"Starman" by Gerry Conway, Mike Vosburg, and Mike Royer

A long time ago, I didn't read superhero comics. It's hard to believe, I know! But compared to many of my interests, I came to them relatively late, not until I was in college. One of my gateways was a blog called, I think, Comics 101 by Scott Tipton (later a writer for IDW's Star Trek and Doctor Who comics). Tipton would do deep dives on different comics characters—his history of Green Arrow was responsible, for example, for my deep and abiding interest in the character. I feel like you can find this stuff all over the Internet these days, but back then it was much rarer.

One of Scott's favorite ever comics was James Robinson's Starman, which he did a multi-part deep dive on, and he made it sound so good and so much to my taste, than when DC released the six-volume Starman Omnibus set a couple years later, I blind-bought the entire run as it came out. These days, the giant hardcover omnibus that reprints everything in order is a common format, but back then it was brand new; in fact, I am pretty sure that Starman was only DC's second foray into it, following the Jack Kirby's Fourth World Omnibus. If you read Reddit's "OmnibusCollectors" sub, you will see lots of people who blind buy stuff they don't actually like based on hype, but I like to think I am actually very good at it, and I loved the run; I even got my wife to read it, and she enjoyed it too. My reviews of it from the time are all posted here on this blog.

One of the things that Robinson was very good at was pulling strands of the DC universe together. Though Robinson's Starman, Jack Knight, was the son of the original Starman, Ted Knight, DC had actually published several unrelated Starman characters between the two. Robinson didn't ignore these, but worked them into his story. I've long been curious about the original appearances of those Starmen: Mikaal Tomas, Prince Gavyn, and Will Payton. So my next comics project is to read the back issues about those other Starman characters, and then reread the James Robinson run, with both those original stories, not to mention twenty years of further comics reading, to inform my understanding. (For example, Starman is heavily connected to the Justice Society mythos, something I knew little about when I originally read it, but now know far too much about!)

Back when I added them to my list of things to check out, the stories of Mikaal and Gavyn were both uncollected, and though they have been collected since, I've still elected to pick up their original publications, since the collected editions in question have a lot of other material that makes them more expensive than just picking up the back issues. The fourth Starman, Will Payton, has never been collected at all. (I do wonder if this will all get collected in a "DC Finest" someday and how it will be handled if so. A DC Finest is typically 500-650 pages long... all of the 1970s Starman content totals less than 150 pages. Would they collect it alongside either the end of Golden Age run, or the beginning of the late 1980s run? It will be a weird collection either way.)

---

So the first post–Golden Age Starman revival was in the pages of DC's 1st Issue Special, a thirteen-issue series where every issue was the "first issue" of a potential run. In most cases, however, the hypothetical run never materialized, and Starman was no exception. Until James Robinson brought him back two decades later, this remained the only outing for this version of Starman.

This Starman is Mikaal Tomas, who is a blue-skinned alien from space. (The lettercol implies he could be from the same planet as Shadow Lass of the Legion of Super-Heroes.) It's the 1970s, and disco and David Bowie are popular, so he wears a shirt with a plunging neckline and a gold medallion. I feel like I often hear this version of the character referred to somewhat dismissively as the "disco Starman"...

...but that must be by people who have not actually read the comic, who have just seen pictures of the character, because this is excellent. Seriously, this isn't high art, but I thought it was masterfully put together. The creative team here is an all-star one: Gerry Conway (Steel, the Indestructible Man, Sun Devils), Mike Vosburg (Starfire), and Mike Royer (Fourth World, Kamandi) were all probably at their heights when they assembled to do this.

If only you didn't suck, Mikaal!

Basically, in eighteen pages, this story does exactly what a good first issue ought to do. Efficiently and effectively, Gerry and the Mikes give us a clear premise and a number of drivers of future conflict. Mikaal comes from a planet whose inhabitants intend to invade the Earth; both he and his girlfriend think this is wrong... but only his girlfriend is willing to do anything about it, and she dies in the attempt. This spurs him to travel to Earth to try to protect its people, only his spaceship crash lands, injuring him. He is rescued by a grocery store owner; meanwhile, a pair of cops investigate his crashed ship, and one of them dies... and then he's followed to Earth! 

So we already have a clear cast of characters with some interesting dynamics, especially the cop who thinks that Mikaal's the bad guy. The action is clear and sharp, the exposition is efficient and not heavy handed at all, it ends on an excellent cliffhanger. I get why this was never followed up on (1st Issue Special was cancelled one issue later, and the Wikipedia page on the series quotes an interview with Conway that does a good job of explaining of why it never really worked as a "tryout" book anyway), but it's a shame. I flew through this issue, and I loved reading it.

Presumably this guy and his beef with Mikaal are never heard from again.

I'll be curious to see how much of it James Robinson actually draws on when he brings back Mikaal; I don't remember much beyond the bare fact of his existence!

(Incidentally, there's no indication this is set in the DC universe: the humans in the story certainly don't act like people who are used to superheroes and aliens flying around. Robinson would place the story specifically in the 1970s, not on DC's sliding timescale, so this actually works out, as it's set in the relatively superhero-light period between the retirement of the JSA and the emergence of the JLA.)

"Starman" originally appeared in issue #12 of 1st Issue Special (Mar. 1976). The story was written/edited by Gerry Conway and illustrated by Mike Vosburg & Mike Royer. 

This is the first in a series of posts about Starman. The next covers Adventure Comics #467-78.

Reading Roundup Year in Review 2025/26

I first started tracking my reading when I went away to college in 2007. Thus, my "reading year" starts about the same time the school year does, and so every September, I crunch the numbers on last year's reading.

Last year, I was distressed to note that I'd fallen a bit below my post-pandemic average of 134 books per year... this year I am happy to note that I had my best year since 2016/17! (I spent that school year commuting via bus, so I got a lot of reading done... on the other hand, my audiobook and podcast listening fell off a lot.) The only other year to beat this year is 2012/13, when I read for my Ph.D. exams.

The key to success is, of course, read more comic books.

SERIES/GENRE/AUTHOR	# OF BOOKS	BOOKS/ MONTH	% OF ALL BOOKS
Star Trek*	31	2.6	16.2%
Doctor Who†	15	1.3	7.9%
Star Wars*	16	1.3	8.4%
Planet of the Apes	1	0.1	0.5%
Media Tie-In Subtotal	63	5.3	33.0%
Oz / L. Frank Baum	14	1.2	7.3%
Naomi Novik	3	0.3	1.6%
Lois McMaster Bujold	2	0.2	1.0%
Nnedi Okorafor*	2	0.2	1.0%
Ursula K. Le Guin*	2	0.2	1.0%
Other Science Fiction & Fantasy	27	2.3	14.1%
General SF&F Subtotal	50	4.2	26.2%
Justice League	7	0.6	3.7%
Supergirl	4	0.3	2.1%
Legion of Super-Heroes	3	0.3	1.6%
The Sandman†	2	0.2	1.0%
Other DC Universe Comics	13	1.1	6.8%
Black Panther	13	1.1	6.8%
Other Marvel Universe Comics	3	0.3	1.6%
Avatar: The Last Airbender	2	0.2	1.0%
Other Comics	6	0.5	3.1%
Comics Subtotal	53	4.4	27.7%
Victorian Literature	2	0.2	1.0%
Other Literature	10	0.8	5.2%
General Literature Subtotal	12	1.0	6.3%
Penguin History of Britain	3	0.3	1.6%
Other Nonfiction	10	0.8	5.2%
Nonfiction Subtotal	13	1.1	6.8%

* Comic books relating to series or authors that are predominantly not comics I don't count under my "Comics" category, but under the main designation.
† Nonfiction about a particular author or series is included with that series, not the "Nonfiction" category.

It wasn't just comics, though; my tie-in reading was in general up this year, as I worked through a bunch of Star Trek and Star Wars books I got like a decade ago but never got around to. (In Star Wars, I have not yet gotten up to the Disney era still!) I can read a 300-page Star Trek book much faster than a 300-page original novel.

(One notes that there are previous years where the total doesn't add up to 100%. Apparently I made some data entry errors, but I am too lazy to track them down; I'm sure this is good enough.) 

This was Star Trek's best year since 2007/08, and Star Wars's best since 2013/14; I had gone five years with no Star Wars books at all. My combined comics went from 9.7% last year to 27.8% this year.

It might look like my general sf&f reading went down, but that's only in terms of share; in raw numbers, it actually went from 47 to 50! I just read more other stuff this year. 

Those are stats I crunch myself (with the help of ChatGPT); here are ones I used LibraryThing to generate. I make different choices between how I enter books on LibraryThing vs. in my personal files, so the total number of books will be slightly different.

Here's their breakdowns by author. (Note that these are about authors, not books by authors, if you see the distinction.) First, what countries did my authors originate from:

My U.S. share is a tad bigger than previous years, I'm guessing because of all the tie-ins I read. My U.K. share is substantially smaller. My nine "others" were one each from Croatia, India, New Zealand, Mexico, Ghana, the Netherlands, Ukraine, and Israel, plus one "not set." (I am a bit skeptical of some of these identifications; Michel Faber, for example, hasn't lived in the Netherlands since he was seven, and has published every single one of his books since moving to the U.K., in English. Is he really a Netherlands author? I feel like some Thingamabarians just want excuses to pick "interesting" countries.)

Most of my books were by living authors, which is normal.

Men always dominate my reading, but this is their highest year since I started doing these graphs, in 2020/21. Again, it's probably all the tie-ins and comics I was reading. I actually read more women authors this year than last (it went from 33 to 39, which actually ties my previous record); I just read more more male authors this year too (from 56 to 87).

One statistic I enjoy a lot on LibraryThing is a breakdown of what you read by pages. This is imperfect: I only enter page counts for paginated books, and many comics and ebooks have no page numbers, and of course page numbers don't perfectly correspond to word counts. Also, multi-author books like anthologies and comic book collections can only be attributed to one person. But still, I find it interesting. Here's my top authors by pages read: 

Keith Giffen did not write all of the two (very large) comic collections where he is first credited author, so once you remove him, David Mack is my favorite author two years running! Joey Cavalieri and Jack Kirby are the other two dubious inclusions in the top ten; Cavalieri wrote just one issue in a 1,438-page collection, but it's the first one, so my method of cataloguing on LT gives him credit for the whole volume. Kirby wrote more of the Panther's Prey Omnibus, but not even a majority of it, I don't think.

The highest-placed author to get on the chart through just a single book is Ada Palmer. She beats out John Jackson Miller, who wrote four!

My tagging on books gives you a sense of genre and series and other attributes: 

Science fiction dominates, of course; lots of comics there, too, and if all comic collections were paginated, it would be even higher.

Finally, here's how many book are on my "To be read" list:

I've sort of plateaued recently. Hopefully I can get it down. At least I haven't crossed the dreaded 700 Barrier!

You can compare this to previous years if you're interested: 2007/08, 2008/09, 2009/10, 2011/12, 2012/13, 2014/15, 2015/16, 2016/17, 2017/18, 2018/19, 2019/20, 2020/21, 2021/22, 2022/23, 2023/24, 2025/25. (I didn't do ones for 2010/11 and 2013/14.)

Christopher Mills (ed.), Noir Quarterly (1994–95)

In order to read one of the Maze Agency prose stories, I picked up the first and only issue of Noir: The Illustrated Crime Fiction Quarterly from Alpha Productions. This actually was not very easy to do: one of those old comics that not so much expensive as unavailable!Published at the comics size, it was mostly made up of detective prose stories with just one comics story, plus a couple nonfiction features. Despite the subtitle and the grand ambitions of editor Christopher Mills (who also edited The Detectives one-shot from Alpha, in which a short Maze Agency comics story appeared), it only lasted one issue. 

Mills later revived the concept under the slightly different title of Noir Quarterly at CFD... though as he admits in a retrospective on his blog, the publication frequency was "way-too-optimistic." There are no Maze Agency stories in those issues, but look—I like collecting interesting and unusual comics, and if tacking two more issues of something onto my order can get me a complete set, then I'll do it. Plus those two issues had beautiful femme fatale–themed covers! (They're both by Brian Stelfreeze, who much much later would be the first artist on Ta-Nehisi Coates's Black Panther run.)

All three issues have similar contents, though not always in the same proportions: predominantly, there are illustrated prose stories, with a couple comics. Most of the material is new, but sometimes there's archival material; one of the comics in Noir Quarterly #1 is from 1955, for example, and a few of the prose stories are reprints, though Mills says he prioritized reprinting stuff that was otherwise uncollected. There are also nonfiction features: book reviews, movie reviews, interviews, and retrospectives on some topic of interest to the audience, like old movies or comics.

from Noir Quarterly #1
(script by Ron Fortier, art by Gary Kato & Steve Crompton)

Like any anthology, the prose stories are hit-or-miss, but there definitely was enough good stuff here to make it all worth reading. Though I was drawn to the series by The Maze Agency, I would say the majority of the prose stories are not detective fiction. Ones I particularly liked included "Surrogate" by Ed Gorman (illustrated by Richard Pace, from Noir vol. 1 #1), where a lawyer confronts the teenager who might be responsible for the death of his own son; "The Love Rack" by Max Allan Collins (illustrated by Blackhawk's Rick Burchett, from Noir vol. 1 #1), a gritty story about a private eye who gets caught in political conspiracy; "A Fine Officer" by C. J. Henderson (illustrated by Michael Harris, from Noir Quarterly #1), where a detective becomes caught in a gang civil war; and "They're Gonna Kill You after a While" by Ron Goulart (from Noir #2), about a detective investigating the disappearance of an actor who once played a Saint-like character in a series of films. This last story was illustrated by Brian Michael Bendis of all people; I had no idea he was also an illustrator! Of the ones I liked, only this last one was a mystery.

I don't read much "noir"; having gotten this sampling, I think what I don't like are ones that seem a bit... gratuitous. I don't mind cynicism, but sometimes it moves over into crassness. I didn't care for "Fool's Gold" by William Marden (illustrated by Darren Goodhart, from Noir vol. 1 #1), which goes on a long time for a mediocre ending, during which you have to read a lot about men ogling a teenage girl. Particularly bad, I thought, were two stories from Noir #2: "Granny Gumption Solves a Murder" by C. J. Henderson & Charles Hoffman (illustrated by Michael Harris) and "The Wallow" by Ardath Mayhar (illustrated by Delfin Barral). The former has a decent joke at its core, about Miss Marple–type characters, but then gets nasty; the latter is pretty obvious from its very beginning but takes a long time to get there.

from Noir: The Illustrated Crime Fiction Quarterly #1
(script by Nicola Cuti, art by Joe Staton)

The amount of comics content varies heavily. Nicola Cuti and Joe Staton's Michael Mauser character appears in both Noir vol. 1 #1 and vol. 2 #2 (he also appeared in the Detectives one-shot). I like the Joe Staton art, of course, but I think the first story suffered from the fact that I have read no other stories about the character than these three: when his brother popped up as part of a solution to a mystery, it was a bit jarring as there was no previous hint he had a brother! But I would read more about this guy. That's the only comics story in vol. 1 #1, but Noir Quarterly #2 has two: a mediocre Johnny Dynamite reprint from 1955, plus a new story by Ron Fortier with art by Gary Kato & Steve Crompton. This same team also supply one of Noir #2's three comic stories, and Ron Fortier scripts another story for issue #2 with art by Randy Buccini & Byron Peneranda. I don't think any of these are high art, but I do think everyone involved very much understood the assignment, and I enjoyed all of them. My favorite was probably the one in Noir Quarterly #1, "Blood Red Lips," about a dangerous femme fatale with a twist.

I did find the nonfiction features pretty interesting, though there were too many of them in Noir #2. This is a whole world I don't know very much, and since I will probably never have enough time in my life to delve into more deeply, I appreciate how other people can give me access. The book reviews were usually interesting enough, but things I particularly enjoyed included the interview with Mickey Spillane by Mills from Noir Quarterly #1 (I've never read any of his work, but he seems fascinating) and the overview of The Saint and Falcon films by Ed Gorman from Noir #2 (what a weird production history... though my guess more interesting to read about than to watch the actual films).

Overall, this was a decent package, though I did find Mills's editorial voice a bit grating at times. Not always great work here, but distinctive and interesting, a nice teaser into a wider world I am unlikely to delve more into, to be honest, but I am glad I tracked down and read it all!

Noir: The Illustrated Crime Fiction Quarterly vol. 1 originally appeared in one issue (Winter 1994). The stories were written Ed Gorman, Wendi Lee, Mike W. Barr, Nicola Cuti, William Marden, and Max Allan Collins, and illustrated by Richard Pace, Terry Beatty, Rick Magyar, Joe Staton, Darren Goodhart, and Rick Burchett, with nonfiction features by Jeff Gelb, Eric Davidson, and Christopher Mills. The series was edited by Christopher Mills.

Noir [Quarterly] vol. 2* originally appeared in two issues (Apr.-Fall 1995). The stories were written by Joe Gill,† William F. Nolan, Robert J. Randisi, William Marden, Ron Fortier, C. J. Henderson (with Charles Hoffman), Nicola Cuti, Ardath Mayhar, and Ron Goulart, and illustrated by Bill Molno & Vince Alascia,† Rick Burchett, Flint Henry, Bob Cram, Jr., Gary Kato & Steve Crompton, Michael Harris, Joe Staton, Mike Harris, Delfin Barral, Brian Michael Bendis, and Randy Buccini & Byron Peneranda, with nonfiction features by Ron Fortier, Christopher Mills, Eric Davidson, Jonathan Anger, Bill Spangler, James Chambers, and Ed Gorman. The series was edited by Christopher Mills. 

* The first issue was titled Noir Quarterly, the second just Noir.

† The Johnny Dynamite story "The Fixer's Mob" (a reprint from 1955) is not credited in the issue; credits courtesy the Grand Comics Database.

Star Trek: The Official Motion Picture Adaptation by Mike Johnson & Tim Jones, David Messina, et al.

Star Trek: The Official Motion Picture Adaptation

Collection published: 2010
Acquired: June 2021
Read: October 2025

I spent some time late last year into early this year reading my way through the various "Kelvin timeline" comics and novels. Some of them I reviewed on the TrekBBS, though I fell behind. But what I never got around to doing was posting those reviews here. So, finally, time to do it!

Star Trek: The Official Motion Picture Adaptation #1-6 (Feb.-Aug. 2010)

screenplay by Roberto Orci & Alex Kurtzman, adaptation by Mike Johnson & Tim Jones, pencils by David Messina with Claudia Balboni, inks by Gaetano Carlucci, colors by Giaovanna Niro and Ilaria Traversi, letters by Robbie Robbins and Neil Uyetake

I was kind of worried about this book going in; in my experience, most comics adaptations of films are not up to much, often reading as a condensed version of a story optimized for a different medium.

To my surprise, then, scripters Mike Johnson & Tim Jones and artist David Messina deliver a pretty strong retelling of the 2009 film. I mean, it won't knock your world alight, but it does a good job of hitting the beats of the story via the medium of comics. Probably this is at least partially because it's been given a whole six issues, so (mostly) it's not too compressed. But also Johnson/Jones/Messina have good pacing instincts, sometimes piling a lot of panels into a single page, sometimes using double-page spreads in order to capture a key moment. Clearly also some thought was put into putting issue endpoints at key moments, so that each issue works as a unit.

from Star Trek: The Official Motion Picture Adaptation #5
(art by David Messina & Claudia Balboni and Gaetano Carlucci)

The result is an adaptation that reads fairly genuinely as a comic. It's probably at its best when covering the earlier parts of the story; it's not until the end of issue #3 that Nero's attack begins, so the early parts of the book have a lot of space to unfold key moments like the destruction of the Kelvin or the reveal of the Enterprise in space. On the other hand, moments like Kirk's and Spock's fight on the bridge are given numerous panels on a single page, really giving them an impact.

Indeed, issue #4 is probably the one compression hurt the most; the drill battle is just not interesting in this format, with key moments like Sulu's unfolding sword going by too quick to leave an impression, and I found the death of Amanda way too rushed to give it the import it needed, and then the issue's final panel, weirdly, does not clearly communicate the destruction of Vulcan. The unnecessary Galaxy Questeseque second climax of the film, where the Enterprise is almost pulled into a black hole, comes across as even more unnecessary here.

from Star Trek: The Official Motion Picture Adaptation #4
(art by David Messina & Claudia Balboni and Gaetano Carlucci)

The book's solicits trumpeted the inclusion of cut material, but this mostly seems to just be a couple lines of dialogue here and there. David Messina is a longstanding Star Trek artist for IDW, and he mostly does a good job here, though I struggled a bit with his likeness for Bruce Greenwood's Pike. Overall I think this does exactly what it needed to, and did it well. Probably the best comics adaptation of a Star Trek film other than the two by Peter David.

This is the first in a series of posts about the Kelvin timeline Star Trek comics (and novels). The next installment covers More Beautiful than Death.

What Is the Pangram AI Detector's False-Positive Rate for LLM-Generated Prose?

As I have mentioned in a previous post, I spent a lot of time this summer battling bots on Reddit. I would like to say I'm pretty good by now at identifying the cadences of LLM-generated prose, especially in Reddit posts—LLM-generated Reddit posts tend to have some very specific tics. (If you want to see them in action, a very useful thing to do is to ask ChatGPT to make you a Reddit post, and you'll see how much it is like all the posts you've seen elsewhere.)

But my sense of how LLM-generated something is is not definitive proof, so in my Reddit bot takedowns, I tend also to link to AI-detector results. Fairly often, someone pops up to claim AI detectors are unreliable. I'm not so sure. I mean, I am sure there are some unreliable ones out there (anyone can create a site and claim it's an AI detector), but that doesn't mean they all are.

My poking around has turned up that it seems like the consensus is the top three are GPTZero, Pangram, and Turnitin. Obviously Turnitin is only useful to people in the academic ecosystem; I've settled on using Pangram in my Reddit takedowns, because you can link to your results page and you get a decent number of tests per day on the free account.

There are two ways to think about the reliability of these tools. First is the false-negative rate. How often do they identify something that is LLM-generated as not LLM-generated? And then there's the false-positive rate. How often do they identify something that is not LLM-generated as LLM-generated? For this post, I'm focusing on the latter. This is important in the context of calling out people on Reddit and also on calling out students: if you make an accusation, you want to know that the positive result of the AI detector is reliable. (In the context of identifying academic misconduct, you also don't want a high false-negative rate: if a ton of students using LLMs are not being flagged, that's annoying... but that's a different issue than the one I'm discussing here.)

Pangram claims a false-positive rate of 0.0041% in their own materials. But, of course, they are incentivized to claim to have really low ones! What I have been curious about, then, is what independent studies turn up. Specifically, I was looking for sources where the writers used Pangram themselves on a corpus of known human-written work, and used that to calculate a false-positive rate. (Or, where I could use their data to calculate it, if they didn't do it themselves.) I excluded anything that was directly published by Pangram itself, or anyone working for Pangram. (Fun fact, though: Adam et al. is published by Pangram competitor GPTZero!) Here's what I have been able to find.* I have sorted the results by denominator, i.e., I start with the study that had the largest corpus of human-generated texts and work my way down to the smallest. To make this table legible, I just included author names in it; full citations are in my Works Cited at the end of this post.

Source	Peer Review?	Specific Citation	False-Positive Rate
Evanko and Di Natale	Conference paper	Calibrated on abstracts from 2021 to 2022 (slide 4).	0.22% (40/18,467)
Adam et al.	Preprint	Tables 1-2 on p. 7 give the accuracy (all texts classified correctly) and recall (true-positive rate) across five domains, each of which consists of 1,000 AI texts and 1,000 human texts; only in one domain did Pangram have 2 false positives.	~0.04% (~2/5,000)
Elazar and Antoniak	Preprint	“We test [...] Pangram on CS papers from 2020- 2022. We find that [...] Pangram did not predict any false positives” (p. 3); they sampled 100 papers per month for the three pre-LLM years of 2020-22 (p. 4).	0.0% (0/3,600)
Saha et al.	Peer reviewed	“For reviews that are fully human written, Pangram flags none of them as ‘AI’ or ‘Mixed’” (p. 3); number of human reviews given on p. 5.	0.0% (0/3,499)
Jabarian and Imas	No	“FPR is essentially zero across all thresholds 0.5 and greater” (p. 10); number of human texts given on p. 5.	0.1% (2/1,992)
Day	No	“~1,000 random abstracts were retrieved from articles published in 2010”; “1 positive detection in 2010 data.”	0.1% (1/974)
Shah and Levy	Preprint	“The pre-AI period (2019–2022) yields 1 detection across 800 documents, a rate of 0.1%” (p. 28).	0.1% (1/800)
Lee	No	“Pangram and GPTZero correctly flagged no human writing as AI.”	0.0% (0/495)
Karr et al.	Conference paper	FPR for 2013-15 control period given in Table 2 on p. 3; number of human texts given in Table 1 on p. 2.	0.0% (0/306)
Saha and Feizi	Conference paper	FPR given in Table 6 on p. 25426.	0.0% (0/300)
Allaham and Diakopoulos	Preprint	“In our evaluation of both Pangram and GPTZero on a curated dataset of 200 human-authored texts [...], we find that Pangram and GPTZero have 100% accuracy in correctly detecting human-authored and AI-generated texts, respectively” (p. 3).	0.0% (0/200)
Lu et al.	No	No AI-generated papers found in the 2022 control group (Table 2).	0.0% (0/159)
Russell et al.	Peer reviewed	FPR given in Table 2.B on p. 5346; “Each experiment consists of 60 articles, 30 human-written and 30 LLM-generated” (p. 5343), and there were five experiments.	2.0% (3/150)
Chakrabarty et al.	Preprint	“Human written text was never misclassified” (fig. 2E, p. 5).	0.0% (0/150)
Ren et al.	Preprint	Ran a control on abstracts posted to arXiv in 2010: “Pangram also perfectly classifies the human abstracts” (p. 42).	0.0% (0/100)
Guppenberger	Preprint	FPR given in Table 2 on p. 10.	0.0% (0/50)
Van Vlasselaer et al.	Peer reviewed	“The results for fully human-written texts show that all four tools classified 100% of human texts correctly as ‘True Negatives’ [...]. This indicates that none of the detectors are prone to incorrectly flagging human writing as AI-generated for this collection of papers” (p. 12).	0.0% (0/40)

You'll see that the majority of independent tests give Pangram a 0% false-positive rate, including one where the corpus included 3,600 texts! If you just crudely average all of the false-positive rates from all seventeen studies, the mean false-positive rate is 0.15%. If you pool all the numerators and denominators, you get 49 / 36,282, which is a false positive rate of 0.14%.†

This would suggest that it would take you 658 submissions to find one false positive according to the crude mean, and 740 submissions to find one false positive according to the pooled average. So overall, Pangram has a very low false-positive rate, and thus you should feel entitled to make LLM-use accusations with a high degree of confidence if you have a Pangram score as support.

Note that many of these studies had other goals in mind than just assessing the false-positive rate of Pangram, so in some cases, it's a very small part of a wider topic. Often, this other stuff is quite interesting: Ren et al. and Karr et al. both discuss, for example, the susceptibility of LLM detectors like Pangram to attempts to evade AI detection, thus resulting in a higher false-negative rate. (It's worth pointing out that some other of these studies, such as Russell et al., Jabarian and Imas, and Van Vlasselaer et al. tried out evasion techniques that Pangram was able to still detect, however.)

One that I found particularly interesting was Russell et al., whose purpose was to compare automated detection tools to human beings' own ability to detect LLM-generated prose. They hired human beings to annotate their pool of texts, and found that

annotators who rarely or never use LLMs are poor detectors of AI-generated text; in fact, they overestimate their own ability to perform the task by providing high confidence scores for their decisions. We identify a subset of five high-performing annotators who frequently use LLMs for writing tasks (e.g., editing, copywriting, creative writing). The majority vote of this subset of “expert” annotators fails to predict the correct label on only one out of 300 articles. (5342–3)

In fact, only Pangram outperformed their group of "expert" annotators. (This might account for my own ability to detect LLM-generated prose; while I have certainly never used ChatGPT to generate content for this blog or academic writing, I do mess with it a lot in other ways, including helping me devise images and Tasks for my Star Trek Adventures scenarios and perform analyses on my iTunes library, so I have seen a lot of its prose in action by now.)

The other study I found particularly interesting was Lu et al. This was done by the chairs of a track at the conference NeurIPS, who "made the decision to require that all papers be substantially human-written, with AI used for only copy-editing or similar peripheral changes..." They partnered with Pangram to enforce this policy, and used papers from 2022 as a control. 0% of those papers were identified by Pangram as AI-written. When they used it on the 2026 batch, they had to outright desk-reject 18.4% of submissions for AI use, and ask for "evidence of substantial human engagement" on another 12.7%. A full 30% of people used AI to write their position papers knowing that this track of NeurIPS had a no-AI policy! It's not just undergraduate students who can't be bothered to obey academic integrity rules, and who don't think about the implications of using LLMs for their writing tasks.

* Full disclosure: this is a combination of sources I found myself and supplemental ones found for me by ChatGPT (and one by the world's most underachieving LLM, Google Gemini). ChatGPT also helped me do the HTML formatting for my table and Works Cited.

† There were two other pieces I chose not to include in my table above, partially because they were just published by individuals on Substack, but they are both also interesting reading. @VeryFinePrint OCRed forty-six undigitized old books, which came out to 8,025 of what Pangram calls "segments," and got false positives on just 4 segments, so 0.05%... but it turned out that in all four segments, their AI-based OCR program had hallucinated text, so Pangram was accurately detecting AI-written segments! The other, by Kelly, states that he tested 96,468 essays and got just 10 false negatives and 1 false positive, but doesn't actually state how many of the 96,468 essays were human-written, so we don't know what the denominator should be. I have included both sources on the Works Cited below.

Works Cited

Adam, George Alexandru, et al. “GPTZero: Robust Detection of LLM-Generated Texts.” arXiv, 13 Feb. 2026, doi: 10.48550/arXiv.2602.13042.

Allaham, Mowafak, and Nicholas Diakopoulos. “Synthetic Sources?: Auditing Generative Search Engine Citations for Evidence of AI-Generated Sources.” arXiv, 22 May 2026, doi: 10.48550/arXiv.2605.23684.

Chakrabarty, Tuhin, et al. “Readers Prefer Outputs of AI Trained on Copyrighted Books over Expert Human Writers.” SSRN, 19 Aug. 2026, doi: 10.2139/ssrn.5606570.

Day, Adam. “GenAI Detection That Actually Works.” Clear Skies, Medium, 18 Aug. 2025, clearskiesadam.medium.com/genai-detection-that-actually-works-edc562581fed.

Elazar, Yanai, and Maria Antoniak. “LLM-Generated or Human-Written? Comparing Review and Non-Review Papers on arXiv.” arXiv, 19 Jan. 2026, doi: 10.48550/arXiv.2601.17036.

Evanko, Daniel S., and Michael Di Natale. “Quantifying and Assessing the Use of Generative AI by Authors and Reviewers in the Cancer Research Field.” International Conference on Peer Review and Scientific Publication, 3 Sept. 2025, peerreviewcongress.org/abstract/quantifying-and-assessing-the-use-of-generative-ai-by-authors-and-reviewers-in-the-cancer-research-field.

Guppenberger, James. “Vernacular Remote Control: Adversarial Robustness Testing of Six Commercial AI Text Detectors on Provenance-Guaranteed Pre-AI Prose.” Research Square, 12 June 2026, doi: 10.21203/rs.3.rs-9596240/v1.

Jabarian, Brian, and Alex Imas. “Artificial Writing and Automated Detection.” National Bureau of Economic Research, Sept. 2025, doi: 10.3386/w34223.

Karr, Jonathan A., Jr., et al. “Why AI Detection Fails for Academic Integrity.” ACM AI Leadership Summit, 20 Aug. 2026. arXiv, doi: 10.48550/arXiv.2608.11256.

Kelly, Trent. “Practical Attacks on AI Text Classifiers with RL.” Substack, 8 July 2025, trentmkelly.substack.com/p/practical-attacks-on-ai-text-classifiers.

Lee, Jaeho. “AI Detectors Rarely Flag Human Writing, but Sometimes Miss AI Text Imitating Real Authors.” Epoch AI, 15 July 2026, epoch.ai/data-insights/ai-detectors-false-negatives.

Lu, Alex, et al. “AI-Generated Papers in the NeurIPS 2026 Position Paper Track.” NeurIPS, 2 June 2026, blog.neurips.cc/2026/06/02/ai-generated-papers-in-the-neurips-2026-position-paper-track.

Ren, Kevin, et al. “Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift.” arXiv, 23 June 2026, doi: 10.48550/arXiv.2606.25152.

Russell, Jenna, et al. “People Who Frequently Use ChatGPT for Writing Tasks Are Accurate and Robust Detectors of AI-Generated Text.” Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics, July 2025, pp. 5342–73. ACL Anthology, doi: 10.18653/v1/2025.acl-long.267.

Saha, Rounak, et al. “Policies Permitting LLM Use for Polishing Peer Reviews Are Currently Not Enforceable.” International Conference on Machine Learning, 23 June 2026. arXiv, doi: 10.48550/arXiv.2603.20450.

Saha, Shoumik, and Soheil Feizi. “Almost AI, Almost Human: The Challenge of Detecting AI-Polished Writing.” Findings of the Association for Computational Linguistics, July 2025, pp. 25414–31. ACL Anthology, doi: 10.18653/v1/2025.findings-acl.1303.

Shah, Anand, and Joshua Levy. “Access to Justice in the Age of AI: Evidence from U.S. Federal Courts.” SSRN, 21 May 2026, doi: 10.2139/ssrn.6766859.

Van Vlasselaer, Marijke, et al. “Who Wrote This? Evaluating the Reliability of AI Detection Tools in Higher Education.” International Journal for Educational Integrity, vol. 22, 29 June 2026. Springer Nature Link, doi: 10.1007/s40979-026-00226-w.

@VeryFinePrint. “Scanning for Pangram Errors.” Substack, 15 July 2026, veryfineprint.substack.com/p/scanning-for-pangram-errors.

The Maze Agency: "The Adventure of the Chinese Dissident" and Other Stories by Mike W. Barr et al.

Somewhat belatedly—I think while writing up the previous post in this sequence—I discovered that there were actually three Maze Agency short stories that had been published outside of the regular Maze Agency comic series. One was a comic story published in a one-shot from Alpha Productions called The Detectives; the other two were prose stories, which could be found in an installment of the quarterly series Noir (a mix of comics and prose) and an anthology called Sex, Lies and Private Eyes. So I tracked them down to cap off this series. (When the hypothetical Maze Agency omnibus is released, hopefully it will include them all.)

The comic story in Detectives #1 (which has a cover by original Maze Agency ongoing artist Adam Hughes), "The Adventure of the Chinese Dissident," is just ten pages, and so can't be as complicated as most Maze Agency stories, but is pretty good regardless. Its main downside is that you only get any clues at all about a page before it's all revealed, so there's not really a chance to "play along" as the reader... but it's not like I was ever going to do that successfully in any case! Paul Pelletier is a solid artist here (I mostly know him from small bits and bobs on series like The Darkstars or Blue and Gold), and I would have gladly seen more Maze Agency from him.

The other two are, like I said, prose stories. It's weird when characters originated in comics make the jump to prose, in my opinion; I think a lot of time there's a kind of flatness. Comics tell you interiority through  mostly external cues, especially in a series like The Maze Agency, which has no narration. I spent some time reading a bunch of superhero-comics-turned-prose-stories, and I found it was hard to imagine a lot of superhero characters as actually having vivid interior lives—not just for me, but obviously for the writers! Mike Barr isn't inexperienced in prose (I actually read a couple of his prose Star Trek works two decades ago now), but he doesn't really go for interiority at all here: the narrator is distant from Jen and Gabe. This can work in a detective story, but it did make me feel like these weren't really playing to the strengths of the medium.

from Noir: The Illustrated Crime Fiction Quarterly #1

Weirdly, both "The Heat Is On" and "Murder Gone Wylde" are about pornography: the former about an adult video store, and the latter about a producer of porn DVDs. I found "The Heat Is On" had a good sense of atmosphere that worked well, but the actual mystery was slight; "Murder Gone Wylde" has an engaging set of characters and suspects, and was a lot of fun... unfortunately, I think it might have been even more fun as an actual comic! Still, as one last way to spend time with Jen and Gabe (and Lieutenant Bliss, etc.), you could do a lot worse. Did Barr ever write any original mystery novels? I feel like he could have done good things with that canvas.

"The Adventure of the Chinese Dissident" originally appeared in issue #1 of The Detectives (Apr. 1993). The story was written by Mike W. Barr, pencilled by Paul Pelletier, inked by Bob Cram Jr., lettered by Susan Dorne, and edited by Christopher Mills. 

"The Heat Is On" originally appeared in issue #1 of Noir: The Illustrated Crime Fiction Quarterly (Winter 1994). The story was written Mike W. Barr, illustrated by Rick Magyar, and edited by Christopher Mills. 

"Murder Gone Wylde" originally appeared in Sex, Lies and Private Eyes (2009). The story was written by Mike W. Barr and edited by Joe Gentile and Richard Dean Starr.

This is the last in a series of posts about Mike W. Barr's The Maze Agency. Previous installments are listed below:

1. The Maze Agency vol. 1 #1-5 (1988-90)
2. The Maze Agency vol. 1 #6-23 (1989-91)
3. The Maze Agency vol. 2 #1-3 (1997-98)
4. The Maze Agency vol. 3 #1-3 & vol. 4 #1 (2005-23)

Two related posts cover Noir Quarterly and Sex, Lies and Private Eyes.

Star Trek: The Destiny Era: Coda: The Ashes of Tomorrow

July 2387
My guess is that I am going to reach book III of Coda and decide Coda was a bad idea. Some might roll their eyes at this declaration, but I just want to put my cards on the table as a reviewer. People are allowed to go into works of art with a suspicion of what they will think about them, contra the opinions of some creators. But, also, maybe I am wrong! Certainly I have my own thoughts on what a "novelverse" wrapup trilogy should have done, but we'll see.

Star Trek: Coda, Book II: The Ashes of Tomorrow
by James Swallow

Published: 2021
Acquired: October 2021
Read: August 2025

That said, if it had to be done this way, The Ashes of Tomorrow is probably about the best it could have been done. James Swallow's novel picks up from the end of book I, Moments Asunder, with Devidians devouring the galaxy. Whereas Moments Asunder mostly focused on the Enterprise-E crew, this book begins with them, but as it goes on, increasingly focuses on the Deep Space Nine cast. Yes, Picard and Worf are there, but the series cast who are all there, getting the spotlight, are the DS9 ones.

It's really effective. One of the great things the original DS9 relaunch did was be willing to move on from the tv characters in the configuration we had known them... but this does mean that it's been eleven years of story time and twenty years of real time since Sisko, Kira, Bashir, O'Brien, Odo, Quark, Jake, and Nog got together to solve a problem as a team. Swallow very effectively hits the nostalgia button here; I loved the scene where they all end up together, trying to figure out what can be done about this problem. Swallow has a very strong command of their voices.

If I have a complaint, it's that while most of the novel-original characters (e.g., Taran'atar, Vaughn) had been killed off by this point, we don't get a return appearance by, say, Shar or Prynn. This is the end of the litverse after all; most of these characters could be seen again... but not Shar! There is at least a lot of Ro Laren. (James Swallow can't make the USS Robinson crew interesting, alas, but he does wisely avoid even mentioning most of David R. George's other post–Typhon Pact recurring cast. Sorry not sorry, Jefferson Blackmer. You will forever be Starfleet's worst and most boring chief of security. Your space station is being purposefully destroyed, and your boss doesn't even mention you.)

Almost all of the characters get great moments. Nog dies—but he dies saving Jake. Jake saves the lives of a bunch of civilians (and we learn he wrote a novel based on the events of Rising Son). Quark and Ro have a genuinely emotional scene about growing old together. Odo gets a couple good gotcha moments. O'Brien's death isn't as impactful as some, but he goes out swinging. I found that Swallow captured Quark best of all. I liked that it was Quark who most objected to the plan to destroy the wormhole, and the scene where he says he's staying behind to protect what's most precious to him, which Ro initially thinks means the bar, but turns out to mean her, is great. Quark always comes through when it matters, and Swallow nailed that here. (Also good: "it's very presumptuous of you to assume that I'd just go along with this, without complaint." "I never assumed you wouldn't complain.")

I wasn't a huge fan, though, of how the Bajoran prohecy was just a maguffin, only of relevance as yet another portent of the apocalypse. At its best, DS9 managed to also use stuff like that to say something interesting about faith, but that's absent here.

This book also does the "flashes of random established characters in action scenes to show the scope of the crisis" thing more effectively than Moments Asunder, with appearances by Martok and Klag, for example. No full-on IKS Gorkon, but I'll take what I can get. (Has there been a da Vinci reference in this series? I haven't noticed one yet.) We also get a lot of random callbacks, which I enjoyed because it fits the tone of the series. Who would have thought a key plot point in this trilogy would reference Keith DeCandido's The Brave and the Bold duology!? Or that there'd a passing reference to The Janus Gate trilogy of all things!?! (Had that ever even been referenced by another "novelverse" book before?) (The thing this book needed, though, to really complete the vibe was comic book–style footnotes explaining this references. Those parts should have had a note at the bottom of the page attributed to "Editing Ed" or something.)

My big complaint is that if this book is our last chance to see the novels' Admiral Riker—and especially, the novels' version of the Titan—I'm pretty disappointed with the choice to go with a "Riker is crazy and has to hunt Picard" subplot. Not only did we (as the book itself points out) do this quite recently in Takedown,* but it means we can't really get any POV scenes from any Titan characters, as it would lampshade how little sense it makes that they're all going along with Riker's plan. Disappointing, given how 1) Titan was one of my favorite novel-original concepts, and 2) Swallow is typically very good at writing the Titan crew.

Longtime readers know of my dislike of the novels' version of the mirror universe, so the end of this book has me worried, but on its own, this was about as good an installment as I could have hoped for.

Continuity Notes:

• Because of the way the novels have jumped around in time, you end up with weirdness like Rugiero being Sisko's first officer for (arguably) longer than Kira was. You'll never make me believe it.
• Here we learn Tom Paris is on Akaar's staff. Tom was in Moments Asunder, but I thought just because he had experience with this kind of malarkey thanks to "Year of Hell"; if he was working for Akaar there too, I missed it.
• Here we learn what nickname Wesley has for his half-brother René. Was this preestablished? I can't remember any meaningful interactions between the two in earlier books, but perhaps it was in Cold Equations.
• There is a dog on Picard's vineyard (c.f., Picard season one). It's funny, because I think of the "novelverse" as so separate from the modern shows, it was quite jarring every time there was an explicit reference to them: for example, that Boreth is a source of time crystals (as seen in Discovery). The most effective one is probably when you realize part of Riker's motivation is this feeling that he ought to have two kids and live in a cabin in the wilderness (Picard: "Nepenthe"); I wish the book had leaned into that more, it could have made evil Riker work.
• Given how many books were referenced here, it struck me as quite pointed that the book makes sure to establish that Saavik and Spock are good friends and great colleagues and no more, contradicting Vulcan's Heart and Vulcan's Soul, even though those books have been previously referenced by other Destiny-era books.
• Alexander and Spock previously met (for some reason) in the Prometheus trilogy. I didn't detect any reference to that here, but I also didn't notice any explicit indication they hadn't met; their first meeting here seems to be between scenes.
• When did Odo resume being constable on DS9? Is this something that happened in another book that I forgot about?
• Look, the Andorian transporter duplicates were dumb. I appreciate this book wrapped that plotline offhandedly, but it should have done so in a way that made it clear what a dumb plotline that was to begin with.
• I didn't buy how little Picard seemed to know about Bashir (p. 287). Surely there was a lot of information on him in the Section 31 files released at the end of Control, and surely Picard would have read all about him!
• Counselor Hegol dies in book I of Coda but pops up here. Whoops.

Other Notes:

• Why is the adjutant of the Federation ambassador to the Klingons a Klingon? That didn't make any sense to me.
• At the beginning of the book, a former disciple of Admiral Dougherty (from Insurrection) gets a lot of setup, but I don't really know why, because he promptly vanishes from the story in favor of the "evil Riker" approach.
• There is a Nick Locarno joke here. James Swallow beat Lower Decks to the punch!
• Jake calls Rebecca his "stepsister" (p. 151), but that's wrong, she's a half-sister.
• Endalla being destroyed in gory detail without a single reference to the "what's up with Endalla" plotline that ran interminably through DRG3's DS9 novels just feels like trolling.
• Does every James Swallow Star Trek novel have a bit where someone says "down and safe" after beaming?

I read Destiny-era Star Trek books in batches of five every few months. Next up in sequence: Coda: Oblivion's Gate by David Mack

* Funny moment: Riker points out Picard gets possessed a lot; Vale retaliates that so does Riker.