Ainsider Weekly AI Newsletter vol. 46

Google Gemini 2.0 is probably the most powerful LLM

Google announced the launch of Gemini 2.0 with several key features that promise to revolutionize how developers and businesses interact with AI.

It’s very powerful with addition of build-in tools and features:
- Whole Agentic approach with use of tools
-Realtime working & streaming API with voice assistant
-Coding Assistant Jules
- Gemini model itself has tremendous updates with accuracy, reasoning and speed.

Here’s Ainsider full article Google Gemini 2.0 an it’s all possibilities:
https://ainsider.beehiiv.com/p/google-gemini-2

After Gemini 2.0 now NotebookLM has huge, powerful update which allow you to take conversations with hosts.
See details at Notebook LM X profile.

OpenAI Sora is released.
But it’s worth 200?$

OpenAI finally released the SORA AI Video model.

Video Generation:

1. Text-to-Video

2. Image-to-Video

3. Video-to-Video

Video Editing:

Storyboard Much like traditional filmmaking, Sora's Storyboard feature gives users the ability to organize and edit sequences of videos on a timeline.

Remix This feature allows for the manipulation of video elements.

Re-cut With Re-cut, users can pinpoint and isolate the most impactful frames from a video.

Loop: Ideal for creating engaging, looping content, the Loop feature trims and adjusts videos so they can repeat seamlessly.

Blend Blend allows the fusion of two different videos into one cohesive clip.

Style Presets Creators can apply and share custom styles with Style Presets.

Here’s full article about Sora AI release:
https://ainsider.beehiiv.com/p/sora-is-here

But it’s overhyped within that price of 200$ for Premium GPT? See tweet →

Pika Labs released their Pika 2.0 model 🎥

With the most important feature 'Scene Ingredients':

It'll allow you to upload images of:

- yourself or other character,

- places/background

- surrounding, items, details

And generate from it one consistent video.

Additionally, Pika 2.0 has much imporived text alignment and visuals details generation.

The fully-controled AI Video models are much closer