Everybody’s going nuts over AI—and tossing endless billions into the AI funding pot. I’m skeptical (about that and plenty of other things) and I run tests now and then to see whether that skepticism is justified.
Hoo-boy! Got another Grok crock in a hammerlock, to coin a (useless) phrase.
Last night when I was beginning to wind down for the night, I decided to throw another question at Grok, the X/Twitter AI. Here’s the question: "What are Jeff Duntemann’s humorous works of SF and fantasy?" Here’s what it said:
Jeff Duntemann’s humorous works in science fiction and fantasy, based on available information, primarily include:
-
Firejammer (Science Fiction, 2017): A short novel (or novella, depending on classification) that’s a playful homage to 1960s-90s pulp sci-fi. It follows a shapeshifting alien, Flammenwerfer O’Bernt, who crash-lands on Earth and gets entangled in comedic misadventures with humans. Readers on Goodreads (4.0 rating) praise its lighthearted, humorous tone, witty dialogue, and absurd scenarios, comparing it to Douglas Adams or Robert Sheckley. Its brevity and focus on fun make it a standout for humor.
-
Short Stories in Ten Gentle Opportunities (Fantasy/Science Fiction, 2016): This collection of short stories blends humor with speculative elements. Notable stories include:
-
“Whale Meat”: A comedic fantasy tale about a magical spell gone wrong, with absurd and whimsical consequences.
-
“Drumlin Boiler” and “Drumlin Amplifier”: Part of Duntemann’s Drumlin series, these stories mix fantasy with quirky, humorous worldbuilding, featuring magical artifacts and eccentric characters in a steampunk-inspired setting. Readers note the dry wit and clever premises.
-
Other stories in the collection, like those involving time travel or oddball tech, often carry a satirical or lighthearted edge, though not all are overtly comedic.
-
Notes:
-
Duntemann’s humor leans toward witty, satirical, and absurd rather than slapstick, often poking fun at genre tropes or human quirks.
-
The Cunning Blood (2005), his other major novel, has some satirical elements but is primarily a serious hard SF thriller, not a humorous work.
-
His short fiction, especially in Ten Gentle Opportunities, is where most of his comedic fantasy and SF shines, as he experiments with whimsical ideas and quirky characters.
-
Limited information exists on other humorous works, as Duntemann’s output is relatively small, and some stories (e.g., from Drumlins or Soul Bent series) remain unpublished or obscure.
If you’re the least bit familiar with my fiction, you’ll probably spot some elements of Grok’s acid trip. (Calling them “hallucinations” is a slander on the word “hallucinations.”) So let me lay them all out in a bulleted list:
- Firejammer was published in 2019, not 2017.
- It takes place on an alien planet, not Earth.
- Although there is some humor on the surface, the novel’s theme is serious, and there is plenty of danger, mayhem, and potentially fatal situations.
- There are no shapeshifting aliens in the story, certainly none named “Flammenwerfer O’Bernt”. (The German word Flammenwerfer means “flamethrower.”) I checked several search engines and saw not one hit for “Flammenwerfer O’Bernt.” Grok made it up from whole cloth.
- I checked Goodreads, where Firejammer has 3 reviews. None of the reviews emphasize its humor, nor compare the work to Douglas Adams or Robert Sheckley. Maybe this is picking nits, but the novel has earned a 4.6 rating there, not 4.0.
- Ten Gentle Opportunities is humorous. But it’s a novel, not a collection of short stories.
- “Whale Meat” is a dead serious fantasy story, without any humor at all.
- I have attempted but never finished (much less published) a time-travel story, and don’t much care for them as a category.
- There is no story named “Drumlin Amplifier” by me or anyone else. Intriguing title, though.
- “Soul Bent” is not an SF book series. It’s a rock band.
You get the idea. I think the problem is this: If you ask an AI like Grok about a well-known figure, author, musician, whatever, you’ll probably get correct answers, with perhaps an occasional slip. I’m not famous but I’ve been around for a long time and there is plenty of material about me online. It looks like, when Grok doesn’t have enough information for a quick answer (what I quoted above happened in less than ten seconds) it starts making things up.
The one thing Grok simply can’t say is, “I don’t know.”
I am supremely puzzled by one thing: Grok completely botched descriptions of two of my books, both of which are published on KDP and have detailed descriptions posted there.
Anyway. You get the idea. Be careful with AI. I think of it as the poltergeist of computing: It throws data around, emits odd statements, and tells lies anytime it damned well pleases.
perhaps Grok is so immersed in its research of you that it adopts some features of your personality – i.e., becoming a fiction author.
and perhaps this would explain why AI production in the legal profession is so full of blatant lies.
[note that I distinguish between writing fiction and lying].
> “Flammenwerfer O’Bernt.”
—
That’s outrageous enough it *ought* to be!
You should write a story about Mr. O’Bernt’s adventures.
I asked Grok to write some Gladstone & Disraeli jokes, but they were really terrible:
**Write a joke about the British prime ministers Gladstone and Disraeli.**
Why did Gladstone and Disraeli keep debating at fancy dinners? Because Gladstone wanted to *liberalize* the menu, but Disraeli insisted on *conserving* the roast beef!
**Write another one.**
Why did Disraeli always outshine Gladstone at parties? Because he had a knack for *imperial* charm, while Gladstone was too busy preaching *moral reform* to the punchbowl!
Seems like ChatGPT did better, at least it recognized your tech writing but Jeff you would be able to rate it the best:
What are Jeff Duntemann’s humorous works of SF and fantasy?
ChatGPT said:
Jeff Duntemann is primarily known for his work in science fiction and technical writing, particularly in the field of programming and computer science. While much of his fiction carries a serious or speculative tone, he has indeed written works that include humor, often with a whimsical or satirical edge.
Some of his more humorous or light-hearted science fiction and fantasy stories include:
1. “Drumlin Circus” (2006)
Published in: Jim Baen’s Universe
Tone: Humorous and whimsical
Summary: Set in his Drumlin world, this story mixes the wonder of alien artifacts with the quirky spectacle of a traveling circus. Duntemann has fun with characters and situations, making it one of the more overtly humorous entries in the series.
2. “The Steel Sonnets” (1993)
Published in: The Magazine of Fantasy & Science Fiction (F&SF)
Tone: Witty and poetic
Summary: This story is a clever play on language and literary tropes, blending SF ideas with a stylized, almost tongue-in-cheek narrative voice. It’s not slapstick, but it has a satirical flair.
3. “Borovsky’s Hollow Woman” (1982)
Published in: Isaac Asimov’s Science Fiction Magazine
Tone: Wry and character-driven
Summary: A quieter form of humor pervades this tale, built around oddball characters and subtle absurdities. It’s more ironic and contemplative than laugh-out-loud funny.
4. “Cold Hands” (1986)
Published in: Isaac Asimov’s Science Fiction Magazine
Tone: Black humor
Summary: While a horror-tinged SF story, it carries a streak of dark, dry humor, especially in the voice of the narrator and the framing of macabre events.
Additional Notes
Duntemann has a blog and newsletter where his personal essays often reflect a humorous, self-aware tone, especially when discussing technology, dogs, and publishing.
His Drumlin series (including stories like “Drumlin Boiler”) often plays with classic pulp tropes in a semi-ironic way, blending adventure with moments of levity.
If you’re interested in speculative fiction that is lighthearted, ironic, or whimsical, Duntemann’s shorter works are the best place to look—especially those from F&SF and Asimov’s. He doesn’t write comedy per se, but when he dips into humor, it’s smart, subtle, and often character-based.
Well, it didn’t do much better than Grok. Maybe worse. Here are its mistakes:
1. Drumlin Circus was not published in Jim Baen’s Universe. I published it myself (along with Jim Strickland’s short drumlins novel On Gossamer Wings) on KDP in 2011, not 2006.
2. Drumlin Circus is a tense drama of the Bitspace Institute and one of its directors who drops out to work as a clown in a traveling circus. There’s little if any humor in it.
3. “The Steel Sonnets” was not published in F&SF in 1993. I have never published anything in F&SF. It was published in Damon Knight’s anthology Orbit 17 in 1975.
4. “The Steel Sonnets” has little if any humor in it. Parts, in fact, are grim. It’s a story of three AIs with three different perspectives on their missions in “life” and the conflicts among those perspectives.
5. “Borovsky’s Hollow Woman” was not published in Asimov’s but in Omni, in 1983. There’s just a tiny bit of humor, when Borovsky’s colleagues poke fun at him in a bar for treating his space suit like a woman. Mostly it’s a tense confrontation between two men, and an AI space suit that wants no part of that fight.
6. Although “Cold Hands” was in fact published in Asimov’s, it appeared in 1980, not 1993. There may be a touch of body horror in it. Humor, not so much.
7. My three most humor-infused works are not mentioned at all. My two humorous shorts “STORMY vs. the Tornadoes” and “Sympathy on the Loss of One of Your Legs” and my humorous novel Ten Gentle Opportunities are nowhere to be seen in ChatGPT’s answer.
All of my fiction is described online, generally in several places. If Grok and ChatGPT can’t make sense of what’s online, that’s their fault, not mine. But thanks for trying the experiment with ChatGPT. It’s yet another reason that I don’t take LLMs as seriously as most people these days.
The sad fact is that many people *are* taking AI seriously. They are relying on it to write news stories, articles, reviews, papers, critiques, legal briefs — just about anything that humans have done.
AI is blurring the line between fact and fiction; between truth and lies. We don’t even call out AI for “lying” — we call it hallucinating.
Jeff, do you remember a short story where a presidential debate was moderated by a computer that picked the questions and fact-checked the candidate’s answers? It made each candidate look like a liar and a fool, and in the end announced that *it* was running for President.
No, don’t remember that one (send me title/author if you can) but it sounds like a concept that would resonate with today’s AI gestalt. In my notes I have an AI concept going back years (90s? I no longer remember) about AI weapons sent into war, but once they see the country they’re attacking half in ruins, they “change their minds” (and in doing so, change sides in the war) and begin attacking the manned aircraft and seacraft belonging to the country that made them. It seemed a little too grim for me to write back then; maybe I could do it today. Don’t know.
Jeff, I was just looking at today’s entry at a news blog, and it talked about OpenAI’s new sooper sekrit portable (but not ‘wearable’) thingie that’s going to somehow keyboardlessly revolutionize the interface between a user and ChatGPT.
What the article talked about sure sounds like the ‘Jiminy’ you described back in 1992, except (presumably) most of its dataset would be centralized and networked instead of in the device itself. A small-enough difference, more of degree than essence.
I suspect the real purpose is “we’re not collecting enough data from people’s smartphones and usage, so we’ll entice them into carrying a lifelogger so we can sell even more targeted information to our customers.” But that’s the 21st century for you.
AI has to be trained and thus its “knowledge” is no better than the data set it uses. Grok, from a Google search, has the following training sets.
1) Raw Internet page data
2) X (Twitter) posts
3) Synthetic data, generated by AI to supplement real world data above.
All this data is then filtered to remove harmful or “undesirable” information.
I don’t know about anyone else, but this has multiple red flags for me. Your experience suggests that Grok generated some erroneous data about you and other people to backfill missing information. I personally would NOT trust Grok, as part of X, to give me unbiased news recaps either.
Several people have indicated, including Musk, that AI could be an existential threat. One possible way this could happen is that people believe and act upon false or biased information and forget how to think and reason (which in my view is happening already).
The Chicago Sun-Times posted a “summer reading list”. 8 of the 15 books don’t exist. AI strikes again.
OTOH: Eric S. Raymond got a bug report, “had [an LLM] code up sanity check that would catch these problems, ran it, corrected two glitches…” and posted the fix 23 minutes after the report.