🚀Google Launches Experimental Reasoning AI

Google launches new experimental reasoning AI model

Key Points 

• Google’s Gemini 2.0 Flash Thinking Experimental AI model is designed for complex reasoning in programming, math, and physics.

• The model pauses to analyze and explain its reasoning before summarizing the most accurate response.

🤖News - Google has unveiled a new "reasoning" AI model called Gemini 2.0 Flash Thinking Experimental, which, as its name suggests, is still in the experimental stages.

Available on AI Studio, Google’s prototyping platform, the model is described as being ideal for multimodal understanding, reasoning, and coding. According to its model card, Gemini 2.0 Flash Thinking Experimental excels at addressing complex problems in areas such as programming, math, and physics.

🧠What it's like - Gemini 2.0 Flash Thinking Experimental builds on Google’s recently launched Gemini 2.0 Flash model and has some similarities to reasoning models like OpenAI’s o1. When given a prompt, the model pauses to consider related information, explaining its thought process as it goes. After taking the time to think, it summarizes what it believes is the most accurate response.

Logan Kilpatrick, who leads product at AI Studio, described it as "the first step in [Google’s] reasoning journey" in a post on X. Meanwhile, Jeff Dean, chief scientist at Google DeepMind, explained that the model is “trained to use thoughts to strengthen its reasoning,” marking an exciting step forward in Google’s AI development.

Instagram previews generative AI video editing feature for 2025

Key Points 

• Instagram's upcoming AI editing tool, powered by Meta's Movie Gen, will allow users to transform videos easily.

• Early previews show AI changing outfits, backgrounds, and even appearances, without distorting during rapid movements.

👨🏻‍💻News - Instagram plans to launch a generative AI editing feature next year that will let users "change nearly any aspect of your videos." 

Powered by Meta’s Movie Gen AI model, this tool aims to make video editing accessible, helping creators transform their content and bring ideas to life without needing advanced editing skills. While no official release date for the Movie Gen model has been shared, Instagram is the first platform confirmed to use this text-to-video technology. 

✨What to expect - The announcement came via a teaser video shared by Instagram head Adam Mosseri. The teaser showcases early AI capabilities, such as changing Mosseri’s outfit, altering backgrounds, and even modifying his appearance—at one point turning him into a felt puppet. 

Impressively, these AI-generated elements remain consistent and don’t distort, even with rapid movements. However, the snippets shown are brief, lasting only a second each, so, like with OpenAI’s Sora model, we won’t know if Instagram’s tools can live up to expectations until they are fully launched.