1. A brutal sell-off in US Treasuries accelerated earlier today, driving 10-year yields to their highest since 2002 and pushing government borrowing costs around the world back to multiyear highs. Yields on 10-year Treasuries climbed 0.04 percentage points to 5.34 per cent, a level last seen almost a quarter of a century ago, as traders warned that the $32 trillion US government bond market, an anchor for global finance, was caught in a “vicious loop” of selling. The pressure on Treasuries reverberated across global markets, with yields on the 10-year gilts rising 0.08 percentage points to 5.51 per cent, the highest level since 2007, and 30-year gilt yields passing 6 per cent for the first time since 1998. (Source: ft.com)
2. An international team of researchers has assembled one of the largest single-cell genetic atlases ever built, profiling more than ten million immune cells from 1,108 Finnish individuals to reveal, in unprecedented detail, how disease-associated DNA variants exert their effects on gene regulation. The study, published in Nature, combines simultaneous measurements of chromatin accessibility and gene expression in the same cells, allowing scientists to trace the full molecular cascade that connects a genetic variant to altered immune function. The work offers testable mechanistic explanations for more than half of known immune disease associations, a milestone in the long-standing effort to convert statistical genetic discoveries into biological understanding. (Sources: bioengineer.org, finngen.fi, broadinstitute.org, nature.com)
3. Proteins designed by artificial intelligence could soon carry a hidden signature of their machine-generated origins, thanks to a watermarking system described today in Nature. The technology, developed by researchers at Google DeepMind in London, borrows a trick that is already used to identify AI-made content across images, video, audio and text. It involves weaving a faint statistical tell into both the amino-acid sequence and 3D shape of computer-designed proteins — without noticeably compromising their function. Known as SynthIDBio, the approach could distinguish AI-generated proteins in biological repositories from their natural counterparts, helping to preserve the integrity of databases that underpin both scientific research and biosecurity screening. (Source: nature.com)
4. Google has released a new version of its most advanced AI model that it claims matches or beats rivals Anthropic and OpenAI in coding and cyber security, as the tech giant attempts to recover its position as a leader in the sector. The company on Wednesday claimed Gemini 4 “Argon” had state of the art performance on industry benchmarks for coding, corporate work, science and maths, and cyber security. Parent company Alphabet is counting on these abilities to galvanize its market share in enterprise AI, the most lucrative use case for the technology. (Source: ft.com)
5. Google posted the news on its blog. The most interesting (and, in our view, most important) paragraph in that blogpost:
Quantum algorithmic optimization: Argon is helping our quantum computing researchers optimize the spacetime resources (qubits × gates) of subroutines that bottleneck important applications. In one example, it beat the published baseline by 40% in a matter of minutes. (Source: blog.google)
6. Meta’s Breakthrough:
