AI Daily Brief@2025-03-17 | Gemini Native Image Generation | Gemini Deep Research | BlenderMCP

On March 12th, Google released their brand new open-source model, Gemma 3. In the realm of open-source large language models, the most impressive achievement was DeepSeek R1, released on January 20th. So has Gemma 3 surpassed DeepSeek R1? The answer is no, because Gemma 3 was designed for a unique scenario.

On the same day, Google also launched Gemini Native Image Generation. What does "Native" mean here? It means that the image generation capability is now native to the model itself.

On March 14th, Google introduced the new Gemini Deep Research, and even better news, it's free for everyone to use!

On March 11th, OpenAI released a new set of developer tools called Responses API.

On March 13th, in a policy proposal to the Office of Science and Technology Policy, OpenAI called for the US government to ban DeepSeek and other Chinese government-supported AI models.

Now that we've covered the major company news, let's look at a recent hot topic: MCP.

One of the most exciting recent MCP services is BlenderMCP. Let's use it as an example to help you better understand what MCP is.

Now let's look at Perplexity's progress with MCP.

This service is called Perplexity Ask. If you pay for an API token on the Perplexity platform and start the Perplexity Ask MCP Server locally, you can connect this service to, for example, the Claude desktop client.

Видео AI Daily Brief@2025-03-17 | Gemini Native Image Generation | Gemini Deep Research | BlenderMCP канала Benjamin Turing