AI news in context

The former OpenAI chief product officer's move to Stoke Space signals growing Silicon Valley interest in reusable rocket startups.

• Kevin Weil, former OpenAI CPO and VP of Science, has joined the board of Stoke Space, a reusable rocket startup.
• Weil left OpenAI after leading Prism, an AI workspace for scientists; his departure was confirmed by WIRED.

A Gizmodo report says the upcoming model will mark the first joint production between SpaceXAI and coding-tool maker Cursor.

• SpaceXAI is reportedly releasing a major new AI model this week, per Gizmodo.
• The model is described as the first joint production between SpaceXAI and Cursor.

Meta's latest AI image generator taps public Instagram content, raising fresh privacy concerns for anyone with an open account.

• Meta has launched an AI image generator that can use content from public Instagram accounts, including user faces.
• An opt-out mechanism exists, but its discoverability and scope are not yet fully detailed.

Meta's new detector can flag synthetic media produced by its latest generative models, though unexplained rate limits may constrain practical use.

• Meta released an AI detection tool for identifying images and video created by its newest generative models.
• The tool includes rate limits that Engadget described as unexplained, raising questions about scalability.

Samsung is tipped to hold its July Galaxy Unpacked in London on July 22, with a wider Fold variant and AI-powered glasses joining the lineup.

• Samsung's next Galaxy Unpacked is tipped for July 22 in London, a break from its usual US/South Korea rotation.
• Expected devices include the Galaxy Z Fold 8, Z Flip 8, and a new Z Fold 8 Wide with a 4:3 display and S Pen support.

A new AI image tool and a hijacked support bot show how Meta's AI features can be turned against ordinary accounts.

• CNET reported that Meta's new AI image tool (tied to 'Muse' AI image/content-reuse capabilities) can be used to create deepfake content of Instagram users, and recommends opting out of a relevant AI setting.
• Cybersecurity researchers found attackers exploited Meta's AI support assistant to hijack high-profile Instagram accounts, including the former White House account (2.4M+ followers), a senior U.S. Space Force member, and Sephora.

Rising costs of third-party AI models are pushing Microsoft to route some prompts through its own in-house models instead.

• Microsoft is reportedly using its own AI models for some prompts in Excel and Outlook to reduce costs associated with Claude and ChatGPT.
• Both ChatGPT for Excel and Claude for Microsoft Office reached general availability in 2026, each priced around $20/month at entry tier.

António Guterres used the inaugural UN Global Dialogue on AI Governance to issue his strongest call yet against lethal autonomous weapons—on the same day his 2026 treaty deadline passed with no agreement in sight.

• UN Secretary-General Guterres called lethal autonomous weapons 'morally unacceptable' and demanded a binding international ban at the inaugural UN Global Dialogue on AI Governance in Geneva.
• Guterres's own 2023 deadline for a treaty banning 'killer robots' expired in 2026 with no treaty and no formal negotiations ever begun.

A provocative headline raises questions about whether political pressure could threaten AI companies like Anthropic, but the full story remains unclear.

• Gizmodo published a provocative piece suggesting Trump could become an adversary to AI companies, possibly including Anthropic's Claude.
• The article body was not captured, so the specific claims and evidence remain unclear.