Revolut has reportedly shifted about 2,000 engineers to a self-hosted GLM model for coding, making Anthropic’s Claude Code and OpenAI’s Codex options for exceptional cases rather than everyday tools. David Abitbol, who leads the company’s AI operations, is cited as saying the system runs on 56 Nvidia B200 GPUs and costs about 30 cents per million tokens - roughly one-tenth of what Revolut previously paid Anthropic and OpenAI. The account also reports a 70% improvement in engineering lead time, though it gives no details on how that measure was calculated or what comparison period was used.
The central argument is that lower costs can change how teams use AI, not merely reduce their bills. When tokens are expensive, engineers may reserve AI for difficult tasks; at the reported price, Revolut’s engineers could use it more broadly, and that shift is credited with producing gains. The experience is presented as evidence that open-weight models could become the default for routine coding, with frontier models kept for cases where their extra capability matters. Replies raise two practical questions: whether maintaining the GPU fleet and staffing it erodes the savings, and whether GLM produces code as good as leading proprietary models. One commenter argues that experienced engineers may find the difference negligible, while juniors could benefit more from Claude’s greater capability.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Astra improved Qwen’s math performance by running the network for 50 passes and editing its weights, not by simply inspecting raw weights. Similar experiments used different methods, and an attempt to improve Gemma 3 failed. (honnibal.dev)
Itch.io is expanding its filters for games and other content disclosed as using generative AI. The filters will appear in more places across the site. (bsky.app)
Sam Altman argues that people should use less RAM so memory can be redirected to AI data centers. Critics say OpenAI’s efforts to secure a large share of global RAM supplies could also drive up prices and question whether efficiency justifies restricting personal use. (techdirt.com)
LLMs often give outdated or incorrect product prices, even when they use web search. The tutorial shows how to build a Python agent that grounds answers in live data and lets code insert verified prices. (blog.apify.com)
A 1995 Pentium Pro running a modern Stockfish port is estimated to play at 3300 Elo, far above Kasparov’s peak rating. The experiment shows how advances in chess algorithms could have beaten Kasparov without Deep Blue’s custom supercomputer hardware. (nicholas.carlini.com)
Gartner says open-weight models can handle about 90% of routine enterprise agent tasks at a fraction of the cost of closed models, while frontier models remain useful for complex reasoning. It expects companies to rely more on specialized open models, making the software harnesses that route tasks between models a key source of value. (theregister.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.