OpenAI models are coming to Google Cloud

Google has opened its SynthID Detector website for general public use, allowing users to upload images, videos, or audio files to verify if they were generated by artificial intelligence. The platform scans media for imperceptible watermarks embedded by Gemini and partner models including OpenAI, NVIDIA, and Kakao. Apple participates in the initiative, though detector support for its creations remains in development.… https:// en.hacks.gr/i-google-diathetei -se-oloys-to-ergaleio-synthid-detector-gia-ton-entopismo-periechomenoy-technitis-noimosynis/ # Google # ArtificialIntelligence # SynthID # OpenAI

We moved our agent backend onto the Agents API a few weeks ago, and I'm curious who else has. Why we bet on it: we watched Chat Completions go from one company's endpoint in 2023 to the shape almost every provider speaks now. OpenRouter, Groq, Ollama, vLLM, even Google and Anthropic for their own models. Nobody voted, developers just had code written against it and everyone followed. Agents feel like they're at the same point, and the Agents API already has the pieces a real agent needs: agents, sessions, turns, events, files, skills. How we use it: we dropped our own SDK and use the official openai package. Our code talks to agents and sessions in the API's shape, so changing where they run is a base URL change, not a rewrite. That alone made it worth it for us. It's still in beta, so I'm