Hagen Blix

@hagenblix

Linguist, Cognitive Scientist, Occasional AI Researcher, Immigrant in NYC, Co-Author w/ Ingeborg Glimmer of 'Why We Fear AI' - out now:

2,393 Following4,129 Followers4,027 Posts

YouTube Bluesky

Is this your profile?Claim it and hagenblix.dirty.social is yours.Claim profile

Hagen Blix@hagenblix· 28d

Even Milton Friedman would go

Hagen Blix@hagenblix· 29d

"Adoption of such tools, continued Lucas, was an inevitability. 'There’s nothing you can do about it. That’s progress, it’s the future.'"

Hagen Blix@hagenblix· 29d

I present to you... General Ludd

Hagen Blix@hagenblix· Jul 14

I think this is the key passage, probably. He doesn't quite want to just agree with the AI booster perspective, but he's basically like "if they're right, maybe this will get us UBI" - which is a pretty common perspective among the "left" wing of Effective Altruism, I suspect

Hagen Blix@hagenblix· Jul 7

The Austrian's call red-brown politics Punschkrapfen, after this desert: Red on the outside, brown on the inside

An Austrian desert called Punschkrapfen, which is pink on the outside, and brown on the inside

Hagen Blix@hagenblix· Jun 16

Lovely company for Why We Fear AI at The Word is Change, Brooklyn! The Edge of Space Time by @chanda.blacksky.app Empire of AI by @karenhao.bsky.social Interface Frictions by Neir Alexander, Marsha by Tourmaline, The New Age of Sexism by Laura Bates

A picture of a book shelf at The Word is Change in Brooklyn, on which there are The Edge of Space Time by Chanda Prescod-Weinstein, Empire of AI by Karen Hao, Interface Frictions by Neir Alexander, Marsha by Tourmaline, The New Age of Sexism by Laura Bates, and Why We Fear AI by me and Ingeborg Glimmer

Hagen Blix@hagenblix· Jun 12

I don't generally post much about cooking here, but the fact that Alliinase (the enzyme that makes garlic pungent) is less active in low pH environments, and that this means that pureeing garlic in lemon juice is amazing is something I think everyone should know. www.seriouseats.com/israeli-styl...

A chart showing that Allinase activity is strongly reduced in acidic environments (with a peak around a pH value of 6.5)

Hagen Blix@hagenblix· Jun 11

Drawing heavily on the works of Simone Browne, Michelle Alexander, @timnitgebru.blacksky.app, and @ruha9.bsky.social here, where we try analyze this predictable unpredictability in terms of racist political terror. We also draw out connections to behaviorism's "variable reinforcement schedules" 4/

The old Jim Crow was obviously a system of political terror.
It reinforced a wider system of segregation and discrimination,
through violence, inflicted in both targeted and random ways. It
inherited many properties from its predecessor, plantation slavery.
Historians have only recently begun to uncover how invested the
plantation slavery system was, not just in the use of violence, but in
the scientific study of violence for managerial purposes, for extract-
ing a maximum of labor from enslaved people.15 That is to say, that
system already relied on scientific approaches to violent behavior
modification. Just as Simone Browne did, Michelle Alexander
reminds us that the system that we live in now, the new Jim Crow,
similarly has its debt to the old version. And we already know what
the old Jim Crow reinforced: a system of racist inequality. What
does the new Jim Crow reinforce? One might suspect—a system
of racist inequality.
In light of that current and historical context, and in light
of the importance of unpredictability for political terror, studies
like the ones by Buolamwini and Gebru, which show that facial
recognition systems are more error-prone for Black people than
for white people, cast a particularly long shadow. When the police
uses facial recognition, errors lead to unpredictable, random acts
of state violence. When the distribution of those errors, of that
violence is not random, but structured, affecting some people
more than others, we should take note. When the patterns in the
errors align so clearly with other racist patterns in policing and the
US system of mass incarceration, we must study that pattern, and
also wonder if the errors really are simply errors, or if they serve
a function. This is, in fact, the point highlighted by sociologist
Ruha Benjamin, who—in analogy to Michelle Alexander’s New
Jim Crow—coined the phrase “New Jim Code” for the algorith-
mic dimensions of the system. What’s new about the New Jim
Co…

Hagen Blix@hagenblix· Jun 11

Right?! Literally, what if we retold the sorcerer's apprentice, but made it about an industrial production process! We're talking about that and the anthropomorphization of LLMs being the anthropomorphization of the logic of capital right away, too (Why We Fear AI, p. 8f)

Why We Fear AI, p. 8f:

We can see this identification at play in an oft-repeated story:
the Paperclip AI. For readers who haven’t come across this little
attempt at a fable/thought experiment, a very brief recap: we
are asked to imagine a “superintelligent” AI whose sole aim is
the production of paper clips. Guided by a prioritization whose
consequences its human makers did not foresee, the AI makes
essentially the whole of Earth into a ball of paper clips, dooming
humans in their desire for office supplies. In short, the story re-tells
the sorcerer’s apprentice, and sets it into an explicitly industrial
setting. One would think a retelling would clarify the telling we
have given above—that it would make clear that the Goethe/
Mickey Mouse/Paperclip fable is not simply about tools and the
abilities they give us, but about tools as capital, and about the
pitfalls of production for the limitless drive to profit. But, alas,
the identification of the AI tool with a “superintelligence” in the
Paperclip AI version of the fable produces the false identification
we sketched above: It is not capital that is understood to be the
issue, but “superintelligence.” It’s not what we do with our tools
when directed by the endless will to profit, that appears as the
problem—it’s just the tool, whose intelligence, it just so happens,
has the very same desire that the boardrooms are cast to enact.

Hagen Blix@hagenblix· May 29

Hagen Blix@hagenblix· May 21

New google drive logo looks like it's saying "This is where you recycle your documents" ... which I guess is what google is for now

Hagen Blix@hagenblix· May 21

As I keep saying, Why We Fear AI somehow has a paragraph on everything 🤷

Generative AI will bring this kind of crowding out of the
mid-level quality products, and the (often good) jobs associated
with them to new parts of the economy. At the same time, the
pattern is a more general one that has been quite typical of cap-
italist development for decades now, and another aspect to the
deskilling drive. From shoes to tools to pots and pans—the medi-
um-quality market has been crowded out by a lower-quality level
that is often much cheaper (and most of the time simultaneously
less durable), while the higher level quality goods are increasingly
out of reach for most people. Perhaps the most obvious exam-
ple is furniture—ready-to-assemble furniture from IKEA and its
copycats are ubiquitous, but furniture that could last for more
than one move is increasingly limited to a high-end market. The
high-end market may be largely unaffected, but the mid-level
quality often cannot compete with the lower end. This is especially
true in a country like the US where the majority of people live
paycheck-to-paycheck and find saving almost impossible. Often,
cheaper goods have to be replaced significantly more frequently,
perhaps even to the point where the cheaper goods end up costing
more money in the long run. It’s expensive to be poor.

Why We Fear AI, page 104

Hagen Blix@hagenblix· May 13

Next step: Reinforcement learning from human feedback (RLHF) for the masses

Hagen Blix@hagenblix· May 13

Oh, I quite liked that! That's quite similar to our starting suspicion, too (in our case, capital as the automatic subject, Fisher's unthinkable centerlessness of capitalism, and the anthropomorphization of language - the unthinkable authorlessness of text, so to speak)

Hagen Blix@hagenblix· May 12

"We think the straight will be opened because it would be bad if it didn't" Well, that is, uhm, an interesting perspective for the head of JPMorgan's global commodities strategy. So good to know that the ruling class totally knows what they're doing, and are totally not of gambling with our lives!

An excerpt from a FT article

The Wall Street bank’s analysts said this could force an agreement between the US and Iran, despite the White House and Tehran remaining at loggerheads over a prospective peace deal.
The alternative would be a further sharp increase in refined fuel prices, which could stoke inflation globally.
“Our conclusion is that one way or another the strait reopens in June,” said JPMorgan head of global commodities strategy Natasha Kaneva, though she cautioned the market would only trust a “clear, credible announcement, ratified and confirmed by both sides”.

Hagen Blix@hagenblix· May 12

Oh, and bonus quote from the article in the OP from the tax lawyer: "'We're in purgatory. We just don't know until we see what's going on.' Klein described his clients as 'very angry', but conceded the new tax was expected to be relatively minor compared to their overall net worth."

Hagen Blix@hagenblix· May 12

3. There was this "Trump sons' war chest targets policy winners" article. "There is no suggestion that Don Jr and Eric Trump knew that Cove was on the cusp of securing a contract from their faster's US administration when they made their initial investments" ... oh?

Hagen Blix@hagenblix· May 12

2. There's whatever this was "What do you mean I'm an NPC wimp, frightened of bold strategy? Let me tell youm Claude, bold strategies backfire all the time"

Hagen Blix@hagenblix· May 12

The rest of the issue was gold, too 1. There's someone literally named Sarah O'Connor asking "Can AI help poor countries leapfrog into the future?"

Hagen Blix@hagenblix· May 5

When we were going through AI dreams/nightmares on a class-by-class basis for our book, we devoted like 3 pages on billionaires, because.. well, exactly what you said

If there is anything of interest to say about the billionaires’
dreams, other than their transparently infantile and self-serving
nature, it is perhaps merely that they range from the mundanely
apologist to the triumphantly narcissistic.

(From Why We Fear AI, page 158)

Hagen Blix@hagenblix· May 5

Sure sounds like we've got a *subprime AI &data center loan market* now

A Financial Times Article from May 4th, 2026 titled: "Banks in danger of 'choking' on data centre financing seek to offload risk"

Hagen Blix@hagenblix· Apr 27

Page 50 - also image gen, we were talking about car crashes (which.. the models like to add lingerie models to)

One might
argue that it really is the system itself that we encounter, reappear-
ing in some distilled, condensed, obscured form. Obviously, spe-
cific bots are deployed for specific purposes, and that will matter
for the way they interact with users. At the same time, all of them
share certain deep properties, rooted not in the needs of those who
deploy them, but in the technology that they are based on. An
LLM is an abstraction over the status quo (or rather, the text of
the status quo) that has, at its core, no other aim than producing
more text that is just like the text already in existence. One result
is that the dominant ideas, the dominant textual patterns of our
time, are recast as “intelligence” per se. We might call the complex
mass of concepts and ideas, the channels into which ideas flow
often and easily, or the boundaries of what is reasonable to debate
in so-called “polite society”—the dominant ideology of an age.
An age’s dominant ideology tends to presuppose that the power
of that age is natural; the hallmark of ideology, in this sense, is
that it is easy to recognize only when it’s past its time. We may
no longer think that an almighty god appointed a king to rule
us, that women are naturally unfit to vote, or that sex outside of
a heterosexual marriage is immoral. But at various times, to think
otherwise (to think of replacing the monarchy with a republic, to
affirm that women are people, or being in favor of sexual or gay
liberation) was the far more difficult, even dangerous thought—if
not to have, then at the very least to put in writing. Ideology, of
course, is never complete, other thoughts are always possible.
But ideology is easy to think, it’s what our thoughts are already
marinated in. And hence, ideology will be dominant in the textual
patterns of an age. Seen in this light, what is called “intelligence”
in a language model, is really that—ideology as a thing that can
pretend to be a person.

Even in AIs, ideology is easiest to see when it goes awry,
when the AI gets the common/ideological conjunction of things
wrong. Shane Jones, an engineer at Microsoft, produced a perfect
example: when Microsoft’s image-generating AI, Copilot Designer,
was prompted simply with the words “car accident,” it produced
women in lingerie sitting atop or kneeling by wrecked cars. The
conjunction of sexualized women and expensive cars is a familiar
trope of patriarchal ideology in the visual realm. It is embedded in
larger structures of meaning that operate in parallel, and it is easy
to use: no advertising design team would have to work very hard
to come up with yet another example of this association (whether
they have an AI helping them or not). Easy to use for people in the
“right” context—and easy to suddenly and immediately identify
as ideology when it goes wrong: the model continued to produce
sexualized depictions even in a decidedly not sales-related context.
Models are ideology, but as anyone who has used generative AI
has likely noticed—they aren’t perfect at it.

Hagen Blix@hagenblix· Apr 25

Obviously, this is where I go "You don't hate Mondays/AI, you hate capitalism". I think this misidentification isn't random. We have, after all, argued that there's very good reasons for this particular "fetish", for this particular way of identifying AI as the problem, rather than the symptom.

But: not everything is a business. Not everything is a loop! The entire human experience cannot be captured in a database. That’s the limit of software brain. That’s why people hate AI. It flattens them.

Hagen Blix@hagenblix· Apr 25

And this class- and power-blind alternative explanation of nihilism that the essay offers is itself the source of plenty of nihilism. You know, other than spending and voting, we should also hope that the rich and powerful wake up one morning and stop acting like the point of life was business.

I also think it’s incredibly important for our politicians and tech executives to make sure our political process makes people feel empowered, not helpless, which is a specific kind of nihilism they have all greatly contributed to.

Hagen Blix@hagenblix· Apr 25

This is a perfect example of the pervasiveness neoliberal nihilism, where all political activity boils down to spending and voting. Nothing else is even conceivable. www.theverge.com/podcast/9170...

f you want to meaningfully oppose AI in a way that lasts, you should speak loudly with your dollars in the market and your attention on the internet, and you should speak loudly with your votes.

Hagen Blix@hagenblix· Apr 23

This is just the weather now

Hagen Blix@hagenblix· Apr 23

Obligatory Foucault crow

A four panel web comic:
1. A crow labeled Foucault saying "things are like prisons"
2. Someone in the audience saying "Booo! Get better material!"
3. A distressed crow labeled Foucault
4. More material with stuff like "factories are like prisons, schools are like prisons" etc

Hagen Blix@hagenblix· Apr 21

Just spotted this on the shelves at The Word Is Change in Brooklyn the other day @tamigraph.bsky.social @enda-b.bsky.social @gleemie.bsky.social

Hagen Blix@hagenblix· Apr 14

Here's a nice random example of "recalcitrantly unintelligible" writing: "You know full well we are exploiters. You know full well we have taken the gold and minerals and then oil from the “new continents,” and shipped them back to the old metropolises.

"You know full well we are exploiters. You know full well we
have taken the gold and minerals and then oil from the “new
continents,” and shipped them back to the old metropolises. Not without excellent results in the shape of palaces, cathedrals, and centers of industry; and then when crisis loomed, the colonial markets were there to cushion the blow or divert it. Stuffed with wealth, Europe granted humanity de jure to all its inhabitants: for us, a man means an accomplice, for we have all profited from colonial exploitation. This pale, bloated continent ended up by
lapsing into what Fanon rightly calls “narcissism.” Cocteau was irritated by Paris, “this city that never stops talking about herself.” And Europe, what else is it doing? And that super-European monster, North America? What empty chatter: liberty, equality, fraternity, love, honor, country, and what else?

Hagen Blix@hagenblix· Apr 14

Surely, the way to solve the affordability crisis is to make workers poorer, right?!