AI model releases dominated the day: Google pushed Gemini 3.7 Flash just three weeks after 3.6, tightening its coding and dev-tooling focus, while Ollama added DeepSeek Harness support for wiring up local and cloud models from one command, underscoring how much frontier capability now runs on localhost. On the infrastructure side Docker shipped a public beta of Docker VMM, a first-party virtualization layer for Docker Desktop, and IBM launched a dedicated OpenAI consulting practice aiming to certify tens of thousands of consultants. Security teams got fresh warnings too, with a campaign hitting Salesforce Aura, Salesforce LWR, and ServiceNow through unauthenticated guest access. For engineers, the throughline is infrastructure catching up to model velocity: virtualization, local inference tooling, and consulting layers are all being rebuilt around a faster AI release cadence.

• Gemini 3.7 Flash ships three weeks after 3.6, focused on coding gains
• Ollama adds DeepSeek Harness for one-command local/cloud model setup
• Docker VMM public beta brings first-party virtualization to Docker Desktop
• IBM launches dedicated OpenAI consulting practice, plans mass consultant certification
• Go-binary campaign hits Salesforce Aura, LWR, and ServiceNow guest endpoints