🟣 Latest AI News Roundup 🏆 Gemini 2.0

🎁 Free access to AI services. A Curated Collection of AI Innovations December 2024

Every Friday, I cut through the AI hype noise.

I spotlight what matters in AI-driven creativity. You’ll see the week’s top innovations, ranked by their impact. Stay ahead and activate your creative powers.

1️⃣ Gemini 2.0

2️⃣ ChatGPT Vision

3️⃣ Meta Llama 3.3 70B

4️⃣ Midjourney Patchwork

5️⃣ Google’s Quantum Willow

6️⃣ Devin is now generally available

7️⃣ Phi-4 Microsoft’s Newest Small Language Model

8️⃣ Sora, text-to-video model. Publicly accessible on December 9, 2024

Gemini 2.0

Google has unveiled Gemini 2.0, focusing on agentic AI capabilities. The model can output images, audio, and use Google tools. It has native image generation and integrated tool use. The experimental Gemini 2.0 Flash version is now available through Google AI Studio and Vertex AI, while chat-optimized versions are accessible to Gemini users globally.

Gemini2.0 overshadowed the 12-day Open AI Marathon. You can get for free, thanks to Google, what chatGPT proposes in its paid version.

ChatGPT Advanced Voice - Vision

OpenAI has launched video and screen sharing capabilities in Advanced Voice mode for the ChatGPT mobile app, enabling real-time visual interactions with the AI assistant.

Now you can chat with ChatGPT via video and voice in real time. Screenshare while using AV for instant feedback on what you're looking at.

Meta Llama 3.3 70B

Meta’s release of Llama 3.3 70B represents a significant advancement in open-source AI models. The 70-billion parameter model delivers performance comparable to Llama 3.1 405B while being significantly smaller in sizecapabilities in certain use cases. Handles up to 128k tokens. You can use it free of charge with over models here

Outperforms OpenAI's GPT-4o and Google's Gemini Pro 1.5.

Midjourney Patchwork

This experimental multiplayer world-building tool combines language and image models with a canvas-based interface. It allows users to create and collaborate on AI image content while maintaining consistent style and character. The platform is currently in beta and requires a Google Account linked to a Discord-based Midjourney profile.

The complete in-depth guide to Midjourney's new Patchwork

Google's Quantum Willow

Google's Willow quantum chip completed a benchmark computation in under five minutes that would take current supercomputers approximately 10 septillion years. The chip represents significant progress in quantum computing technology and error correction capabilities.

Real, world-changing quantum power… closer than ever.

Devin AI Software Engineer

Devin AI is an autonomous AI software engineer designed to assist with various programming tasks. It can write, run, and test code, aiding software engineers in both personal and team projects. Devin now costs $500/month, making it too expensive for casual users. Only committed users will pay this price, which hides the tool's flaws.

Microsoft Phi-4

This 14B parameter model specializes in complex reasoning and mathematics, achieving unprecedented scores on mathematical competitions. Phi-4 scored 91.8 points out of 150 on American Mathematics Competition problems, exceeding Gemini Pro 1.5's 89.8 points. The model is currently available on Azure AI Foundry.

OpenAI's Sora

OpenAI's text-to-video model has been released to ChatGPT Plus and Pro users. The model can create seamless video loops, animate static images, extend video length, and blend different video clips.
Unfortunately, its price - $200 for 400 5-second videos - is inadequate, but you can find better-quality paid alternatives, or even free equivalents, I recommend this article.

🟣 BONUS

Get started with the best AI tools and services right away. I'm really happy to tell you all that the AI Engineer Pack now includes free access to AI platforms! I just wanted to give you a heads-up that it's really popular, so you might want to hurry!