Month: April 2026

“The overview effect is when you’re looking through the cupola and you see the Earth as it exists with the whole universe in the background. You see the thin blue line of the atmosphere, and then when you’re on the dark side of the Earth, you actually see this very thin green line that shows you where the atmosphere is. What you realize is every single person that you know is sustained and inside of that green line and everything else outside of it is completely inhospitable. You don’t see borders, you don’t see religious lines, you don’t see political boundaries. All you see is Earth and you see that we are way more alike than we are different.”

– Artemis II Mission Specialist Christina Koch

The Overview Effect: Astronaut Perspectives from 25 Years in Low Earth Orbit – NASA

In a randomized trial of 1,000 high school students, those using GPT-4 performed much better on practice problems but scored 17 percent worse once the AI was removed. The likely culprit was overreliance: students asked for answers, copied them, and moved on. But the researchers also showed that a more structured prompt that offered step-by-step hints rather than direct answers largely eliminated the negative effects. In some ways, that was the most important finding: the design of the AI tutor mattered enormously.
— Read on www.aei.org/technology-and-innovation/ai-works-in-education-when-it-makes-learning-harder-not-easier/

In a world where your economic value can evaporate overnight, “Become a whole person” sounds like advice from someone who has never worried about next month’s rent. The utilitarian case for education has the force of necessity behind it. For millions of people, making yourself useful is what responsibility to their families demands.

Yet if the response to being replaceable is always to train for a different function, you have entered a race you structurally cannot win. The principle that makes your education valuable is the same principle that makes you disposable the moment the function migrates.

The scramble into computer science was an early sign of the trap: students rushed toward the field that seemed safest, and then AI began destabilizing the very functions it trained them to perform. The flight to function looks rational from inside it; that is what makes it a trap.
— Read on blog.cosmos-institute.org/p/you-are-not-a-function

Within hours Carlini found numerous techniques to infiltrate systems used around the world. Once Carlini was back in Anthropic’s downtown San Francisco office, he discovered Mythos was able to autonomously create powerful break-in tools, including against Linux, the open-source code that underpins most of modern computing.

Mythos orchestrated the digital equivalent of a bank robbery: getting past security protocols and through the front door of networks, and breaking into digital vaults that gave it access to online treasures. AI had picked locks, but now it could pull off an entire heist.

Carlini and some of his colleagues began alerting staff to what they’d found. And each day they continued to discover high-severity and critical bugs in the systems Mythos probed, the kind of flaws normally uncovered by the world’s best hackers.

How Anthropic Discovered Mythos AI Was Too Dangerous For Release – Bloomberg

The “expensive” bag costs less. But VF Corp doesn’t want you to do this math, because the $35 bag creates a repeat customer every eighteen months. The $200 bag creates one transaction and zero follow-ups. From a shareholder’s perspective, the bag that falls apart is the better product.
That’s the business model. Repeat failure, repeat purchase, repeat revenue. The quality decline isn’t a side effect. It’s the strategy.
— Read on www.worseonpurpose.com/p/your-backpack-got-worse-on-purpose

The central claim is this: when you scroll a social media feed, watch algorithmically recommended content, or interact with any digital interface that monitors your behavior, you are not merely consuming a free product. You are simultaneously producing a distinct economic output — behavioral data that serves as a critical input to AI training, targeted advertising, and algorithmic product development. This production is not compensated by wages. It is not recorded in any national accounting system. And its aggregate value may be large enough to reshape our understanding of the global economy.

The six hours and forty-seven minutes on my screen time report are not wasted time. They may be uncompensated labor. And the difference between those two interpretations has consequences that reach far beyond my Sunday morning guilt.
— Read on navvaidhyanathanatvysdomai.substack.com/p/dark-gdp-the-economics-of-entangled

THIS MONTH, USA Today published an excellent report that revealed how US Immigrations and Customs Enforcement delayed disclosing key information about the impacts of its detainment policies. The authors used the Internet Archive’s Wayback Machine to compile and analyze detention statistics from ICE and track how the agency had changed under the Trump administration. The story is one of countless examples of how the Wayback Machine, which crawls and preserves web pages, has helped preserve information for the public good. It was also, Wayback Machine director Mark Graham says, “a little ironic.”

USA Today Co., the publishing conglomerate formerly known as Gannet that runs both its namesake paper and over 200 additional media outlets, bars the Wayback Machine from archiving its work. “They’re able to pull together their story research because the Wayback Machine exists. At the same time, they’re blocking access,” Graham says.

The Internet’s Most Powerful Archiving Tool Is in Peril | WIRED