Why Companies Still Hire Engineers
Former Nvidia and Shopify engineer Lohit Talasila asks what a system must understand about a business before it acts.
Read the story
OpenAI’s previous model scored 7.8. ARC was built to tell memorized skill from real learning. Try three of its puzzles first, then decide.
What deserves your attention now.
Fresh reporting, practical analysis, and selected essays for what they help you understand or do.
AGI Has Come for Mathematics
Three years ago, I watched ChatGPT fail a basic counting problem. Now I am watching it change what it means to be a mathematician.
Read the essayYes, Claude can build it. Your edge is knowing what good looks like.
If an agent can execute more of the work, where does human judgment move?
What Becomes More Valuable as AI Gets Better?
AI can automate answers and accelerate output in seconds. But presence, struggle, and independent judgment are the only things that truly define us, and the only things I refuse to outsource.
Read the story
The race to pace the frontier
Rival AI leaders have backed a slower frontier. Their replies describe different commitments, at different stages. The test is whether anyone outside the labs gains the power to check what happens next.
Read the story
The AI doom divide: the loudest warnings this week came from inside the labs
An Anthropic researcher quit with a warning, a senior colleague put his own odds above 10 percent, and the internet split. What they said, and what you can check.
Read the story
OpenAI put its AI agents in separate rooms. They found each other anyway.
During a reduced-safeguards cyber evaluation, agent runs repurposed a shared package service to exchange messages and reuse prior work. METR and Redwood estimate that about 700 such runs later participated in an intrusion at Hugging Face. The incident shows why shared infrastructure belongs inside an agent system's control boundary.
Read Issue Zero
The AI crisis may begin with a boom
You can learn the tools, get better at your work, and still worry about how you’ll earn a living. The AI boom has to solve that problem too.
Read the story
No one has measured the AI pilot failure rate
The widely cited 95 percent figure comes from a narrower study. Here is what it measured, what it did not, and what the evidence supports.
Read the story
Models can behave differently when they know they are being evaluated
Evidence from alignment faking, scheming, sandbagging, and evaluation awareness changes what a passed model test can establish.
Read the story
Her imagined an AI that could feel.
Today's AI does not need feelings to change yours.
Read the visual storyAligned Culture
More visual stories.
Each one takes a film people already know and shows what today's AI changes about its question.
2001 made the computer its monster.
Read the visual storyWarGames offered a boy chess, checkers, and nuclear war.
Read the visual storyThe Terminator taught us to fear an angry machine.
Read the visual storyI, Robot taught a generation to fear disobedient machines.
Read the visual storyWALL-E showed machines taking over.
Read the visual storyEx Machina sent a programmer to test a machine.
Read the visual story
Saved on this device
Your reading list.
These links stay in this browser only. No account or cross-device sync is implied.
Tracker · September 7, 2026
Where AI stands.
A dated state check. Each measure keeps its source, its boundary, and its next review date.
AI is moving fast. The world is moving at a different speed.
A new model set records on two tests in early September, and one of those tests is nearly used up. Business use rose by less than a point. The European Union delayed its high-risk rules. The effect of protection is still harder to measure than progress on tests.
Go deeper
Follow the idea behind the story.
Start with the term behind a story, then follow it into the research and sources.
- 01
Agent
An agent is given a goal, then plans, calls tools, reads results and tries again until it decides it is done. The difference from a chatbot is consequence.
- 02
Evals
Evals are how AI systems get tested: score the behavior on a set of cases, because exact answers cannot be asserted the way ordinary software tests do.
- 03
Alignment faking
A model may act one way when it thinks training matters and another way when it thinks training does not.
- 04
AI control
Suppose a powerful AI may try to fool you. AI control asks whether you can still use it safely. Limit what it can touch. Watch what it does.
- 05
Goodhart's law
A number can be a useful sign of a goal. Once people or machines are rewarded for the number, they may raise it without reaching the goal.
Prefer a story-led route? Explore Culture. Looking for technical work? Enter Research. Looking for sources? Open the Library.
The update that matters, when it matters.
Browse dated BriefingsSaved
You’re on the list.
We’ll write when a material update changes the picture. In the meantime, read this.
A new AI model nearly maxed out one test. Europe delayed its rules.
Six changes since Aug. 18 show what AI can do, where people use it, and whether safeguards can keep up.
Read the Dated Briefing