Christopher Berry
banner
cjpberry.bsky.social
Christopher Berry
@cjpberry.bsky.social
Perpetually mystified | Views are mine.
Here’s the Canada-Alberta MOU. It’s worth a read. And it’s fine.

See, you and me, the reader, could both author a MOU that the sum of angles of a triangle add up to 1269 degrees.

And parts of this MOU read like that.

www.pm.gc.ca/en/news/back...
Canada-Alberta Memorandum of Understanding
PREFACE At this pivotal global moment, Canada and Alberta, working closely with Indigenous Peoples and industry, must work together cooperatively, and within their respective jurisdictions, to foster ...
www.pm.gc.ca
November 29, 2025 at 2:31 PM
The core of the UCP is likely out of step with the balance of Albertans.

The ICR shops closed. Cod collapsed. So too, hydrocarbons.

Transition now or pretend to be so shocked and surprised.
Danielle Smith got booed at a UPC rally for saying she

"supports a strong and sovereign Alberta within a united Canada...

and I hope people today feel a lot more confident that 'Canada works' than they did a few days ago"

youtu.be/UjSCc_MoAMA?...
CBC News: The National | Smith draws boos at UCP convention
YouTube video by CBC News: The National
youtu.be
November 29, 2025 at 1:46 PM
Reposted by Christopher Berry
Astonishingly, sea ice still hasn't started to reform over the Barents Sea area yet of the #Arctic. This includes record low conditions around Svalbard as well. It's been a very bad few months across the region.

Check out more graphics here: zacklabe.com/arctic-sea-i...
November 29, 2025 at 1:13 AM
Introduce yourself with 5 animals you’ve seen in the wild:

Beaver
Raccoon
Deer
Canadian Goose
Mountain Goat

Happy Friday Friends!
Introduce yourself with 5 animals you’ve seen in the wild:

Moose
Stoplight parrotfish
Box turtle
Elephant
Wild horse
Introduce yourself with 5 animals you've seen in the wild:

Caribou
Ravens
Seals
Kea
Turkeys
November 29, 2025 at 3:14 AM
Reposted by Christopher Berry
"Put another way, the median mobile page is now 70 times larger than the total storage of the computer that landed men on the moon."

infrequently.org/2025/11/perf...
The Performance Inequality Gap, 2026 - Infrequently Noted
Embedded in this year's network and device estimates is hopeful news about the trajectory of devices and networks. It has never been easier to deliver pages quickly, but we are not collectively hittin...
infrequently.org
November 28, 2025 at 6:09 PM
Reposted by Christopher Berry
It seems likely that Opus 4.5 is actually very good at sensing how many tokens it's generated, tending to overestimate, and tending to overestimate more on longer generations. Very preliminary, but r = 0.9747 and p_value = 1.74e-06!

github.com/Embedding-Sp...
Azimuth_II/box_5/shared/experiments/token_vibe_test/token_vibe_calibration.ipynb at 5a4329afc0b2c8178d1428760ab6956f39752cb3 · Embedding-Space/Azimuth_II
Contribute to Embedding-Space/Azimuth_II development by creating an account on GitHub.
github.com
November 29, 2025 at 3:02 AM
Reposted by Christopher Berry
Hi, my name's Tim, and I've used Bayesian structural time series while also being befuddled as to fundamentals of Bayesian methods overall. I had a prior that the latest episode of the Analytics Power Hour podcast would clear things up a bit for me, and it did! analyticshour.io/2025/11/25/2...
#285: Our Prior Is That Many Analysts Are Confounded by Bayesian Statistics
Before you listen to this episode, can you quantify how useful you expect it to be? That’s a prior! And "priors" is a word that gets used a lot in this discussion with Michael Kaminsky as we try to de...
analyticshour.io
November 25, 2025 at 3:40 PM
The upsignal/downsignal components in the emails are interesting. Bias comfirming that many never leave prep school.

We’re going to get some great science out of the dataset. And that’ll be a small amount of justice. Maybe.
November 23, 2025 at 1:27 PM
I look forward to the meeting to discuss the venue for the meeting to consider to caterer for the meeting to discuss the composition of the steering committee for the voluntary agreement to begin discussions on a roadmap to an eventual phase-out of fossil fuels.
Countries meeting in Brazil for 2 weeks could manage only a voluntary agreement to begin discussions on a roadmap to an eventual phase-out of fossil fuels, & they achieved this incremental progress in the face of implacable opposition from oil-producing countries.
www.theguardian.com/environment/...
End of fossil fuel era inches closer as Cop30 deal agreed after bitter standoff
Wealthy countries agree to triple funds for countries to tackle climate impacts, but deforestation and critical minerals blocked from final deal
www.theguardian.com
November 22, 2025 at 6:01 PM
Just finished Michael Lewis’ “The Fifth Risk” … What a treat. Just a gem of a book. Love it. I’m gushing.

There’s a real craft for storytelling. I positively love the DOE stories. Just great stuff.

Most of my friends will love the book.
November 21, 2025 at 11:09 PM
The trivialization of what happened at The Pulse Nightclub is rather disappointing. Yet ultimately unsurprising.
Kash Patel's personal firing of the Pride flag-displaying employee was always absurd and unconstitutional, so I'm glad David Maltinsky is suing, but, wait.

It was a flag FLOWN BY THE FBI that was given to him BY THE FBI?! Come on.

Complaint: storage.courtlistener.com/recap/gov.us...
November 20, 2025 at 10:42 PM
Matthew Desmond’s “Poverty By America” is factual, for readers, and explains the design of American poverty. It makes good arguments.

For Canadians I’d rec Elsbeth Heaman’s “Tax Order and Good Government” for the roots of the Canadian design.
November 20, 2025 at 2:01 AM
Tonight at Weird Wednesday at Trajectory Labs with Jason Yung. Luddites! Sherwood Forresty.
November 19, 2025 at 11:39 PM
Did an old style CBC user interview on Queen Street last night. Once a public broadcaster, always. And met the rare causal-factual/entertainment-escape segment. They’re around 1:8, 1:10 in Canada, and a real pleasure to talk to.

They’re terribly underserved by both the privates and the publics.
November 19, 2025 at 5:37 PM
Reposted by Christopher Berry
"We have a 2000s housing bubble level of financial engineering on top of a 1920s level of private unregulated lending on top of something bigger than a 1990s internet (or 1870s railroad) level of technology and infrastructure build-out." prospect.org/2025/11/19/a...
The AI Bubble Is Bigger Than You Think - The American Prospect
It’s not just OpenAI that looks overhyped. There’s a whole mountain of sketchy financial engineering underneath.
prospect.org
November 19, 2025 at 5:03 PM
The danger here is not only data misuse but the erosion of trust. When every digital interaction feels like surveillance, the risk is dulling the authenticity that makes such tools valuable.”
One reason LLMs are so popular is the promise of seemingly private, judgment-free interaction. But once users internalize that their data is being recorded, studied, or reused, it could lead to a shift in how people think, speak, and even feel when using AI, says computer scientist Koustuv Saha.
The Risks of the 'Observer Effect' from Being Watched by AI | TechPolicy.Press
Dr. Koustuv Saha says if users begin to feel their privacy is compromised, they may stop using AI in the very ways that make it useful.
www.techpolicy.press
November 19, 2025 at 1:52 PM
Trajectory Labs has been hosting this book chat on Monday evenings. Great conversations - thanks Georgia for facilitating!
November 18, 2025 at 12:33 AM
Reposted by Christopher Berry
Great perspective from #CDSM2025: causal inference is 'what-if' analysis. You don’t have to get everything perfect—or how dare you use the c-word. What matters is laying out your assumptions transparently and showing us what happens when they're violated.
November 15, 2025 at 6:47 PM
My clothes are still soaked from Milton Mayer’s “They Thought They Were Free”. That bit where the Americans believed they were causing the Germans to accept responsibility for the Nazis at Nuremberg, and that 9 in 10 Nazis were completely devoid of shame.

Sticky. Sticky stuff.
November 15, 2025 at 4:26 PM
Reposted by Christopher Berry
Over on the Gelman blog, there's a post about this histogram of z-scores in published papers. I'm always baffled by the discourse around it. Why is this plot a symptom of a problem? Under what model would we expect this to be a bell curve?
November 15, 2025 at 3:31 PM
Finished Gati’s “Zbig”. Got him in a Grand-Grand Strategy way. I’ve never experienced Bloodlands, I don’t have his intuition, just the fuzzzziest Anglo-Acadian intuition about the value sets that make institutions, reinforce, combine to cause power. Hinterlandish.

Good read. A few bangers.
November 15, 2025 at 2:40 AM
Reposted by Christopher Berry
@alphaxiv.org released quickarXiv

Swap arxiv → quickarxiv on any paper URL to get an instant blog with figures, insights, and explanations.
November 14, 2025 at 11:52 PM
Reposted by Christopher Berry
According to signaling theory, some signals must be costly just to be costly—that's how you get a separating equilibrium. Think peacocks and their oversized feathers. So even if AI removes one costly signal, it doesn't mean we should stop technological progress — we'll just find new ones.
Is AI making job recruitment less meritocratic? We're getting some v interesting research studies on this question now, and the news is... not good. @jburnmurdoch.ft.com & I dive in, in the latest edition of our newsletter The AI Shift www.ft.com/content/e5b7...
November 14, 2025 at 11:52 AM
The twist:

Colleague is a Turniptologist
November 14, 2025 at 1:05 AM
“Moreover, we present a dataset comprising 25,368 samples (CausalDR), where each sample includes an input question, explicit causal DAG, graph-based reasoning trace, and validated answer.”
Researchers have developed CDCR-SFT, a novel method that enhances causal reasoning in language models, achieving an impressive 95.33% accuracy while significantly reducing logical hallucinations. https://arxiv.org/abs/2508.12495
Mitigating Hallucinations in Large Language Models via Causal Reasoning
ArXiv link for Mitigating Hallucinations in Large Language Models via Causal Reasoning
arxiv.org
November 14, 2025 at 1:02 AM