weekend ai reads for 2026-07-24

My goal was to understand the entire state of so-called tokenomics. I wanted to know what monitoring systems exist, how teams govern their AI spend, and which optimization tools and practices actually work, both in the papers and in the real world.

The token savings are therefore not gained by switching models or ‘rerouting’ to cheaper LLMs, but by changing the starting conditions of the doc review.

I dont think we understand what two people fighting each other with 1 billion tokens a minute will look like. Small property disputes will look like megacap antitrust lawsuits

While the internet has opened up new doors for musicians, the increasing corporatization and algorithmic takeover of online spaces has done irreparable damage to organic music discovery and independent communities.

“She had a few wrinkles. My mother was ageing,” he says, typing AGEING in all caps on text to underline his shock. He hadn’t considered the way many video-calling apps now apply face-smoothening filters by default.

The first kind of world model is a renderer. A renderer outputs observations in the form of pixels meant for human eyes, and the quality that matters most is visual fidelity.

The second kind is a simulator. A simulator outputs state: a geometrically, physically or dynamically faithful representation of the world that humans and computer programs can both compute on and interact with.

The third kind is a planner. A planner outputs actions.

Customers are actually buying judgement, experience and confidence. They’re buying better decisions, reduced risk and faster progress.

AI changes the cost of delivering those things, but it doesn’t diminish their value. If anything, it increases the importance of knowing when to trust AI, when to challenge it, and how to turn its output into something genuinely useful.

Armed with generative AI, a B student can produce A work while turning into a C student.

Eggers argues that the advent of generative AI is destroying an animating force of his work: the act of learning by doing. He told OpenAI’s staff: “The effect of ChatGPT on educators’ lives is catastrophic. Whether you intended to do it or not, you’ve made every teacher’s life infinitely more difficult than it was two years ago. So, just let that settle in . . . 

“If students are using it to compose, which is the biggest tragedy of all, they’ll never learn to write. And their voice is stolen from them. They’ll never have the ability to say their truth and tell their own story. And that’s silencing an entire generation or two.”

AI is changing what students need to learn and how schools prove it. Explore 12 provocations for building assessment that measures human capability.

1. Don’t start with cheating. Start with the purpose of school.

2. Measure capacities rewarded by the new economy.

3. Non-traditional teaching strategies can make thinking visible.

4. Cognitive offloading is the enemy, not AI use.

“Pulling teeth to get him to get proactive with strategy and now this is what we get,” one texted to another. The exchange, which surfaced in public records obtained this month by The Chronicle, was first reported by the Milwaukee Journal Sentinel.

Soon after, President Jay Rothman would be out of a job.

This is where most benchmarks cut a corner we refused to. Preparing labels took more than importing human feedback. We had the engineers who wrote the PRs annotate candidate findings, then compared three sources against each other: the human annotations, the original candidate findings, and an agentic judge. Human feedback was valuable but frequently wrong: reviewers missed valid issues, accepted weak claims, or read a historical PR’s context differently than the next reviewer would. Where the three sources disagreed, we re-reviewed the evidence by hand and resolved it, and those resolved cases became calibration data for the judge.

The visual dictionary of UI. Describe the thing in your own sloppy words; you get the real name, the API symbol, and a paste-ready prompt for your coding agent.

Last updated: approximately as recently as your CEO's five-year plan.

A TTF font that hides what you’re typing from AI. Type a message where each letter contains a decoy.

The faces and voices of 32 real people were used under licensing agreements, and human artists created the concept art.

American service providers, private and public, increasingly run like ghost ships.
share

Japanese technology and design seemed in greater service to humans, their journeys, and enterprises.

It’s possible I believe in files partly because I know how things are built in the cloud. Building things on clouds even sounds bad.