AI hotline, Anthropic-OpenAI test deal signal forced cooperation

The article frames three seemingly separate AI news items through the lens of the 1984 film Ghostbusters: "dogs and cats living together" as a sign that things are working, not ending. In each case, parties expected to be rivals are quietly collaborating.

In AI models, OpenAI and Anthropic — described as the Coke and Pepsi of AI — are close to a deal to stress-test each other's frontier models, even as both prepare for mega-IPOs. The author sees this as an attempt to create a duopoly and to pace frontier model development, a slowdown he calls Blip 3.0, partly to help governments feel more comfortable regulating.

On the geopolitical front, the US and China agreed to set up an AI hotline ahead of Thursday's Trump-Xi meeting, following a meeting between Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng. The author reads the hotline as a signal that the two sides remain far apart, and cautions against framing AI as a Cold War-style existential race. He argues AI is not a button-pressing doomsday technology and that the PC-to-smartphone transition took forty-five years, while leaders position themselves as if it will happen in three to five.

In e-commerce, Amazon is blocking Meta's Muse AI agent from shopping on its site. The author calls it an opening bid in negotiations over how AI agents will transact, noting that agents could grow the overall pie but each company wants its share. The stakes hinge on whether these collaborations are genuine or tactical, and on how quickly regulators and companies adapt to agents acting on behalf of users.