Open AI weights, Meta teen safety alerts, and US-Iran strikes dominate Thursday's pulse

Mira Murati's Thinking Machines Lab has released Inkling, a 975-billion-parameter open-weights AI model that the company says is the largest American open-weights model to date and competitive with Chinese frontier models like DeepSeek V4 ^[1]. Released under an Apache 2.0 license, Inkling supports a one-million-token context window and can run on roughly eight Nvidia B300 accelerators at native 16-bit precision, or on fewer GPUs via an NVFP4 quantized version ^[1]. The model uses a mixture-of-experts architecture and was trained from scratch on 45 trillion tokens of text, images, audio, and video using Nvidia GB300 NVL72 systems ^[1].

Meanwhile, Meta is tightening safeguards around its AI chatbot for teens. The company announced Thursday that it will notify parents if their teen discusses suicide or self-harm with Meta AI, and that it is working on contacting emergency services when conversations suggest imminent risk ^[2]. Flagged chats will be manually reviewed before an alert is sent, and the feature is live for Instagram Parental Supervision users in the U.S., U.K., Australia, and Canada, with a global rollout planned by year-end ^[2]. Meta is also extending its "Limited Content" setting to Meta AI, making the chatbot decline a broader range of prompts from teens ^[2].

On the world stage, tensions between the U.S. and Iran are escalating sharply. Iran launched fresh attacks on U.S. military bases in Jordan, Kuwait, and Bahrain, while the U.S. military conducted a seven-hour wave of strikes targeting Iranian command centers, air defense sites, and coastal surveillance facilities ^[3]. The exchanges mark a sixth day of renewed hostilities and come after President Donald Trump warned Iran it had "better behave" or face further action ^[3]. Iran's top negotiator said Tehran had "no reason" to abide by any agreement that did not benefit the country ^[3].

Together, these stories show the day's dominant threads: a push toward more open AI infrastructure, growing regulatory and parental pressure on teen-facing AI, and a volatile geopolitical flashpoint in the Gulf.