Full Summary
This Tuesday morning, Google claims its new Gemini 3.5 Flash runs four times faster than rival AI models, a development confirmed by both *Interesting Engineering* and *Business Insider*. This model, designed for coding and autonomous task execution, is rolling out immediately across the Gemini app and AI Mode in Google Search. Alibaba is also making waves, with *Startup Fortune* and *South China Morning Post* reporting new previews of its Qwen 3.7 models, Qwen3.7-Max-Preview and Qwen3.7-Plus-Preview. These models are already ranking highly globally on the crowdsourced Arena platform, scoring 13th in text and 16th in vision capabilities. Alibaba is set to announce a "heavyweight new friend" at their Cloud Summit on May 20. Meanwhile, Cursor has released Composer 2.5, an AI coding model that performs similarly to top models like GPT-5.5 but costs less than one dollar per task, significantly undercutting competitors. *AIBase* highlights its 79.8% score on SWE-Bench Multilingual. In Sri Lanka, Dialog Axiata and Rime Labs have launched Arcana, the nation's first enterprise-ready Sinhala text-to-speech model, as reported by *Adaderana Biz English*. This model aims to make digital services more accessible for the approximately 16 million Sinhala speakers. Apple has announced WWDC 2026 for June 8-12, where they're expected to unveil a revamped Siri, iOS 27, and an AI health assistant, potentially powered by Google's Gemini AI, according to *ETV Bharat*. Cloudflare's tech boss warns that their advanced AI model, Mythos Preview, is too powerful for public release without stronger safety measures, as it can combine software flaws into serious attacks. This highlights the urgent need for robust safety protocols. This surge in AI development means faster, more capable tools are becoming available, from coding assistants to specialized language models. However, the rapid advancement also raises concerns about AI safety and the need for robust safeguards before powerful models are released to the public.