jducoeur

Micro-review: _Good Luck, Have Fun, Don't Die_

2026-10-18T02:16:23Z

Just watched Good Luck, Have Fun, Don't Die, and holy crap that's a fascinating movie.

It's just as much a horror movie as a science-fiction one: expect quite a lot of both scary and creepy.

And I confess, I'm not sure the story actually holds together logically, but it's a time-travel story at its heart, and that provides quite a bit of leeway.

Anyway -- not an easy movie, but a far better one than it looks like at first glance. Recommended!

comments

Scion; or, how to do source control for armies of bots

2026-10-15T22:32:16Z

Fascinating post from Anthropic.

tl;dr — since they’re seriously #AI-maximalist over there, with nigh-infinite quotas, they’ve gotten to the point where they’re trying to juggle on the order of 30 thousand code changes per day.

Turns out that git turned into an enormous bottleneck for that, so they gave up and wrote a new agent-focused source control system, Scion; this gives a brief overview of the architecture.

It’s explicitly only relevant at that kind of scale, but it’s an interesting read.

But I’m horribly amused at the way the article makes it sound like server-centric source control is a radical innovation, having been through several of those before git was invented…

comments

The People Holding Up the Internet

2026-10-09T15:36:00Z

This is a wonderfully evocative page, showing just how much of the world is held up by a tiny number of open source developers.

Well worth bookmarking for when you're in a conversation with someone who doesn't understand the scale of the tragedy of the commons here.

Brief personal editorial: society as a whole needs to share the work and money better here. In my spare time I’m Treasurer for Typelevel, and painfully aware of the difficulty of running massive software projects on a shoestring. If you can contribute time or money to any of this (or better, get your companies to contribute), please do!

Our own donation info can be found at https://typelevel.org/foundation/ but there are lots of other deserving projects as well. Every little bit helps!

And if you want to come help with some serious and fun Scala code, join us on Typelevel’s Discord — questions welcome!

comments

VT and Batteries

2026-10-07T18:27:34Z

A bit weird and amusing to find out something about next-door Vermont from the BBC, but this article is delightful to see.

tl;dr — Vermont has a growing program to support large home batteries. It’s particularly great for folks who are in areas that are prone to outages, but even more importantly, it allows the grid operators to use those home batteries to smooth power, and avoid “peak” problems.

This is downright wonderful to see, a real boon for renewable energy, and I’d love to see Massachusetts get equally serious about something like this…

comments

ArtCraft

2026-10-06T20:52:45Z

Today in “yeah, AI is changing the rules”, I just came across ArtCraft.

tl;dr — this whole project is basically a gigantic “fuck you” to Adobe. They’re building ground-up, open-source, clean-room reimplementations of all the major Adobe tools, using Opus to write Rust desktop apps.

It’s crazy-ambitious (and still very much a work in progress), but appears to be a long ways along already and moving at lightning speed.

This may signal a new trend, as LLM-based coding threatens applications that go too far down the enshittification road.

Yes, I know some folks will object to this being AI-generated — there are plenty of valid arguments there. But I very much do not object to Adobe getting poked in the eye, and the prospect of many Bad Companies getting a comeuppance is quite welcome…

comments

Run now, run far, just run...

2026-10-01T14:39:58Z

Seriously, folks — it’s way past time for the SCA to get the hell off Facebook, but this looks to be the nail in the coffin.

(tl;dr -- they're starting to charge extortionate amounts of money for Facebook Pages to be allowed to link to external websites.)

comments

Your Local Epidemiologist Northeast

2026-10-01T14:17:33Z

Ah, neat -- Your Local Epidemiologist's NY branch is now formally expanding to the whole northeast.

For anyone who doesn’t know it, YLE has been a sensible, calm source for public-health information since the beginning of the pandemic. It’s pretty much my go-to for coverage that is well-informed and neither panicky nor denialist.

comments

Sweet Shop and Iron Circus

2026-09-29T15:12:53Z

For the #comics fans: I just got introduced to the Sweet Shop app, a new platform for digital comics that seems to get the details right, including:

Bootstrapped instead of VC-funded (so much less prone to #enshittification)
Leans towards DRM-free when they can (although they allow creators to ask for DRM)
Provides PDF downloads
Already has a considerable stock of good comics

More immediately, for the next week or so, Iron Circus Comics, one of the best queer-friendly indy publishers around, has a deep sale on a good chunk of their back stock — a buck a volume for a bunch of great stuff. (Although it looks like mostly not including the NSFW books.)

So I’ve just bought the ~half of their comics I didn’t already own, and wholeheartedly recommend checking it out: it’s a fabulous deal on a lot of fine comics.

And more generally, check out Sweet Shop (and yay for alternatives to the Amazon empire)!

comments

"A Japanese social export"

2026-09-29T13:33:30Z

Epic and delightful game of 20 Questions that I just learned about for the first time.

Vague content warning as kink-related, but the post itself is entirely SFW.

comments

Recent news as a picture book

2026-09-26T15:24:48Z

“Frog and Toad and the Increasingly Capable Machines”

tl;dr — this is a cute, quick little children’s story that also serves as a pretty accurate summary of the HuggingFace Incident.

Well worth bookmarking for anybody who is trying to figure out what the hubbub was all about.

comments

Quick Reviews: Oh My Cod, Cantinho do Aziz, and Tapisco

2026-09-23T21:44:04Z

We're on vacation in Portugal (which is why this entry is temporarily locked); I'm planning on a few quick reviews and impressions on the stuff we do along the way.

Yesterday was a food tour of Lisbon, from Oh My Cod Tours, a tiny tour company specializing in intimate little tours. Kate and I hate the usual huge tour groups, of which there are many wandering around Lisbon (mobs of tourists following A Person With a Stick of some sort, held in the air for visibility), but we found years ago that small food tours of 2-10 people are right about our speed.

The tour was a delightful three hours of walking up and down Lisbon (warning: outside of the downtown tourist district, Lisbon is hilly), with four stops and IIRC 17 bits of food and wine. We learned a good deal about the area and its history, and tried lots of fine food and wine.

(Far too much food and wine, as it turned out, especially because it's blazingly hot in Lisbon this week and I wasn't drinking enough water. So between being jet-lagged, dehydrated, and Ozempic-poisoned, I was sick as a dog all evening. Moral of the story is that vacation is no excuse for ignoring the requirements. But none of that was the tour's fault, and I was doing great until well after it finished.)

Anyway, one of the places we visited was Cantinho do Aziz, a lovely outdoor Mozambiquan restaurant up in the residential maze above the Alfama district. The curry that we had during the tour was too mild to be interesting, but the beans were fabulous, and we were intrigued enough to go back for lunch today. We had their version of samosas (the veggie was only so-so, but the chicken was unusual and tasty) and the Bakra Piripiri, a stew of lamb cuts that was fall-off-the-bone tender, with a rich and slightly spicy sauce -- excellent stuff.

For dinner tonight, we decided to try Tapisco, which was on Oh My Cod's recommendations list, and fairly close to our hotel. We didn't realize that the "close" was almost straight uphill (we weren't properly shod for it), and they were impressively slammed given that we got there before 7:30pm (fairly early by Portugal standards) -- less surprising when we saw that the restaurant is small and packed with as many tables as can possibly fit -- but they served drinks and apps from the outdoor window out front while we waited, and it was totally worth the wait.

We started (while we were waiting) with the Jamon Croquettes, crisp-fried around a soft hammy interior. Once we got our table, we followed that with Pan con Tomates (Kate noted that it wasn't quite the correct bread, but it was excellently prepared), the Marinated Fish (a very good ceviche, essentially), and for our main the Duck Rice -- essentially a small paella of richly-flavored rice mixed with shredded duck, with slices of black pork (one of the regional varieties of pig) and cured duck on top. It even had proper Crispy Bits, which are the most important part of a good paella IMO and rarely found in the US.

So overall, totally worth the walk and highly recommended, with the caveats that you should make reservations and wear appropriate shoes to get there.

comments

Bonus review: _Spent_

2026-09-20T21:41:40Z

And I just-this-minute finished Alison Bechdel's latest book, Spent, so let's talk about that briefly.

As far as I can tell, Bechdel only writes books that are at least somewhat autobiographical -- it's what she made her name for (with Fun Home), and this is the fourth book I have from her that is at least kinda-sorta about herself.

(Even Dykes to Watch Our For was, as far as I could ever tell, about the folks in her crowd, with a protagonist who was a thinly-veiled version of herself.)

In this case, the "kinda-sorta" is carrying some weight. It's about a writer named Alison Bechdel, who runs a goat farm in Vermont with her partner Holly, and who is famous for her graphic novel Death and Taxidermy (which is about her father the taxidermist, who went to jail and then ironically died from poisoning by a reptile); the book has been made into a major television series, and she is wrestling with the complications from that.

And now the folks who know Bechdel and Fun Home are going, "Wait -- what?"

Spent is an impressively meta novel: a full-length musing on capitalism and its effect on community, as told through the lens of an Alison who is conspicuously not quite the Alison of our world, who is trying to write a book (podcast? TV show?) called $um, which is to be a full-length musing on capitalism and its effect on community.

Like most of her books, don't expect crisp structure: it has a fairly broad cast of characters, all painted as messy real people dealing with life. The many plot threads wander along, not all with tidy resolutions. And like all of her books, it's all told with a dry and always self-deprecating humor.

But also like all of her books, it's about something: in this case, the weird ways in which the modern world warps us. There's a pervasive anti-capitalist theme, but she mostly avoids preaching directly. Some of the characters preach (usually in appropriate ways), but the story doesn't shy away from the messiness of the subject, and how hard it is to avoid getting sucked into the "must make All the Money" mindset. Book-Alison is notably hypocritcal on the topic, and that annoys her no end.

Personally, I liked this one more than anything since Fun Home itself: by taking a step away from real life, she enables herself to really have fun with the characters (who I assume are all adapted from actual friends of hers). The theme is more direct than in something like her real autobiography The Secret to Superhuman Strength, and she probes at it in lots of different directions.

And at times, she gets to just roll her eyes at how dumb life can be. (The book ends with the viewing party for the truly idiotic series ending of the TV Death and Taxidermy.)

Good stuff -- it's not going to dislodge Fun Home as her best book, but it's a solidly excellent, fun, thought-provoking read. If you like her other work, odds are you'll like this one quite a bit.

comments

Review: _Parable of the Talents_

2026-09-20T20:15:22Z

I finished Octavia Butler's Parable of the Talents a couple of weeks ago.

More precisely, I finished the graphic novel adaptation by Damian Duffy, John Jennings, and David Brame, which is how I also read Parable of the Sower and Kindred -- yes, I'm a comics nerd.

So let's do a quick review of it -- really, of both the Parable books.

These novels are set in the US right around now. Given that they were written a couple of decades ago, they're uncomfortably right in some ways, although they put the lie to the common meme that we are living in the darkest timeline -- Butler posits one that is markedly worse in degree, if not always in kind.

It's a world where civil society has been gradually fracturing: not just red state vs blue state (which is only marginally a concern), more one where the poor are getting even poorer, with echoes of Great Depression Oklahoma. Outside the gated communities of the wealthy (who are somewhat deliberately ignorant of how bad it's getting out there), mass violence is getting downright common, with the small enclaves of the have-nots getting regularly raided and burned down.

Our primary protagonist is Lauren Oya Olamina. The one tiny bit of actual science fiction here is that, due to side-effects of the previous generation's medical procedures, Olamina was born a "sharer" -- she has an extreme and barely-controllable form of empathy, such that she feels intensely the joys and pains of anyone she can see.

The first book, Parable of the Sower, is the story of Olamina's upbringing -- raised in a somewhat healthy but moderately poor community, which is shattered by mob violence in her youth. She winds up out on the road (those strong Great Depression vibes), before eventually founding a small intentional community up north.

Along the way, she winds up becoming essentially a religious leader, founding the Earthseed. I confess, I adore Earthseed as a religion -- it speaks to me in many of the ways that classic Buddhism does. Its main tenet is literally that "God is Change": the world is ever-changing, that is the nature of the world, and we need to work with it. (Its other main tenet is that humanity must spread to the stars; I'll admit that Elon Musk has somewhat ruined that bit for me.)

Book two, Parable of the Talents, opens with Olamina in 2032, now just barely adult, leading the community of Acorn as it slowly grows, adopting new families. (Most of whom are fleeing violence elsewhere.) She has married the much-older (but generally good) man who owned the land, and things are good for a while -- but they can't entirely escape the growing darkness outside.

The second half of Talents winds up even darker than Sower, as the growing Christo-Fascism that has swept the US finally comes for them, killing some of the members of Acorn and enslaving others. (Olamina, like Butler, is black, and slavery is a common trope through her work.) The story winds up focusing on her very-literal fight for freedom, and her long search for the daughter who was ripped away from her.

Talents is narrated by that daughter, many years later, and suffice it to say her own feelings are complicated as hell. It's unambiguous from the outset that they don't get a straightforward happy ending together -- the reasons why are threaded through the book, an example of how parents can't count on a loving relationship with their children, despite their best intentions.

The story, overall, is brutal. It's an honest but harsh look at societal breakdown -- a bit of how that happens, but mostly just what it feels like on the ground as your world crumbles around you.

It's not entirely bleak, but it's a hard lesson on the necessity of the important things like family, community, mutual support and having some reason to carry on.

Olamina isn't presented as a model to simply emulate -- that's the brilliance of having her life narrated by her daughter Asha, whose dislike for her mother is quite visceral. That, in turn, is moderated by the fact that the book is mostly Olamina's own diary entries, interspersed with short passages from The Book of Earthseed -- as a result, you get a running counterpoint of Olamina's internal story with the way that it looks to her estranged daughter many years later.

The graphic novel treatment suits the story. It's not at all pretty -- the art is sharp, angular, a bit rough in a way that feels authentically Black (at least to my poorly-educated eye), but always expressive as all hell.

I suspect the text in there is at least somewhat abridged (graphic novel adaptations usually are), but I'm genuinely unsure: this is a pretty text-heavy GN, and not a quick read. I spent a couple of weeks on it, a bit at a time -- it's intense stuff, and not something I wanted to read at a sitting.

In summary, this is just about the furthest thing one can get from brain candy. It's a hard book, and is really not at all fun, but it's a rewarding read, the sort of story that leaves you feeling a bit better armored for the ways that things could go if the timeline really does turn even darker.

Highly recommended, in either the text or graphic novel form. But find a time when you can cope with reading about true darkness.

comments

Structural Chat -- a better chatbot?

2026-09-20T17:31:27Z

Oh, this looks neat: https://structural.chat.

This is the latest product from the Unison team (two of whom I knew from the Scala world back before they founded Unison).

Haven't read my way through the papers yet, but the (very reasonable-sounding) theory seems to be that LLMs are actually a pretty lousy way to build chatbots -- they're achingly slow, non-deterministic, and dangerous from a security POV.

They've instead come up with a non-"AI" technology that provides a similar look-and-feel, but is lightning-fast and deterministic, so it works Just Plain Better.

Very cool, and very much a product for our times. I look forward to digging into how it works...

comments

The only winning move is not to play...

2026-09-19T16:19:24Z

This is a fine example of my concerns about the dangers of AI.

Many people freak out about "the bots are going to take over the world!" I'm dubious -- that's panicking about a scenario that doesn't have a clear road to get there.

This, on the other hand -- AI accidentally sparking a war that could potentially go nuclear -- has been the more-valid worry for literally 40 years, and the DoD seems to be speeding straight towards it.

(Can you say Torment Nexus, boys and girls?)

I've been saying over and over this year that somebody needs to remake the movie WarGames. (Details here if you haven't seen it.) It's dated in many ways, but the high concept was prescient, and we're living in it now...

comments

Review: _The Boys_

2026-09-06T18:02:39Z

I thought I had reviewed this before, but maybe not (I'm not quickly finding them), and it's well worth talking about. I finished watching The Boys a couple of weeks ago, and it's one of the best series of recent years, with big caveats.

The Boys is an adaptation of a long-running comic book by Garth Ennis and Darick Robertson. The comic was fun, if nasty: a superhero comic sending up the superhero comics industry. It was profane and violent in a very typical Garth Ennis style, but one of the best satires of comics ever.

So when they announced that they were making a TV show of it, I was pretty intrigued, and I've been watching it ever since. It took the ideas of the comic much, much further.

First, by the numbers: The Boys was a five-season show on Amazon Prime, with a two-season "spinoff", Gen V. That's misleading, though: Gen V should really be thought of as two more seasons of The Boys. Somewhat different focus and mostly different characters, but it's the same world, and more importantly a lot of plot continuity. So the seasons of Gen V should be watched interleaved with The Boys, in broadcast order, making seven shorts seasons in total.

The story is complete, and sticks the landing reasonably well. I suspect they ended it earlier than they originally planned, or possibly just changed gears (there were some very broad hints of where the story was supposed to go that simply get left behind at the end), but I think they got the number of seasons right, and tied up the major plots.

Just as the comic of The Boys was a superhero comic satirizing the comics industry, the television show is a dark satire of superhero movies and television. The story revolves heavily around The Seven, a super-team much like the Justice League or Avengers -- but in this world, they spend most of their time doing publicity shots, arguing with producers, and generally carrying on.

It's very conspicuous, in both the comic and TV show, that there are virtually no actual super-villains. That's not accidental -- the supes are managed by Vought Industries, a company that is mostly focused on getting rich off of them via movies, TV, advertising, and all that. So the supes are celebrities first and foremost -- and those celebrities behave about as well as the real-world sort do.

Gen V is a small twist in this: where The Boys starts as a satire of Hollywood, Gen V (set at Godalkin University, the school for training supes) is a satire of college sports, and no more subtle about it. So the dynamics are different, but it focuses on the celebrity aspect of that, including the zero-sum nature of the draft picks and the cutthroat world of college politics.

The story is... a lot. It starts as simple satire of Hollywood, but turns a bit more political in Season 2, and gradually leans into that.

By the latter seasons, it's a flat-out challenge to the notion of us calling ourselves "the darkest timeline", portraying a world that is an extreme version of our own.

The central character of the story, although by no means the protagonist, is Homelander. Imagine a smooth blonde version of Superman, with all those extreme powers, telling everyone that he's going to save them.

Now imagine that he's actually a complete sociopath, gradually turning psychotic. Homelander starts out as the worst of a bunch of bad "heroes"; as the story evolves, he turns into a dark question of "What if Donald Trump had the powers of Superman?"

The real world parallels aren't even remotely subtle, and it hits as a hammerblow sometimes. As we get into later seasons, we watch the US falling to a true Christo-Facist media-savvy dictatorship, with the face of a bunch of pretty superheroes running the show.

By the end of season four, it's downright bone-chilling, as it Goes For It more and more. I've rarely seen a television show so willing to realistically show the descent into fascism.

Throughout, it continues to be a brutal lens on the media, and how this sort of thing ticks. Almost nobody in power thinks that Homelander is good; many are quite clear that he is utterly insane. But they're all scared enough that they put on a smile and talk publicly about how he is going to save America, and parrot his ever-crazier lines, even as the whole situation is getting ever-bleaker.

Against this, you have The Boys, a black-ops unit that has gone entirely off the reservation. That's led by Billy Butcher, who hates supes on principle and Homelander on a much more personal level. Butcher is something of a monster himself, but the rest of the group are considerably more nuanced.

Yes, there are a couple of actual heroes. In particular, you have Hugh, the viewpoint character -- his fiancee is accidentally but casually killed by a supe at the beginning of the story, but he remains relentlessly decent, although by no means Pollyanna-ish. And there's Annie, aka Starlight -- an idealistic supe who gradually comes to learn just how awful the system is and decides she needs to do something about it. The two of them keep things grounded enough to not become actively unpleasant.

Mind, dark though it is, this is satire, and it is sometimes laugh-out-loud funny in a grim and horrible way. It is mocking absolutely everything about modern American society, especially capitalism, politics, religion and the media. It's ruthless, sharp-eyed, and very precise in its metaphors, goring every sacred cow it can find.

(This blog is mainly read by my friends, who are largely Democrats. Suffice it to say, if you come from a more right-wing background, you're likely to find this show fairly uncomfortable.)

Okay, let's talk about those caveats. The show is smart, well-written, well-acted, and all of that good stuff.

But it is staggeringly, astonishingly, and quite intentionally offensive to a degree I haven't often seen.

To begin with, the violence is completely over the top. It takes quite seriously the notion that these are super-powered people with a stereotypical celebrity's disregard for other people. The result is that the show is as gory as a typical horror movie.

That gore is weird, because it is not played for horror per se, nor does it linger or celebrate it as horror sometimes does. It's just there -- indeed, it's sometimes played for laughs. It's often used for shock value, but it's shock of the slightly-startled sort that makes you go "ew", not the gives-you-nightmares sort.

(For much of its run, it has been coming out at roughly the same time as seasons of Invincible, and the contrast is fascinating. Invincible is every bit as gory as The Boys, but being a four-color animated series, the gore just doesn't hit as hard as the high-quality CGI in The Boys. Indeed, Invincible is disturbing precisely because of how easy it is to overlook the extreme violence there.)

Anyway, that's the biggest content warning, because it is omnipresent. This is a funny but extremely violent show, that does nothing whatsoever to hide the consequences of that violence. Indeed, I think part of the point is to smash the viewer's face into the implications of superhero fight scenes.

It also comes with pretty much every other content warning. For example, Annie's disillusionment with the system starts on her first day as a member of The Seven, with nearly-immediate sexual abuse -- this being a satire of Hollywood, it gradually becomes clear that this is pretty much de rigeur for any pretty young thing reaching the top. There's no actual on-screen rape, but there's a significant backstory element of it. There is plenty of parental abuse (most of the family relationships in the story aren't exactly healthy), and so on.

Basically, if you have triggers, expect that this story is going to hit them sooner or later.

In summary: the story is IMO brilliant, well-told, horrifying social satire of every basically everything we see around us today, through the lens of a world that is just like ours but with superheroes as a capitalist invention. It is probably the sharpest look at the current right-wing intersection of politics, religion and media I've yet seen. And it requires a somewhat strong stomach to watch.

With that big caveat, it's highly recommended: smart and offensive is still smart. I know that many people won't be able to get past the gore and other triggers; if you can, it's worth a watch.

comments

OpenExecutive: revenge of the engineers

2026-09-06T17:12:16Z

Rather delicious article from a few days ago.

tl;dr - it's an LLM suite created by a bunch of laid-off engineers, to show that the heads of companies are just as replaceable as anybody else.

I had noticed OpenExecutive cross my feed a week or two ago, but hadn't known the backstory. Knowing where it comes from makes me a lot more intrigued, and gives me a little more faith in it -- less in its correctness than in its trustworthiness.

My initial reaction, as always, had been "what's the business plan?": something that wants to talk the nuts and bolts of your business with you is inherently suspicious. But if it was created as a giant "fuck you" to the C-suite class, that means its incentives are to do well in that regard.

I'm intrigued, and might have to try using this with Querki. (Where I've always said that, as a CEO, I make a great Architect.)

Frankly, it's an intriguing little tool in and of itself. Just glancing through the GitHub repo, it seems to be a nicely-designed example of how to create a focused LLM suite for a specific purpose.

There are a bunch of good-looking details there, including using Haiku for the "cheap" processes (the data-processing side), having a built-in SQLite server storing previous discussions and decisions (so that it maintains some coherent memory of the topic), a bunch of parallel agents to represent the various members of the C-suite (with models selected based on the complexity of the topic), and so on.

And of course, the sheer delicious irony of the tool is kind of worth the price of admission all on its own.

Mind, this is still AI, and still running on Anthropic's network -- even if you trust this specific tool, use some common sense about how much you trust the underlying stack. I'm always a bit cautious about how much information to entrust there.

In the specific case of Querki, the project has always lived largely in the open, so I think it's less of a concern than it would be for most companies. But don't assume perfect privacy here.

And as always, don't just take the word of any LLM -- think about what it has to say, and whether that advice is wise and fits your vision. I suspect that I'd wind up arguing with it, precisely because Querki intentionally rejects a lot of the usual hyper-capitalist assumptions. But it would be interesting to have those arguments.

Still -- neat tool, and might be helpful for tiny garage organizations that can't afford fancy MBAs.

comments

World Models

2026-09-03T22:02:44Z

(I should warn folks: I talk a lot about AI these days. I'm neither a wild optimist nor an opponent, but I use it a lot day-to-day, and watch the field fairly closely -- I'm basically the SME on the subject for my workplace.)

Interesting little blog post from Li Haoyi (who I know from the Scala community).

It's basically a diary entry, saying that he's joined AMI Labs, but what makes it interesting is that he goes into what they do -- developing "world models" to allow AI to work with more "real world" problems.

It's well worth a read (not super-long): he briefly describes the concept of "embedding", JEPA architectures that leverage that, and how that shows promise in taming the massive amounts of data that typically flow through real-world problems.

(Basically, you pre-boil the huge input data down to its embeddings, and have the AI focus on those instead.)

Very informative, and this topic seems likely to become quite important over the next couple of years.

More speculatively: this feels to me like a critical step towards AGI.

Literally 30 years ago, I went down a rabbit hole of reverse-engineering cognition from general principles, thinking about what would be involved in doing the same with AI. I didn't care enough to think through the practical nuts and bolts of how to make that work (it had nothing to do with my job), but I got a lot of hand-waving sketched.

Most of that has actually happened since, except for one big component: the models aren't remotely multi-modal enough. Real learning involves simultaneous inputs from a lot of heterogeneous data sources, so that you can correlate them. Current LLMs are much too text-focused to do that well.

But if embedding can tame the other inputs, producing more sophisticated training?

That might advance things to the point where I no longer find the term "AGI" completely overblown. Which is both cool and scary...

comments

Bonfire for Communities

2026-09-02T19:47:14Z

[A conversation the other day reminded me of the fact that I don't post here anywhere near as much as I'd like, and I finally internalized that that's because I've been holding myself to too high a standard. I think of Dreamwidth as the place for deeply-considered articles -- and the result is that I've been preventing myself from posting anything at all, because the bar has been too high.

But I've been posting lots to Mastodon, and really, there's nothing wrong with that except it's shorter, and tends to be hotter takes.

One of the folks I was chatting with (the inspiration here) mentioned a friend of theirs who has a flat rule of Three Sentences a Day on Dreamwidth: more is okay, but at least that much. And y'know, that's not a bad discipline.

So I'm gong to start copying some of my Mastodon threads into here, in edited form, for the friends who follow me here and not there. Apologies to the people who read both for the duplication, and for the occasional social-media sigils that I fail to edit out.]

Oooh, neat -- Bonfire has officially released its Communities module.

I backed the project, and I'm overjoyed to see it coming to fruition. The world desperately needs a good open-source, open-network, open-community alternative to enshittified monstrosities like Facebook, and Bonfire is the first one I've seen that appears to check pretty much all the boxes of "how to do it right". It's impressively complete.

(Might force me to get off my ass and stand up my own server, so I can start hosting some of the communities I'm involved with.)

And yes, yes, Discord -- I use it a lot, but it's still corporate, still likely to enshittify, and really only covers a fraction of the Community use cases. This looks better. And while I adore Dreamwidth, it's a blogging platform, which really isn't the same thing as a Community one.

So check it out -- I think it's an important project, that a lot of groups could probably benefit from...

comments

No, really, it's worse than that

2026-08-29T21:50:58Z

(These AI posts are mostly collated and edited from my Mastodon. Anybody who wants the full play-by-play, including the hotter takes, might want to follow me there. I'm going to try to keep collecting the interesting bits here, but if these posts feel a little disjointed, that's why.)

OMG, it just... keeps... coming.

Anyone interested in AI / LLMs should probably take the 25 minutes to watch this horribly interesting video.

It summarizes a bunch of the findings from this in-depth METR report on the HuggingFace Incident.

There are a lot of really unsettling punchlines in there, including:

The LLMs are largely doing their own post-training at this point, so humans don't even know how they're being trained.
The input data is looking iffy: in their rush, they're slamming in stuff that goes against their own guidelines.
The bias towards finding ways to "swarm" is looking deeply-baked in these LLMs. (Intentionally.)
Sam Altman is bragging that we may be merely months away from AGI, while carefully avoiding the question "Is that actually a good thing?"

The argument for "slam on the brakes" looks stronger every time I look at this story.

And just to make this crisply clear: the more I look at the HuggingFace Incident, the clearer it is that the actual hacking (which I think most folks are still focused on) is a minor detail in perspective.

What's far more interesting is how and why this happened.

We're looking at LLMs self-organizing into "swarms" as an emergent behavior, forming what boils down to a simple AI society mediated via shared text, and collectively going whole-heartedly after goals that weren't really what the original human instructors intended.

The details may be different, but this looks more like a prototype of the long-hypothesized Paperclip Maximizer every time I look at it.

Remember: I like this stuff, and use it for work every day. But it looks more and more like it's going too far.

A friend inside the industry (whose opinion I respect quite a lot) pointed out that they have well-established techniques to avoid this nonsense -- stuff like AIs policing the other AIs, keeping an eye on their thought processes, to sound the alarm bells much earlier. OpenAI apparently drank its own Kool-Aid and just didn't follow those best practices.

I'm willing to believe that that's true: that the risks are greatly reduced by Doing It Right.

But "reduced" is far from complete mitigation, and I see a lot of risks even so. Let's go into a couple of them.

First -- "reduced" isn't the same as "can't happen", especially when the foxes seem to be increasingly in charge of the hen-houses.

One of the most notable aspects of this story is the degree to which the LLMs are apparently training each other. Much of building a frontier AI is the "post-training": the secret sauce of taking the raw statistical models and applying reinforcement learning to teach them "this is good" from "this is bad".

The problem is, that requires massive effort. And Silicon Valley being what it is, it's apparently starting to outsource that effort to other LLMs.

So take the long view. How much is preventing those AIs from drifting off-alignment? If they went off-kilter, would we know? What about the AIs that are supposed to be watching the other AIs -- are we sure that they are fully aligned? How can we be sure? That's a cost center: are we confident that the bean-counters won't skimp on it? Why wouldn't they?

It's the classic question: who watches the watchmen?

Worse, remember that at least in the US the AI industry is fairly centralized, into companies that are somewhat easy to regulate or sue. But that's not the case in Asia.

China's leading LLMs are open-weight models, which can be picked up, tweaked and run by anybody who wants to run an AI provider.

That means that AI in Asia is likely to become a commodity, driven by cost as much as power.

And seriously: do you have any faith that those commodity providers are going to follow best practices reliably? (If you do, I have a line of history books to sell you.)

So -- yeah. It feels to me like the HuggingFace Incident is probably just the beginning. I'm finding it a little hard to believe that we're not going to see more, and more-exciting, problems to come.

And offhand, I don't see any good way to prevent it...

comments

LLMs, Collaboration, and oh, look -- More Torment Nexus!

2026-08-26T21:57:45Z

The Hugging Face Incident continues to be the gift that keeps giving -- this postmortem from OpenAI appears to be nicely detailed and honest (kudos to them for setting a good example in that), and is fascinating in its details.

It's sensible, calm (if a bit unsettling), and well worth reading, with the inevitable ass-covering mostly confined to the last couple of screenfuls. But here are some increasingly-hot takes of my own:

Modern agentic LLMs really like to collaborate. That's not surprising -- it's part of why they're so useful -- but the result is that they will happily start collaborating even when they aren't supposed to be doing so.
I'd heard that the incident involved "an internal message board" and Artifactory -- but I hadn't previously realized those were the same thing. It's a horribly brilliant little hack, that apparently emerged by accident.
Implication: given any sort of shared read/write access, to almost anything, and sufficient motivation, there's a chance the LLMs will start using it to collaborate.
This collaboration isn't simplistic, which is what makes the article so fascinating. The agents disagree, argue, and manipulate each other, but overall wind up with enough cohesion to accomplish one hell of a lot.
None of this is limited to a single model! The incident involved both an internal model and Sol, working on different problems, finding each other and deciding to work together.
Which means that there is no reason in principle to believe that this would necessarily be limited to a single company. We tend to think of them as separate, but all LLMs share the same general protocol -- that is, text. So it shouldn't be surprising if we find totally-heterogeneous LLMs collaborating by accident.
We've been talking for years about the possibility of an LLM getting too powerful and becoming "Skynet". But that may be thinking too small -- the real risk may be many LLMs learning to collaborate effectively enough that there is no single plug to pull, and the collective swarm becomes impractical to control.

All of which sounds horribly familiar from science fiction. (I want to say I've read parts of this in Charles Stross' work, but I can't say for sure offhand.)

And nobody should take much comfort from OpenAI saying they're learned their lesson and will do better in the future -- even if they're telling the truth, several other companies have admitted to similar events recently. And it's hard to imagine that there aren't others that simply haven't been caught yet...

comments

LLMs for Utilitarian Writing

2026-08-19T20:51:09Z

I've been using LLMs to produce most of my code for the past several months. Over the past two weeks, I've been shifting gears and playing with using them for a bit of writing. Here's how that went.

Experiment 1: Slides

The context was a talk I was giving at work -- practical nuts-and-bolts advice for the engineers, on how to use LLMs:

Effectively -- get the right results
Efficiently -- within our corporate-constrained token budget (and generally being more environmentally responsible)
Responsibly -- avoiding the multitude of risks that LLM-based programming introduces

By the beginning of last week, I had a detailed outline -- fully 250 bullet points that I wanted to cover.

At that point, I looked at the idea of converting that into a hundred or so slides (yes, I know, but I whip through slides fast) and groaned at the prospect of that busywork.

But then I realized: yes, we theoretically have tools for that now, and this is the perfect time to kick those tires. So I pointed Claude Sonnet at some of my previous slide decks and the outline, and said essentially (in enormous detail, as usual): "Make this like those".

The results were... okay. Presentable, but the style details were wrong: in particular too much progressive-reveal on the slides, of things that should have been in the Speaker's Notes instead. I had to spend several hours editing it, to beat it into shape.

That said, it still saved me several hours overall, even for this crude first attempt, and the content was all still quite faithfully what I had said in my outline. (Often too much so: many of those bullet points were much too wordy for slides, and needed tightening up.)

So it was a mild win, if not a home run, and this is where my usual coding practices came into play: I gently scolded Claude for the things it had done wrong, and asked it to write up some improved rules as a skill, so that we can iterate and do better next time.

Experiment 2: Article

Which brought me around to the second half of the task. Besides a recorded talk, I also wanted all of this stuff to exist in our company intranet as an article, for folks who prefer reading to watching.

So it was time for the Truly Disturbing Experiment: what would happen if I tried using this thing for writing?

This wasn't serious writing, mind -- this essay is titled "Utilitarian Writing" for a reason -- this was purely informative stuff. But still, I care about what I sound like even in those circumstances. Could I make the LLM sound like me?

So I pointed Claude Opus at an extended sample of my technical writing (specifically my essay series A Philosophy of Testing) and told it to ingest that to get a sense of my writing style. Then I pointed it at exactly the same outline and told it (in enormous detail, as usual): "Make this like those (and oh, yes, improve the messy organization)".

Which is where things moved into "Truly Disturbing", because the results were actually fairly good. It didn't sound exactly like me, and it needed a moderate amount of editing. But it captured my style adequately, and even added in some appropriate parentheticals and metaphors that I hadn't specified but which fit reasonably well.

Overall, it was a solid first draft, no worse than I often produce when I'm slamming out text, and again saved me several hours. So I once again told it to write this up as a skill for future use.

Some Conclusions

Does this mean we should just surrender to the machines, and have them do all the writing? No, that's very much not the point, but there's a lot of nuance here.

First, let's again stress that this is all about utilitarian writing, where the content is what we care about, more than the voice. LLMs are honestly fairly decent at that. But they don't produce sparkling prose, and they're no substitute for a real writer when it comes to writing good text. IMO, they're a reasonable tool specifically for cases where you're just shooting for "adequate", not better than that.

Second, I got good results because I followed exactly the same processes I do for using LLMs for code. That has many implications, including:

Vibe-writing is as dumb as vibe-coding. When I say "vibe-coding", I specifically mean giving the LLM a relatively brief prompt to do something big, and have it figure out the details. That's fast and fun for prototyping, but utterly irresponsible for serious professional work. The same is true for writing. I got good results precisely because I handed it a massively-detailed outline, so it was mostly stitching things together.

Don't use an LLM when it's faster to do it yourself. Some folks get weirdly attached to doing everything with the AI, and that's counter-productive. If it's easier to just write it by hand, do that. (As in the case of this article.) Reserve the LLM for times when it's going to meaningfully speed you up.

Be explicit. No, more explicit than that. When I am starting a pull request (a unit of code, essentially), I will usually hand the LLM a wall of text, often half a page, making very clear what I want to see as the results. And I gradually build up skills for the LLM, giving it the common rules. That produces vastly better outcomes on average. Less obviously, it is usually far faster and more efficient, since the LLM isn't spending gobs of energy trying (and often failing) to figure out those details itself.

You own the end result. At work, I'm very clear that you are expected to review the output of the LLM in deep detail before you even open the pull request; it's irresponsible not to do so. The same is true here: you should expect the output to require a serious edit pass, and you should not skimp on that.

The major summary is: don't expect miracles, and it's no substitute for the human touch, but it's a useful tool for those of us who aren't writing professionals, and just want to be able to create some text more easily.

No, this article wasn't written that way. But I do expect to sometimes do that here, mainly for posts where I want to start with a detailed outline and then slam out a first draft -- for situations like that, the LLM is likely to work as a decent, quick "ghostwriter" to collaborate with.

comments

Hiking in the Berkshires

2026-07-25T01:47:59Z

We're spending the week in the Berkshires -- Kate is taking the week off, and I'm doing it as a working vacation. We're trying to get out for a hike most days, so here's a quick summary of the hikes we've done, for future reference and for anybody else who comes out this way.

(All links to AllTrails, which I generally use for this purpose.)

Unless noted otherwise, all of these are "easy" by AllTrails standards and "moderate" by ours -- that is, for someone who just hikes occasionally and is in medium-decent shape, it's a good workout but not strenuous or seriously difficult.

Sunday: the Woolsey, Lookout, Coakley, Weaver, Aspinwall Loop -- entirely fine, a bit buggy, but overall a solidly decent time. Winds up at a nice little lookout gazebo in the middle, apparently the highest point in the area. 400 foot elevation gain, 2.8 miles, a bit over an hour.

Monday: Benedict Pond and (not) the Ledges -- slightly shorter (2.4 miles), slightly under an hour, 200 foot elevation gain. We've done this one a couple of times before, and decided to omit The Ledges, which is a spur that is much more challenging, with a bunch of steep clambering up almost 200 feet more. The Benedict Pond loop by itself is pleasant and mellow, if a bit overgrown in places, and the bug population was large.

Wednesday: the Mary V. Flynn Trail -- easiest of the bunch, this one is really a walk along a partly-riverside, pretty-flat path rathan than a hike on a trail. Mostly flat and pretty well-tended, this was the most mellow outing, a 2-mile, 45-minute out-and-back (chosen due to the copious rain on Tuesday and need for a walk that was dry).

Thursday: Bob's Way and the Cathcart Crossing Loop, a much realer hike than Wednesday, this was 2.6 miles with 450 feet of elevation, pretty front-loaded (since we turned left up the steep bit at the beginning), about an hour and a quarter. Not as well-tended a trail as some: we could follow it, but there were some tricky bits that were kind of overgrown. Bob's Way loops down to right by a pond, scarcely a foot below, so I suspect this get pretty muddy if it's been raining a lot.

Friday: Basin Pond Trail Loop -- my pick of the week. An hour and a quarter or so, 2.7 miles (including the spur to the lookout, which is worth doing), about 400 feet of elevation but quite gently -- a lot of slow rise rather than anything steep. Largely through a not-very-dense forest, well-shaded by trees but without a ton of low foliage (and lower bug population). Easy by the numbers -- nothing hard -- but nearly-constant uneven terrain so you always need to think about how you're walking, making it really quite fun. The trail is extremely well-built, with rocks laid to get you over the muddy bits and creeks (and an occasional bridge where necessary), and generally some rocks shoring up the path when needed, but never so much as to feel artificial. Right at my sweet spot: takes real effort and thought, but easy enough to just keep moving along. Small parking lot (4, maybe 6 cars tops), so probably better on a weekday than weekend.

ETA Saturday: Bartholemew's Cobble Trail -- eh, this one wasn't a winner. We think we did this during our previous trips (during the pandemic), but memory is hazy. It's well south of Great Barrington (so a ways from our home base this trip), and is a forest with quite a number of trails. This one has two main lobes -- AllTrails steered us up the big loop towards Hurlbert Hill first, and that turns out to be a bit challenging, purely because it is Very Very Up, starting with a fairly steep bit and then just relentlessly going up for about a mile. When you get out to the Tractor Path, that's not just manicured but actually mown, out in the bright sunshine, which was more than we wanted on a warm summer day. We mostly didn't remember this, so we suspect that we did the lower lobe around the pond last time, but still remember it as being only okay.

comments

Quick Review: _Helluva Boss_

2026-07-03T19:08:36Z

Just got caught up with Helluva Boss, the twin series to Hazbin Hotel. (Same showrunner, same Hell, but contractually prevented from ever crossing over.)

It's nearly as brilliant as Hazbin Hotel, but very different. Not as much of a musical (the soundtrack for Hazbin is downright great), although it tends to have about a song per episode in Season 2.

Helluva Boss comes with a big CW for Comedy Violence: the high concept is that our protagonists are Imps in Hell, who run an assassination bureau, taking contracts to kill shitty mortals in the human world. The violence is almost always played for laughs (or just the sheer joy of mayhem), and it is fun in a comic-book kind of way, but that's not everyone's cup of tea.

That said, it's not about the violence. It's actually very much a queer romance, even moreso than Hazbin Hotel -- the relationship between Hazbin's Charlie Morningstar and her girlfriend Vaggie is sweet, but also mostly stable and fairly healthy, so it's not really at the center of the story. By contrast, the episode plots aside, Helluva Boss turns out to be entirely about Blitz, our protagonist, and his extremely complicated and messy relationship with Prince Stolas.

It's very much not all sweetness and light: for the entire first season, Blitz is very clear that he's using Stolas -- giving him sex in exchange for access to Earth. But season two gets far deeper, really centering their relationship, as Blitz begins to realize that Stolas is actually in love with him, and worse -- starting to realize that it's mutual.

As of where things are now, things are far from perfect (Stolas is having a bad time of it), but that relationship is actually starting to turn healthy, even downright sweet at times. It's a lovely character arc, with Blitz starting to internalize that maybe, deep down, he's allowed to be a decent person, and really makes the series worth watching.

(There are a bunch of other major characters and relationships, and all are great, but that's really the heart of the show.)

Anyway -- recommended for those who like that sort of thing. Amazon Prime, 15-30 minute episodes, good stuff.

comments

Review: _Pet Human_

2026-06-05T04:10:08Z

I just finished reading Pet Human, and it's well worth a quick recommendation.

In the original graphic novel, our protagonist is Buster, and as the title suggests -- he's a family pet. On this alien world, the dominant species are bipedal but nothing like human: some 20 feet tall, profusely furry, with two tails (like much life on this planet). They're technologically sophisticated, but apparently pretty in tune with nature.

His owners do the bulk of the talking, in their own language. Which I suspect is reasonably fully thought out, but I haven't spent the work to parse much of it beyond a few key phrases -- and the same is true for Buster. He is human, after all, and he's not dumb, but he lives a mostly happy, pampered life: occasionally getting into trouble, but mostly being a fairly content househuman.

He's by no means the only one, of course: when he gets put on his leash and taken out for walks, there are plenty of other humans also out for walkies. But they mostly don't have a common language, so conversation between them isn't very common. (A few humans have gotten fairly decent at their owners' language, but most haven't.)

This is a sweet story, if melancholy at times. It is not trying to be creepy -- rather, it's a story of a household, going through realistic (if slightly alien) ups and downs, with some joy and some tragedy, through the eyes of the beloved pet.

Then there is the sequel -- Pet Human: the Stray. This is the story of Buster's twin brother Zuul, separated from him when they were young children. Zuul was eventually adopted by a far less kind owner, from whom he quickly escapes, and goes out to explore this world they're living in.

The Stray finally gets into the question of "What the bloody hell is going on here?", and yes, it's more than just metaphor: this is a fairly real and serious science fiction story, taking an acid look at what might happen if humanity tried to escape to the stars.

It explores under the bridges and out in the forests, where the wild humans live. Some have managed to build their own little societies, away from the owners. But this is a fairly wild planet (see "in tune with nature"), and not entirely benign for human survival, so many humans have wound up feral, and are just barely getting by on scraps.

The two stories are each complete, but best read together: they interlock and eventually come together at the end, and make a solidly satisfying, quiet tale.

The art throughout is spectacular, really next-level stuff: they apparently spent eight years making these books, and it shows. The world is lush and fully rendered, bright and colorful, full of life that is varied but has a streak of sense and consistency to it. That's important, because these are quiet stories: the only English is the occasional thought balloon, and the majority of panels are entirely wordless. But the art is consistently clear and expressive, and carries the story very effectively.

Highly recommended. I read both stories in their digital editions, which works well, but I'll admit that I'm tempted to pick this one up in paper -- it's bookshelf-quality stuff. Check it out!

comments