December 2024 AI Highlights

Hey, AI Explorers! And welcome back to The AI Stream 🔥

Your next AI update is here, and it's packed with insights you won't want to miss.

This week's breakdown:

• Sonus AI has launched the Sonus-1

• Samsung Electronics invests in Rainbow Robotics

• The AI space in December 2024

• 5 AI tools worth testing

• Actionable guides we loved

Time to level up your AI game! Let's dive in 🚀

AI News

Sonus AI has launched the Sonus-1

Overview: Sonus AI has launched the Sonus-1 family of language models, featuring impressive benchmark scores across reasoning, math, and coding tasks with their flagship Sonus-1 Pro (w/ Reasoning) model.

Key Points:

• The lineup includes three variants: Mini (for speed), Air (versatility), and Pro (top performance), with an enhanced Pro version featuring advanced reasoning capabilities

• Their top model achieves remarkable scores including 90.15% on MMLU, 97% on GSM-8k math tests, and 90% on HumanEval coding challenges

• Sonus-1 Pro shows particularly strong performance in mathematical and reasoning tasks, positioning it as a competitor to leading proprietary models

• The models are now accessible through chat.sonus.ai, with plans for additional model releases focused on tackling more complex problems

Why It Matters: Sonus AI's entry into the competitive LLM space with impressive benchmark scores signals growing competition in the AI industry. Their focus on reasoning capabilities and strong mathematical performance could make them a significant player in specialized AI applications.

AI News

Samsung Electronics Invests in Rainbow Robotics

Overview: Samsung Electronics is set to become the largest shareholder in Rainbow Robotics, increasing its stake to 35% and establishing a new Future Robotics Office reporting directly to the CEO.

Key Points:

• Samsung is expanding its initial 14.7% stake acquired in 2023, making Rainbow Robotics a subsidiary in their consolidated financial statements

• The partnership combines Samsung's AI/software expertise with Rainbow's robotics tech, focusing on intelligent humanoid development

• A new synergy council will be formed to drive global robotics business strategy and market assessment

• Dr. Jun-Ho Oh, Rainbow Robotics founder and KAIST professor, will head Samsung's new Future Robotics Office

Why It Matters: This move signals Samsung's serious entry into the advanced robotics and AI space, potentially creating a powerful competitor in the humanoid robotics market. It shows how tech giants are increasingly viewing robotics as the next frontier for AI application.

AI Agent News

Cursor & Replit Updates

1. Cursor: Cursor has upgraded their tool to process over 5,000 tokens per second, significantly improving speed across their chat, composer, and agent features.

2. Replit AI: Replit's Agent now supports xAI integration, allowing developers to easily add Grok capabilities to their applications through simple prompts or direct API integration.

The One AI Thing

The AI space in December 2024

1. ChatGPT Pro: OpenAI has introduced ChatGPT Pro, a $200 monthly subscription offering users enhanced access to its most advanced models and tools, including the o1 model and o1 pro mode.

2. Sora: OpenAI released Sora, a text-to-video AI model that generates videos from textual prompts, now available to ChatGPT Plus and Pro users.

3. OpenAI o1 & o1-mini Models: OpenAI has introduced the o1 and o1-mini models, designed to tackle complex reasoning tasks in areas such as mathematics and coding, with o1-mini offering a cost-effective solution for applications requiring focused reasoning capabilities.

4. OpenAI has unveiled its latest reasoning model, o3, which promises significant advancements in multi-step reasoning and problem-solving capabilities, marking a notable leap in AI technology compared to its predecessor, o1.

5. Veo 2: Google DeepMind has unveiled Veo 2, a state-of-the-art video generation model capable of creating high-quality videos with realistic motion and up to 4K resolution.

6. Gemini 2.0: Google DeepMind introduced Gemini 2.0, a new AI model designed for the "agentic era," featuring native image and audio output capabilities and tool use.

7. Gemini 2.0 Flash: Google's Gemini 2.0 Flash is an AI reasoning model that explicitly shows its thought process, enhancing transparency in AI decision-making.

8. Gemini Stream Realtime: Google's Multimodal Live API for Gemini enables real-time multimodal interactions, allowing for more natural and interactive AI conversations that blend text, audio, and video inputs seamlessly.

9. ElevenLabs Flash: ElevenLabs released Flash, an AI-driven text-to-speech model that delivers high-quality voice synthesis with unprecedented speed.

10. Agentforce 2.0: Agentforce has updated its AI platform to version 2.0, introducing advanced autonomous agent capabilities for complex task automation.

11. MidJourney Moodboards, a new feature designed to help users organize and visualize creative concepts & allows users to curate collections of images and inspirations.

12. DeepSeek-V3: DeepSeek has released DeepSeek-V3, an open-source AI model that rivals top closed-source alternatives, offering significant improvements in performance and speed.

13. Phi-4: Microsoft introduced Phi-4, a 14-billion parameter small language model that excels at complex reasoning in areas such as mathematics, in addition to conventional language processing.

14. Devin AI by Cognition: Cognition made Devin AI publicly available, introducing an AI-powered coding assistant that helps developers write, debug, and optimize code more efficiently.

15. Quantum Chip Willow: Google unveiled its quantum computing chip, Willow, designed to advance AI capabilities through enhanced computational power.

16. Large Concept Models by Meta: Meta introduced Large Concept Models, a new class of AI models capable of understanding and generating complex concepts across diverse contexts.

17. Android XR: Google announced Android XR, a unified platform for extended reality headsets and glasses, aiming to provide users with more device choices and access to familiar apps.

18. Amazon Nova: Amazon launched Nova, an AI-powered virtual assistant designed to enhance user experience across its ecosystem of services.

19. Alibaba QwQ: Alibaba Cloud introduced QwQ, an open-source AI reasoning model with advanced analytical capabilities, along with a new multimodal image editing tool.

20. Microsoft Copilot Vision: Microsoft unveiled Copilot Vision, an AI tool that enhances web browsing by allowing its AI assistant to see and interact with the webpage a user is viewing, providing real-time insights and assistance.

Useful AI Guide

Cursor Agent

This is a full step by step guide on how to use Cursor Agent to build whatever you want with zero coding.

Content Corner

Useful Links & Resources

Soon, anyone will be able to hum a tune, describe the vibe you’re aiming for, and have an AI-powered producer turn your ideas into reality.

50+ LLM apps, AI Agents, and RAG tutorials with step-by-step instructions.

Stanford just launched STORM, an AI research tool.

New AI Tools

Top 5 AI Tools To Try

1. Perplexity > An AI search engine that gives you direct answers and cites its sources - like having a research assistant who shows their work.

2. Suno > Create original songs by describing what you want - just type your idea and get a complete song with lyrics and music in minutes.

3. HeyGen > Turn your text into professional videos using AI avatars that speak naturally in multiple languages - no filming or voice recording needed.

4. Pika > Create and edit videos using text commands - turn rough ideas into polished videos by simply describing what you want to see.

5. Framer > Build beautiful websites without code by describing what you want - AI turns your text into fully working web pages in seconds.

Before You Go

That's it for today!

That’s all for today’s issue. Thanks for reading & being here 🖖

Liked what you see? Subscribe here.

See you next time 👋

Best, Vlad