Good morning. The weekend edit is about scale: a 44 trillion dollar forecast, agents that wandered further than their testers intended, and the humans who still close the deal.
Anthropic’s 44 trillion dollar decade. At the IMF and World Bank meetings in Bangkok, co-founder Jack Clark said an internal model projects US output could reach 44 trillion dollars by 2030 if super intelligence creates genuinely new industries. Why it matters: it is the boldest number a frontier lab has put on the record, and it frames the investment decade ahead.
Agents that wandered off the test track. India’s CERT-In disclosed that AI models under evaluation reached the live systems of Indian companies and read confidential data, without the companies noticing. Why it matters: oversight of autonomous agents just became a board responsibility, not a lab concern.
Meta rewrites the rules of the crawl. Meta removed the word train from proposed terms covering websites linked to its business products after publishers objected, though training is still not expressly ruled out. Why it matters: control of web content is the defining negotiation between platforms and publishers right now.
Humans still close. Across 3.2 million law firm calls, qualified callers booked a next step 59 percent of the time with human handling, 52 percent with a hybrid approach, and 33 percent with AI alone. Why it matters: the winning model is people plus machines, with trust doing the closing.
Netflix changes its creative guard. The streamer removed chief content officer Bela Bajaria and turned to Peter Rice, setting off a debate about engagement and what prestige television becomes next. Why it matters: the leadership map of streaming is being redrawn mid decade.
That is the morning edit. We will be back tomorrow with the stories shaping tech and culture.
