#ai-coding

The Register reports that Revolut says it cut its AI coding inference costs by roughly 90 percent in under a year by making self-hosted models its default rather than relying mainly on commercial services such as Claude. Speaking at the AI Engineer conference in New York, ops and infrastructure lead David Abitbol said the fintech rolled AI coding tools out to about 2,000 engineers, adopted Z.ai's open-weights GLM-5.1 and now serves GLM-5.3 on 56 Nvidia B200 GPUs. Self-hosting cost about $0.30 per million tokens, roughly a tenth of Codex or Claude Code, and he said cheap tokens let engineers hand the AI debugging and ticket work, which is where the biggest gains came from. He also called very long context windows "a scam" in practice.

About 1,900 upvotes and 340 comments in a 329,000-member community. The top reply says this is why running models locally is not a waste of time; another argues that self-hosting means a company knows exactly which model it is running, while a cloud provider can change it without notice. One commenter who says they know people at Revolut describes its engineering culture as cautious.

The post is an image, titled "Who's excited?", about the size of the upcoming Home Assistant 2026.11 release. German Home Assistant blog simon42 reports the 2026.11 beta is out with the final release planned for November 4, and that it reads like three releases at once, with HACS (the community add-on store) becoming a built-in marketplace and support for Matter cameras.

The top reply, from user frenck_nl, who says the screenshotted post is his, explains that the release is big because of a large number of contributions, that bringing HACS in was only a few pull requests, and that AI shows up in many submissions. He writes that the team has added static checks and tuned automated Copilot reviews, that his analysis of 10,000 reviews found about 40% of review comments were repeats worth automating, and that human judgment still decides. Other commenters say they'll wait for a later point release, or doubt that a large count of merged changes is good news in the AI era. About 2,600 upvotes in a 386,000-member community.

The post is a screenshot of a Linux kernel mailing-list reply in which Linus Torvalds calls the idea of "don't use AI for X" rules "idiotic" and, mockingly, proposes banning compilers and text editors for kernel submissions. He writes that he expects people "to understand what they are doing," and that this has "very little to do with what tools they use." The Next Web reports the reply came on October 13 after kernel maintainers moved to drop the "Assisted-by" tag that had asked contributors to disclose AI coding tools.

The top replies agree that quality work should be accepted regardless of where it came from, while others stress the other half of his message: contributors must understand their code. Several worry that AI makes it easy to flood open-source projects with low-quality pull requests that take maintainers time to sift. About 3,100 upvotes in a 329,000-member community.