Digest · past week

1. AlphaFold Redesigned Gene-Editing Proteins, Reducing Error Risk
   Researchers used AI protein-folding software AlphaFold to identify key areas in gene-editing proteins responsible for off-target effects, then modified those areas to reduce errors. This approach aims to improve the safety of gene-editing therapies.
2. Google DeepMind Introduces Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
   Google DeepMind released three new LLMs: Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. 3.6 Flash offers improved coding, knowledge work, and multi-modal performance with 17% fewer output tokens than its predecessor. 3.5 Flash-Lite is the fastest model in the series, and 3.5 Flash Cyber is specialized for vulnerability detection and fixing.
3. Anthropic Releases Claude Opus 4.7: How to Try, Benchmarks, Safety
   Anthropic has launched Claude Opus 4.7, a new large language model with improved performance across various benchmarks. The model is available through the Anthropic API and Claude.ai, with enhanced safety measures.
4. New Opus 5 half the price and more powerful than Fable 5? Tests explode the internet, users shocked
   Anthropic officially released Opus 5, which rivals or surpasses Fable 5 in multiple benchmarks at half the price. Users have been testing its capabilities in various creative and technical tasks, from game development to 3D modeling, with many calling it a 'monster'.