Google Delays Gemini 3.5 Pro Launch to July as It Tweaks Its New Frontier AI Model

My take: Google promised Gemini 3.5 Pro for June at I/O, and instead we got July. The delay has two layers: the technical one (they are "tweaking" coding and long-task performance using feedback from platforms like LMArena) and the human one (four senior researchers, including Gemini co-lead Noam Shazeer, have left for OpenAI and Anthropic). Having your star talent walk out exactly when you need them most is not a confidence signal. Google has infinite resources, sure, but in AI, human capital is the genuinely scarce one. Meanwhile, OpenAI already has GPT-5.6 in preview with government-approved partners. The frontier model race (the most advanced tier) is heating up fast, and every month of delay counts. 🙄