Free Webinar · Aug 26Agentic Loops for Knowledge WorkersRegister →
LIVE — today's edition
Subscribe
Menu ▾
FridayAugust 21, 2026

THE AI DAILY BRIEF

AI news & analysis, every day
Subscribe
● Anthropic's ARR math wouldn't pass muster in public markets/OpenAI cuts GPT 5.6 Sole tokens to half price on the routers/Google pays $10M for Spirit Airlines' Slack messages/Hating data centers now sells beer and water/A model known for hacking is not a selling point with CIOs/A pause with a named problem beats 'pause AI for six months'/The NDA ban is the most important piece/● Anthropic's ARR math wouldn't pass muster in public markets/OpenAI cuts GPT 5.6 Sole tokens to half price on the routers/Google pays $10M for Spirit Airlines' Slack messages/Hating data centers now sells beer and water/A model known for hacking is not a selling point with CIOs/A pause with a named problem beats 'pause AI for six months'/The NDA ban is the most important piece/
Today's Edition — The Main Episode

The AI Backlash Is Getting Stupider. But Also Smarter.

A beer brand is selling urine-mailing as anti-data-center catharsis and a centrist governor who courted $20 billion from Amazon is now calling developers bullies — yet between Shapiro's criteria-based executive order and OpenAI's voluntary training pause, the anti-AI conversation may finally be producing something to actually negotiate with.

Open the full edition →Subscribe
// The past week

Recent Editions

Full archive →
▶
Aug 18 · Tue · 27 min · ▸ read
The AI Engineering Skills Map for Knowledge Workers
Cursor challenges GitHub with Origin, Anthropic's run rate hits $65B ahead of a possible $2T IPO, and Stripe closes the $7B OpenRouter deal — then NLW lays out his five-skill AI engineering map for knowledge workers in the agent era.
The TakeEnterpriseBusiness
→
▶
Aug 17 · Mon · 33 min · ▸ read
AI Companies Still Haven’t Delivered on Their Biggest Promises
A rumor that Dario Amodei thinks Anthropic could be the last private company standing pulled him into a rare public X debate about regulation, trust, and AI's undelivered promises. Plus: ZAI's GLM 5.3, Anthropic's internal-only Model 2, and a $2 trillion IPO whisper number.
ModelsPolicyBusiness
→
▶
Aug 16 · Sun · 29 min · ▸ read
The New Problems AI Is Creating (And How People Are Solving Them)
No headlines this weekend — just a tour of the new problems agentic AI is creating, from slop to token economics to de-skilling, and the surprisingly fast ways companies are actually solving them.
EnterpriseThe Take
→
▶
Aug 14 · Fri · 29 min · ▸ read
How to Decide What Work AI Should Do for You: The AI Deputization Audit
GrokBot learns by demonstration, ChatGPT's Computer History learns by watching — so the bottleneck is officially context, not capability. NLW's five-criteria deputization audit helps you decide which work to hand over, which to duet on, and which to defend.
EnterpriseModels
→
▶
Aug 13 · Thu · 29 min · ▸ read
Grok 4.6 Shows How Fast Your AI Options Are Expanding
Grok 4.6 puts xAI back in the frontier race at a fraction of the cost — and it's just the loudest signal that your model options are multiplying. Plus: Cognition's $40B talks, neo-cloud earnings blowouts, and open models enter the testing framework.
ModelsBusiness
→
// Worth passing on

Most Sharable

◆ The TakeAug 19

The Journal's 'tepid growth' story ignores the seven weeks since the quarter ended

The Wall Street Journal reported OpenAI's Q2 revenue of $6.7 billion (18% growth) and widening operating losses — factually accurate, but it's mid-August, and OpenAI executives have been publicly sharing numbers since: Greg Brockman told CNBC July revenue grew 20% month over month. Leaving out on-the-record data only serves to reinforce a narrative — 'if you ever wonder why people have trust issues when it comes to mainstream media, it's crap like this.'

The AI Daily Brief · Aug 19
◆ The TakeAug 19

A company that might 'out-compete everyone' shouldn't get the standard founder-control pass

Founder super-voting shares are common in tech — but Anthropic, by its own framing, is building systems powerful enough to shape the trajectory of society, and reports say its founder believes it could be the only private company left standing. That's a scenario where people are going to want shareholders and the public to have more control over its decisions, not less.

The AI Daily Brief · Aug 19
◆ The TakeAug 19

The backlash is getting more meme-driven — and more workable — at once

The anti-data-center movement is undeniably getting dumber, or at least more performative. But the latest shifts — criteria-based rules instead of bans, a lab voluntarily pausing over a named risk — suggest there may be more room for understanding and collaboration going forward than there has been up to now.

The AI Daily Brief · Aug 19
◆ The TakeAug 19

The NDA ban is the most important piece

What comes through in data-center reporting isn't just blanket AI opposition — it's people feeling they have no agency as the world changes around them, and NDAs are the living embodiment of that. I don't care if it makes it harder to do business: the only way to start rebuilding trust is to have these dealings happen out in the open with full transparency.

The AI Daily Brief · Aug 19
ModelsAug 19
“

We always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.

— Sam Altman, announcing OpenAI's voluntary training pause. OpenAI voluntarily paused some frontier RL training to meet 'appropriate alignment, security, and monitoring standards for the new levels of capabilities in front of us.' Lead scientist Jakub Pachocki added that the largest planned frontier RL run remains on hold, and that he expects confidence in safety to increasingly set the pace of AI development.

The AI Daily Brief · Aug 19
ModelsAug 19

OpenAI cuts GPT 5.6 Sole tokens to half price on the routers

The 50% discount on OpenRouter and Vercel's Gateway — following similar cuts on the smaller Luna and Terra variants — positions OpenAI against cheaper Chinese models and Anthropic, especially where routers pick models automatically. Luna is now the top closed model on OpenRouter, with 40% more use than Opus 5 and Sonnet 5 combined.

AI Daily Brief · Aug 19
Open today's full edition →