About to smash the Guinness world record for: Most-cropped photo in Spanish history.

Rav@rvbdrm.com · 2w ago

LMAO

Sometimes the better team loses. I mean, not today, but it does happen sometimes.

Stuck following the game this way because of poor cell reception but -- wow -- this stat. Never thought I'd see this.

The lunacy of letting 50+1 break a union. When the reality sets in, that 50 shrinks into a minority, and the majority is trapped in an unwanted new political arrangement. Today's example: Brexit a decade later. www.nytimes.com/2026/06/23/w...

Here today to teach my kids about parliamentary democracy and What. A. Morning.

Today's daily AI digest - an Acemoglu paper argues how simple math could lead to AI favoring autocracy.

We have a new Exhibit A for those AI sovereignty advocates arguing that other countries need to scale up alternatives to the U.S. models. At the very least as a backup plan

Alejandra Caraballo@esqueer.net · 2mo ago

The US government just issued an export restriction on Claude Fable/Mythos preventing any foreign national from accessing the model even in the US and Anthropic employees. As a result, all access is cutoff for the foreseeable future.

Life sometimes throws you a curveball -- like watching Canada playing in the ⚽️ World Cup, in Toronto, from your new home in Ankara, Turkey. Full of surprises! 🇨🇦

Switching owner of the other site from block to mute now that he has enough money to buy every industry and subnational region where people he dislikes could ever work or live

Canal Hipocritas Medo

ALT: Canal Hipocritas Medo

static.klipy.com

I'm proud of this. A few years ago, I couldn't use Excel. My wife taught me after the 2016 election, when she saw me tallying results by hand. Today, Apple approved my iPhone app. If you're traveling and want a break from over-touristed sites, this finds beautiful places nearby, off the beaten path.

Today's AI digest: Anthropic Backtracks On Silent Sabotage Strategy --> New Data Backs Up Youth Jobs Fear --> OpenAI's IPO Plan --> The Value Of Chatbots That Challenge

Anthropic's powerful new Mythos-level model stopped while rendering my Daily AI briefing note. Its security flag went up.... simply because I mentioned its own system card!

Striking paper from Wharton: AI must increase productivity 2.7x -- quickly -- or tech companies risk bankruptcy. For context: this is how a quickie 2.7x compares to historical precedent. Useful context for OpenAI reportedly talking to the US government about a bailout (ahem, I mean ownership stake)

Canada's new AI strategy does not skirt the big picture: While Canadian researchers played a massive role in developing modern AI - from language to vision to learning techniques - the country is more hostile to it and slower to adopt it than most peer nations. But it also holds several advantages.

Three takeaways from this extraordinary document about new practises in handling data in an organization. 1) Wiring models to real, verified internal knowledge solves most of the hallucination problem. Which takes work -- in curating, editing. Which brings us to... claude.com/blog/how-ant...

A huge day in AI news, summarized in my automated daily briefing: --> Anthropic Files for IPO at a Valuation Exceeding Most National GDPs --> Leaks on Canada's Al Strategy --> NY Times Publisher Calls for Fight Against Al Theft --> OpenAl Insists: We're Not Backing the GOP

Hey, on the bright side, Gary Bettman will be happy: it's another Stanley Cup final involving two places without ice or snow. It's what he lives for.

A thoughtful piece that argues that people who want humanity to endure -- ie most of us -- need an answer to an influential ideological cult whose ideology leads to eugenics. This answer involves redefining humanism to celebrate diversity in intelligence.

Jennifer Ouellette@jenlucpiquant.bsky.social · 2mo ago

The people who actually want AI to replace humanity www.vox.com/future-perfe...

My dailyaidigest.net today: --> Highlights of the pope's effort to reshape the global AI conversation with a 42,000-word manifesto. --> Plus several studies on the trouble of using desktop LLMs at institutional scale.

My latest: alexpanetta.substack.com/p/heres-why-...

Here's my school project this semester. An alternative travel algorithm. Had fun building this --> 233 countries & territories --> Thousands of beautiful-yet-less-crowded destinations --> Searchable by theme or by country --> With info on how to get there --> And an MCP layer --> wanderingwell.site

Thus it is written: Once a decade, my hockey team shall pull a Shaquille O'Neal-sized horseshoe out of its posterior. And I, for one, embrace it. #gohabsgo

It's hard to hate the Buffalo Sabres. North of the 40th parallel. Been suffering forever without a Cup. Eliminated the Bruins. They're practically family. But we can all hate a relative for a while, can't we?

Great piece on the AI politics in Washington, featuring naked corruption and killer turns of phrase www.theatlantic.com/magazine/202...

The Venture-Capital Populist

How David Sacks and the new tech right went full MAGA and captured Washington

theatlantic.com

DeSantis also says datacenters are unpopular on both the left and the right. If you have followed this guy's career, this is something worth taking note of. He would not be coming out early like this if he didn't see a big political problem coming down the road.

Okay, what's the truth about A.I. and jobs? A rich portrait in my automated daily AI digest today, courtesy of 3 papers just released by the U.S. National Bureau of Economic Research. The most detailed is a study of businesses by researchers at the US Census Bureau dailyaidigest.net/archive/2026...

3:49 am in Ankara, Turkey. How I follow the Stanley Cup playoffs nowadays

From the department of ambiguous omens, ahead of tonight's big Habs game. I'm in central Anatolia. Middle of Ankara. I have never seen a Habs logo here ever -- not outside my house. I just spotted a woman covered in them -- in rhinestone. A bunch of sparkly Habs logos. /

I just published this primer on a hack that's become wildly popular among techies. It has applications for broad professional uses, and is ridiculously easy to set up. I explain the setup, its 80-year history, its limitations, and why it’s catching on now. alexpanetta.substack.com/p/why-person...

Why Personalized Wikipedias Are Suddenly Catching Fire

It involves an 80-year-old dream of building a 'second brain.' It's now doable in 10 minutes

alexpanetta.substack.com

Amusing irony given US administration trying to destroy Claude: researchers test AI models for bias and find Claude is the only one not biased against free-market economics. arxiv.org/pdf/2604.213...