giving a prerelease to all the tech biggies, so they can attempt to get their codebases fixed up. In the hands of hackers, it is feared to be a major chaos agent. From Gary Marcus:

What should we take from Anthropic’s (possibly) terrifying new report on Mythos?

Everybody’s talking about Anthropic’s new but unreleased model Mythos (and the related Project Glasswing) and how it might undermine or even devastate cybersecurity.

The New York Times has a similar take. DuckDuckGo "Antropic Mythos", every tech outlet has a story.

Anthropic Claims Its New A.I. Model, Mythos, Is a Cybersecurity ‘Reckoning’

The company said on Tuesday that it was holding back on releasing the new technology but was working with 40 companies to explore how it could prevent cyberattacks.

I guess this is laudable of Antropic. But, on the other hand, the last couple of years have conditioned me to wonder, how much of this is just more bullshit marketing hype?

---

Here's the home page/directory for my posts on Bullshit. This is post #114.

No Man's Land

"No Man's Land", by Richard K. Morgan, 2026, 496 pages. Hmmm, 2 books in a row of 496 pages?!?!?

I have greatly enjoyed Morgan's prior work, I just wish there had been more of it. But, he has also done gaming & screenplay things, so fewer novels.

I think I had already read the start of this novel before. It is after WWI. The depravity & destruction of this war destroyed faith in scientific rationalism. In it's place, the old forests spring back into being overnight, populated by various kinds of fae, including totally bad-ass immortals, led by 6000 YO Queen Mab.

Our hero is a WWI veteran who has become a Woodsman. He rescues human children kidnapped by the fae & replaced by changelings. Kind of a good cheap detective noir feel, as he becomes embroiled in intrigue involving the highest levels of the British government & the fae.

Very gritty, very cheap detective noirish. Definitely a page turner. I think this will be a series, there's lots more ways he can go with this.

Children of Strife

"Children of Strife", by Adrian Tchaikovsky, 2026, 496 pages, 170k words. Book 4 of the Children of Time series.

I think I have mentioned a few times that Tchaikovsky is my fav new hard sci fi author of the last decade or so. The Children of Time series is my favorite of all he has done. Uplifted hunter spiders, octopuses, corvids, mantis shrimp, plus a sentient bacteria with atomic level storage. What's not to like?

That said, I think book 4 is my least favorite of the series. We don't get a new cool species to enjoy. We get a tale told across 3 eras, with 3 different groups of people. It's a fairly long book, I thought it dragged a bit in the middle. The conclusion is well done & satifying.

I would guess he will continue this series? Presumably it sells pretty well? There are lots of different directions he could head in. I guess we'll see.

Now That is Some Bullshit!

1st up, from John Kwarsick:

The Machine That Can't Explain Itself

Why AI adoption is outrunning accountability and what happens when the auditors arrive

From 2024 to 2025, the Foundation Model Transparency Index published by the Stanford Institute for Human-Centered AI dropped from 58/100 to 40/100. Oops.

In my very 1st anti-LLM rant, I described LLMs as "indecipherable oracles" - I don't see how that's going to change.

---

This next I find almost unbelievable. LLMs will tell you what's in an image without seeing the image?!?!? They just confabulate what's likely to be in the image?!?!? Holy crap, now that's some bullshit! From Gary Marcus:

The mirage of visual understanding in current frontier models

When a model achieves a “top rank on a standard chest X-ray question-answering benchmark without access to any images” you know something is deeply wrong.

Marcus is reporting a research paper just released by Stanford:

MIRAGE: The Illusion of Visual Understanding

First, Frontier models readily generate detailed image descriptions and elaborate reasoning traces, including pathology-biased clinical findings, for images never provided; we term this phenomenon mirage reasoning. Second, without any image input, models also attain strikingly high scores across general and medical multimodal benchmarks, bringing into question their utility and design. In the most extreme case, our model achieved the top rank on a standard chest X-ray question-answering benchmark without access to any images. Third, when models were explicitly instructed to guess answers without image access, rather than being implicitly prompted to assume images were present, performance declined markedly.

Another great new term in the dream world of LLM hallucinations: "mirage reasoning".

This tech is clearly NOT ready for prime time, but, $T invested, tough shit. You will take your bullshit sandwich & enjoy eating it, nom-nom!

I've downloaded the PDF from arxiv, reading through it, it's scary but also LOL. I shouldn't laugh, the Bullshit Apocalypse is not funny, but laughter has always been a part of what keeps me sane. So, laugh, dumbass, laugh!

---

Here's the home page/directory for my posts on Bullshit. This is post #113.

Yay! Wow, 2 Yays in a Row! The Tide Is Turning!

A big YAY yesterday, another big YAY today! The tide is turning (or not). From the most excellent 404 Media:

Wikipedia Bans AI-Generated Content

“In recent months, more and more administrative reports centered on LLM-related issues, and editors were being overwhelmed.”

1 of the worst things? The worst thing? About AI slop is, it overwhelms human capabilities to fact check, edit, stave off the Bullshit Apocalypse. Wikipedia has totally done the right thing. Just shut the slop out completely. Kudos.

---

Here's the home page/directory for my posts on Bullshit. This is post #112.

Yay! Disney Bails On OpenAI-Generated Slop Video

After only 3 months, Disney bails on a $1B investment in OpenAI to allow anyone to create slop videos using Disney characters. The concept was apparently a lead zeppelin. From 404 Media:

Disney's Sora Disaster Shows AI Will Not Revolutionize Hollywood

It turns out when you try to serve slop on a product people pay for, no one wants it.

---

Here's the home page/directory for my posts on Bullshit. This is post #111.

Notably Mixed Results

1st, a tale of woe, from The Leadership Lighthouse:

I Went All-In on AI. The MIT Study Is Right.

My all-in AI experiment cost me my confidence

...

The company gets excited about AI. Leadership mandates AI adoption. Everyone starts using AI tools. Productivity metrics look great initially. Then something breaks, or needs modification, or requires actual judgment, and nobody knows what to do anymore.

The developers can’t debug code they didn’t write. Product managers can’t explain decisions they didn’t make. Leaders can’t defend strategies they didn’t develop.

---

Next, a good review of the state of things. Interesting that he picks Yann LeCun as "the voice of sobriety about AI in today’s world". LeCun was a major "AGI soon" cheerleader until fairly recently. Ditto for Geoffrey Hinton, also profiled in the article, still a cheerleader to my knowledge. From PhD computer scientist David William Silva:

I'm Sorry to Burst Your Bubble: You Are Being Fooled About AI, and You Will Soon Feel Really Stupid

The overall inability to distinguish what is real from what is fake is fueling the frisson.

This I had not heard before, but it is very believable & a matter for concern:

And here is what I’ve found: every time I had an idea that was genuinely outside the box, AI discouraged me from pursuing it. When I persisted, it didn’t just push back, it practically begged me to stop. To not continue. To not insist.

---

3rdly, from 404 Media, a link to a website dedicated to exposing hype:

Ridicule as Praxis (with Emily Bender and Alex Hanna)

Why ridicule works to keep big tech’s claims in check, and what makes us hopeful for the future.

Here's the website:

The Mystery AI Hype Theater 3000 podcast

“deflates AI hype and draws attention to the real harms of the automation technologies we call ‘artificial intelligence’.”

---

Finally, some good news! Generative AI problems solved! A guest article from the AWS blog:

Overcoming LLM hallucinations in regulated industries: Artificial Genius’s deterministic models on Amazon Nova

Artificial Genius uses SageMaker AI to perform a specific form of instruction tuning on Amazon Nova base models.

This patented method effectively removes the output probabilities. While standard solutions attempt to ensure determinism by lowering the temperature to zero (which often fails to address the core hallucination issue), Artificial Genius post-trains the model to tilt log-probabilities of next-token predictions toward absolute ones or zeros. This fine-tuning forces the model to follow a single system instruction: don’t make up answers that don’t exist.

This creates a mathematical loophole where the model retains its genius-level understanding of data but operates with the safety profile required for finance and healthcare.

The authors state that this approach is much better than RAG, which was already supposed to fix hallucinations. I guess we'll see. Like RAG, it looks like after-the-fact bubble gum & duct tape to me.

The primary author is Paul Burchard, PhD, "Founder and Partner of Artificial Genius". He has an interesting resume.

---

Here's the home page/directory for my posts on Bullshit. This is post #110.

Wendig Triples Down

I think that most excellent sci-fi author Chuck Wendig has an attitude towards generative AI at least as negative as mine. In fact, his attitude is probably much more negative than mine because, as a professional author, he has serious skin in the game. He doubled down on AI-bashing in December, he now triples down.

Shy Girl, AI In Writing, And A New Perniciousness

I am against AI because it steals our work, which it then uses to steal our jobs, which it further uses to steal our water and our electricity.

Which is to say, it is here to steal our future.

So, I’m against it! It sucks moist open ass.

...

And don’t use the AI slop-shitting artbarf techbro bullshit.

Wendig's post talks about "The Terminator" movie & includes a pic, which I have copied. Who doesn't like a pic of a terminator walking through fire?

---

From The Register, which appears to now be a part of MSN?

AI still doesn't work very well, businesses are faking it, and a reckoning is coming

Codestrap founders say we need to dial down the hype and sort through the mess

Note, most of the content of this article comes from an interview with "Dorian Smiley, co-founder and CTO of AI advisory service Codestrap", who I'm gonna guess is selling something. The content is decent. But still, capitalism FTL or FTW, you decide.

---

From MindPrison:

How Verifier Loops Made AI Coding Useful | The Rise of Vibeware, Abandonware and Technical Debt as Consequence

How did unintelligent, hallucinating AI progress from nearly useless to writing entire, complex applications and at what cost?

This is a long article, lots of good news, bad news. Despite the positive headline, I'm going to posit that the bad news outweighs the good news. For example:

“Vibe Coding makes you 10x faster at making 10x more technical debt.” — Franziska Hinkelmann, PhD

From the 80s on, we called it "maintenance debt", not technical debt. Same thing. Writing the software is not the hard part, handing it off & maintaining it is.

---

Here's the home page/directory for my posts on Bullshit. This is post #109.