Trends

Google has introduced Gemma, an open-source family of LLMs models. With 2B and 7B model sizes, it outperforms competitors like Mistral and LLaMa. Gemma is accessible, running on laptops, and available to all developers globally, unlike OpenAI’s closed models.

NVIDIA’s market value soars to nearly $2 trillion as demand for AI chips skyrockets, making it the 3rd most valuable company in the US and 4th worldwide. With 80% control of the AI chip market, NVIDIA’s co-founder Jensen Huang is now worth $69.2B.

Google’s AI Cyber Defense Initiative is aiding the fight against cybercrime by open-sourcing Magika. This AI-powered tool aids defenders in identifying malware, transforming cybersecurity and addressing the “Defender’s Dilemma” through AI.

Researchers at Amazon have developed a groundbreaking text-to-speech model called BASE TTS, using 100,000 hours of public domain speech data. With emergent abilities, it enhances natural speaking and could push AI technology beyond the uncanny valley.

Introducing Sora, OpenAI’s groundbreaking text-to-video model that generates smooth, dynamic, and consistent videos up to 1 minute long. Sora excels in animation style, mood, camera angles, and supports various aspect ratios. With language models like GPT, Sora transforms basic prompts into powerful prompts, delivering high-quality videos. Sora also has the ability to use images and videos as inputs, allowing it to animate images, extend videos, edit scenes while preserving characters, and automatically fill in-between frames when connecting two videos. This advancement marks a significant milestone for generative AI in video production.

Gemini, Google’s new umbrella name for AI tools, combines chatbots, voice assistants, and coding assistants. Replacing Google Bard and Duet AI, it offers a versatile solution for various needs.