Issue #26
OpenAI's GPT-Red, Thinking Machines' first open model, and Apple Intelligence clears China
Thursday, July 16, 2026
Today covers OpenAI's new offensive-security model GPT-Red, Thinking Machines' debut open model Inkling, and bets from Anthropic and Blackstone on AI implementation services. Also in this issue: Apple Intelligence gaining China approval with Alibaba's Qwen, and a report on Microsoft's sales strategy against rival AI labs.
1 of 7 · MIT Technology Review
Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer
OpenAI developed an internal tool called GPT-Red, an LLM designed to probe its models for security weaknesses. The company used it to train GPT-5.6, its latest flagship release, which it says is its most robust model to date against cyberattacks.
Read the full story →2 of 7 · MIT Technology Review
The Download: OpenAI unveils GPT-Red and heat pumps rise in the US
OpenAI has developed an internal large language model called GPT-Red, designed to function as a hacking sparring partner that stress-tests and improves the safety of its other models. The newsletter also covers the rising adoption of heat pumps in the US.
Read the full story →3 of 7 · TechCrunch AI
Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling
Thinking Machines released Inkling, its first open model, marking the company's first public product after roughly 18 months of building AI infrastructure privately. The release signals the company's focus on customizable AI rather than a single general-purpose model.
Read the full story →4 of 7 · TechCrunch AI
Anthropic, Blackstone bet the next trillion-dollar AI business is implementation, not just models
Ode has launched with backing from Anthropic and Blackstone, focused on embedding forward-deployed engineers inside enterprises to speed AI adoption. The bet reflects a view that implementation, rather than the models themselves, may become the larger business opportunity in enterprise AI.
Read the full story →5 of 7 · TechCrunch AI
Inside Ode with Anthropic, the startup betting AI services are the future of enterprise
Ode with Anthropic, a joint venture backed by Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs, is placing forward-deployed engineers inside enterprise firms to deliver AI services. Its leaders discussed the model on TechCrunch's Equity podcast.
Read the full story →6 of 7 · TechCrunch AI
Apple Intelligence approved for launch in China with Alibaba’s Qwen AI
Chinese regulators have approved Apple Intelligence for launch in China, with features powered by Alibaba's Qwen AI model. The approval clears a path for Apple's AI capabilities in one of its most important markets.
Read the full story →7 of 7 · TechCrunch AI
Microsoft is reportedly training salespeople to talk down OpenAI and Anthropic
Microsoft is reportedly training its sales staff to position its in-house AI models as more efficient and cost-effective than those from OpenAI and Anthropic. The move signals a shift toward promoting its own models over competitors it has partnered with.
Read the full story →That is today's roundup. See you tomorrow morning.
Morning Roost Issue #26