Anthony Moser
banner
anthonymoser.com
Anthony Moser
@anthonymoser.com
(He/Him) Folk Technologist • [email protected] • N4EJ • http://www.BetterDataPortal.com • baker in The FOIA Bakery • http://publicdatatools.comhttp://deseguys.com • #1 on hackernews when you search for "hater"
Pinned
I considered writing a long carefully constructed argument laying out the harms and limitations of AI, but instead I wrote about being a hater. Only humans can be haters.
I Am An AI Hater
I am an AI hater. This is considered rude, but I do not care, because I am a hater.
anthonymoser.github.io
variable reinforcement learning, that's the term
Also works the other way around, when you think about AI as a tool for political terror, used by the police, the state department, etc. Terror works better when unpredictable.
We've got a section on the variable reinforcement learning you describe & AI in that context in Why We Fear AI, too
November 26, 2025 at 9:47 PM
i agree with this
People are calling this “lazy” which I think is kind of missing what’s going on here. Increasingly convinced that the comparable technology, in how it interacts with the human nervous system, is the slot machine.
November 26, 2025 at 9:25 PM
Reposted by Anthony Moser
new Catholic childhood guilt just dropped
Pope Leo XIV told students not to use artificial intelligence for homework, saying that AI ‘won’t stand in authentic wonder before the beauty of God’s creation.’
Even God Is Worried About ChatGPT
Pope Leo XIV told students not to use artificial intelligence for homework, saying that AI ‘won’t stand in authentic wonder before the beauty of God’s creation.’
www.vulture.com
November 26, 2025 at 3:52 PM
Reposted by Anthony Moser
I can't speak for everyone on Bluesky, but this by @anthonymoser.com speaks for me.

anthonymoser.github.io/writing/ai/h...
I Am An AI Hater
I am an AI hater. This is considered rude, but I do not care, because I am a hater.
anthonymoser.github.io
November 26, 2025 at 12:35 PM
it wasn't for thanksgiving but i once helped make a meat pie that turned out pretty well.

it was for a "church night" variety show sketch called the true story of the last supper. instead of bread Jesus shows up on a crutch with one leg missing holding a meat pie
November 26, 2025 at 5:21 AM
here's a tune I wrote back in early 2011. like many of my songs i wish it had stopped being relevant between then and now
November 26, 2025 at 2:30 AM
Reposted by Anthony Moser
Reposted w/ alt-text
November 26, 2025 at 1:47 AM
this is actually the correct response whenever someone tries to read you synthetic text
November 25, 2025 at 10:45 PM
"dear carlos, thank you for bringing heavy equipment to force people out of the park and back onto the streets, minus their possessions. xoxo comic sams"
FOIA tuesday snack

a text from Alderman Samantha Nugent to CEO of the Chicago Park District, Carlos Ramirez Rosa

she is thanking him for a job well done for a violent removal of unhoused people in legion park
November 25, 2025 at 8:59 PM
haven't read this yet but i like to say they are "structurally indifferent to truth" so i think I'll like it
In "Rethinking Error," historian Johan Fredrikzon goes to the very heart of a large language model's incapacity to "know": a problem the industry likes to call hallucinations, but which Fredrikzon calls "epistemological indifference."

Link: read.dukeupress.edu/critical-ai/...
Rethinking Error: “Hallucinations” and Epistemological Indifference | Critical AI | Duke University Press
read.dukeupress.edu
November 25, 2025 at 7:38 PM
gen ctrl-z
It's weird to not only have lived through an information revolution but also now living through its undoing, all within less than a generation.
Google at its peak was basically the best information retrieval system in human history and they and every competitor decided going from there to “you didn’t want answers you wanted half-assed auto-complete 80%-wrong hallucinations” in a few years was the right idea
November 25, 2025 at 3:12 PM
And now you're baaaaack
from outer space
November 25, 2025 at 1:24 PM
Reposted by Anthony Moser
I should really write about this in more length but I have been increasingly bummed out by people's reactions to robotaxis, like "well, at least they're safer than human drivers". Even if that were true (it's, at best, complicated) it's such an impoverished view of what is possible. (cont'd) /
incidentally, this is precisely what everybody said would happen with robotaxis, huge numbers of empty vehicles clogging up streets.
Waymo privatized another public street:

Chanel approaching 4th, San Francisco

Possibly queued for a Billie Eilish show at Chase Center ~half mile away.

The light rail train on 4th seen passing in front of this roboherd has more passenger capacity than all of them combined.

OP: .tiktok.renaspam18
November 24, 2025 at 10:21 PM
Dammit. I retract any clever memes
the big problem with that Adversarial Poetry versus AI paper is that they did half the work with chatbots and not humans.

they translated prompts into poems using chatbots.

then they judged how well the poems jailbroke the chatbots using other chatbots to do the judging?!
>>
November 24, 2025 at 7:17 PM
Reposted by Anthony Moser
The blood-dimmed tide is loosed, and everywhere the ceremony of innocence is drowned, Charlie Brown
The grapes of wrath are filling and growing heavy, growing heavy for the vintage, Charlie Brown
At least the four of them are safe at last.

AM will be all the madder for that. It makes me a little happier. And yet ... AM has won, simply ... he has taken his revenge ...

I have no mouth. And I must scream, Charlie Brown.
November 24, 2025 at 7:10 PM
Reposted by Anthony Moser
researchers, press and non-profits in this city slept on CPS being a driver of slow death, they fixated on fast death by CPD

Adam Toledo was a kid who was operated upon by systems of power, who didn't get supports for disability in his first elementary school or his second and was foreclosed upon
November 24, 2025 at 3:53 PM
"the problem with surveillance tools is that I am not guaranteed access to surveillance data for research" says researcher

this piece is a layer cake of bad frames, bad inferences, and bad faith and i'm surprised it's from today instead of ten years ago
Rewiring city's technology ties following ShotSpotter saga
At a moment when an authoritarian president is seeking every tool possible to target his perceived political enemies, there is no better time for Chicago to rethink how it does business with technolog...
chicago.suntimes.com
November 24, 2025 at 2:29 PM
I can see the confusion. He thinks it's a machine that makes fun of people but actually it's a mockery of human life
Everything about this might be the saddest thing I’ve ever seen
November 24, 2025 at 1:38 PM
Reposted by Anthony Moser
We were somewhere around Barstow on the edge of the desert when the drugs began to take hold, Charlie Brown
The grapes of wrath are filling and growing heavy, growing heavy for the vintage, Charlie Brown
At least the four of them are safe at last.

AM will be all the madder for that. It makes me a little happier. And yet ... AM has won, simply ... he has taken his revenge ...

I have no mouth. And I must scream, Charlie Brown.
November 24, 2025 at 1:26 PM
Reposted by Anthony Moser
Media literacy is down, so to be clear, when the WSJ writes “Ultimately, the fact pattern Meta relies on to meet its conflicting objectives strains credibility” about your accounting practices and runs an accompanying flowchart, that is the equivalent of a 500-foot neon sign reading “FRAUD”
This feels like one of those stories you're going to look at a year from now and say to yourself, "Why didn't I move more of my portfolio into cash?" **

** (This post is for informational purposes only and should not be construed as financial or any other advice.)

🎁

www.wsj.com/tech/meta-ai...
AI Meets Aggressive Accounting at Meta’s Gigantic New Data Center
Favorable treatment off the balance sheet hinges on some convenient assumptions.
www.wsj.com
November 24, 2025 at 1:18 PM
The grapes of wrath are filling and growing heavy, growing heavy for the vintage, Charlie Brown
At least the four of them are safe at last.

AM will be all the madder for that. It makes me a little happier. And yet ... AM has won, simply ... he has taken his revenge ...

I have no mouth. And I must scream, Charlie Brown.
His feet are light and nimble. He never sleeps. He says that he will never die. He dances in light and in shadow and he is a great favorite. He never sleeps, the judge. He is dancing, dancing. He says that he will never die, Charlie Brown.
November 24, 2025 at 1:04 PM
Reposted by Anthony Moser
Like, the very first rule of doing anything at all creative is *you have to give a fuck about it*

Using genAI is proof you don't.
November 24, 2025 at 11:27 AM
somebody get me a livestream of bear stearns
We’re in the “speculative-AI-debt-hits-Page-One” part of the cycle.

@wsj.com $HYG $LQD
November 24, 2025 at 12:39 PM
Reposted by Anthony Moser
The water usage argument is a political/marketing argument disguised as an empirical one. It makes my brain glaze over, so while I think it's important to understand the resource usage of LLMs, it's not my ministry.
November 24, 2025 at 3:16 AM
this correction might be a compelling reason to change my position on AI if:

a) environmental harm was the only issue AND
b) it was otherwise super useful AND
c) it was easy to get right instead of hard to estimate at all

but none of that is true so I don't think this error changes anything
main thought about the Empire of AI correction situation is that it sure seems really damn hard to get straight info about these datacenters and their resource footprints if it took piecing together hundreds of disparate sources over like 5 years to arrive at what turned out to be inaccurate numbers
November 24, 2025 at 2:39 AM