Anchit’s updates

Anchit’s updates from Dispatches, the newsletter of the Newspeak House 2026/27 fellowship candidates. Newest first. About Anchit.

Dispatches #1

14 October 2026 · from Dispatches #1: The Lab is here, and so is our first newsletter

I haven't been able to attend many events these past few weeks, but I have been reading from the house library, and there are several strands it's activated.

First is from Thomas Malone's Superminds, which I have been reading alongside a series about AI on The Rest Is Politics podcast. In that conversation, Rory Stewart described a world in which AI takes over jobs, people and profits thin out, and a few American corporations absorb the value. 

That is a strong argument for sovereign AI, but sovereignty has a price, and he put it at around £500 billion for Britain. The figure seems to assume we must join an arms race to build the biggest and most general model. However, Malone's account of collective intelligence points the other way: groups grow smarter through the variety of the minds in them.

That makes me wonder whether the sovereign intelligence we need might come from a variety of smaller open models fine-tuned to work alongside people. This is the entire manifesto of Mistral, which recently launched an open weights model. I would love to hear from people who have used open models in ways that further these ideas; please reach out at anchitaxom@gmail.com.