העולם הולך להשתנות. כל מה שקרה ב-AI השבוע!

Summary

This video covers the latest AI advancements, starting with Google Gemini's new feature, Memories, which allows the system to remember user preferences. It also highlights Claude's integration with Google Drive and the new capabilities of GPT-4, including sound and video processing. The video explores AI tools like AI avatars, music creation with Sonu, and Microsoft's new screen recording feature. It discusses how AI agents can replicate user behavior and voice and how tools like 11 Labs are pushing AI capabilities further. The host emphasizes the importance of staying updated with AI developments and offers insights on various AI platforms.

Takeaways

• 😀 Gemini introduces a new 'Memories' feature that lets the AI system remember user preferences, similar to ChatGPT's memory functionality.
• 😀 Users can interact with Gemini by telling it what they like, and it will remember those details to provide more personalized responses.
• 😀 Claude, a competing AI system, now offers integration with Google Drive, allowing users to upload and interact with files directly in the AI interface.
• 😀 ChatGPT's new 'Sound Model' will soon allow users to interact with the AI via live video, providing real-time assistance based on visual input.
• 😀 The 'Sound Model' in ChatGPT is available on both mobile and web platforms, offering a more seamless experience for paid subscribers.
• 😀 Pickle allows users to create digital avatars to represent themselves in Zoom calls, enhancing virtual meetings with personalized, AI-generated representations.
• 😀 The Pickle platform offers different subscription plans that allow users to have one, three, or unlimited avatars and varying amounts of conversation hours.
• 😀 Suno, a music creation AI, has released a new version (v4) with improved sound quality, allowing users to generate music with more precision and style.
• 😀 Suno can generate custom songs, such as a playful tune about 'Bourekas,' and users can fine-tune lyrics and music to their liking.
• 😀 AI platforms like Labs let users create personal voice agents trained to speak and respond like them, using uploaded audio samples to replicate their voice.
• 😀 Mistral, a free AI search model, offers basic functionalities like web searches and content generation, but is slower and less advanced compared to other tools like ChatGPT and Claude.

Q & A

What is the new feature in Google's Gemini system called?

The new feature in Google's Gemini system is called 'Memories' (or 'MemoRise'), which allows the system to remember important details about users, similar to the memory feature in ChatGPT.

How can users enable the 'Memories' feature in Gemini?

Users can enable the 'Memories' feature by either writing it in a message, conversing with the system over time so it learns important preferences, or manually entering preferences through the settings under 'Saved Information'.

Is the 'Memories' feature in Gemini available in all languages?

Currently, the 'Memories' feature is mainly available in English, so users must provide settings in English for it to work.

What is the difference between Gemini and Claude regarding the 'Memories' feature?

Gemini has the 'Memories' feature, while Claude does not yet support it. Claude has introduced other features like Google Drive integration, but it lacks the memory capabilities available in Gemini.

What is the new feature available in Claude related to Google Drive?

Claude now allows users to integrate their Google Drive accounts, enabling them to upload and search files directly within the Claude system, making it easier to interact with stored documents.

What upcoming AI feature could allow video interactions with AI models?

There is an upcoming feature in ChatGPT's advanced sound model that will allow users to interact with AI models through live video. Users will be able to use their camera to capture real-time footage and receive AI-generated responses based on what they see in the video.

How is the advanced sound model in ChatGPT expected to improve user interactions?

The advanced sound model in ChatGPT will allow real-time, context-aware interactions where users can simply show something through video and get detailed, accurate advice, such as troubleshooting car problems or receiving cooking guidance.

What is the Pikal system, and how does it allow users to create digital avatars?

Pikal is a system that enables users to create digital avatars for virtual meetings. The avatars can be customized with different outfits, and users can speak through them, making it appear as if they are in a physical meeting even when they are somewhere else.

What is the difference between the basic and advanced subscriptions of the Pikal system?

The basic subscription for Pikal costs $24 per year and allows up to 15 hours of virtual meetings per month, with one avatar outfit. The advanced subscriptions provide more hours and additional outfits, with the $48 per year plan offering 40 hours and the $96 per year plan providing 100 hours.

What is the purpose of the 'Create AI Agent' feature in 11 Labs?

The 'Create AI Agent' feature in 11 Labs allows users to create AI agents that can converse in their voice and style. Users can customize these agents with specific information and train them to perform tasks or assist with conversations in a more personalized manner.

What new feature was added to the Sono music creation platform?

Sono added a new feature, V4, that improves the precision and quality of generated music. Users can create lyrics and sounds with greater accuracy, and the system allows for more detailed customization of the music style and tone.

What is the 'Recall' feature introduced by Microsoft in Copilot, and how does it work?

The 'Recall' feature in Microsoft Copilot allows users to review and rewind their computer screen activity. This feature records screen actions and enables users to revisit previous actions or content, helping them remember what they did in case they need to retrieve information or correct mistakes.

What is the main limitation of the Misteral AI model mentioned in the video?

The Misteral AI model, while free and functional for searching the web and generating content, is less efficient and slower than other models like ChatGPT and Claude. It may be suitable for users who need a free alternative but is less practical for those seeking higher performance and more advanced features.