Lumian Gen AI Newsletter Issue #30

⏱️ News Flash

The 2-Minute Scoop to Keep You in the Loop

1/ Google introduces Lumiere, a multimodal AI video model

Google has introduced a new AI video generation tool called Lumiere, which can create videos in one go using a unique "Space-Time UNet" architecture. This method offers more consistent and seamless videos compared to previous techniques that relied on creating keyframes and then interpolating between them. Lumiere has shown promise in user tests, outperforming competitors like Runway Gen-2, Pika Labs, and Stable Video Diffusion in quality, text alignment, and motion. There's speculation that this technology could eventually be integrated into YouTube, particularly for its video inpainting capabilities, which merge reality with generated content.

2/ Microsoft creates new AI team for small models

Microsoft is shifting its AI focus towards developing smaller, more efficient AI models with a new team called GenAI. These smaller models, known as SLMs (Small Language Models), aim to reduce costs and computing power requirements compared to larger models like GPT-4. The GenAI team is distinct from Microsoft's Turing team, which works on larger models. The move towards SLMs, exemplified by Microsoft's Phi models, is gaining interest for their potential to offer similar capabilities as larger models but with lower resource demands, making AI more accessible and economical.

3/ Sam Altman is raising money for building semiconductor factories

Sam Altman, the head of OpenAI, is planning to raise billions of dollars to build a global network of semiconductor factories, a move that diverges from the common tech industry practice of designing chips but outsourcing their production. By creating these factories, Altman aims to ensure a steady chip supply for the future, anticipating a potential shortage as AI technology becomes more prevalent. He is in talks with companies like G42 and SoftBank for funding, and Microsoft, OpenAI's major investor, is reportedly interested in the plan.

🤖 The Future of Prosumer

The Rise of “AI Native” Workflows

a16z’s latest article on the future of prosumer is a great read! In essence, founders have a unique chance to build AI-native companies, leveraging AI for enhanced creativity and efficiency in work tools. Such AI-native platforms will automate low-skill tasks, allowing users to focus on higher-level thinking.

Examples in Action:

• Vizcom for photorealistic renderings

• Durable for AI-generated websites

• HeyGen for talking video avatars

• Pika for iterative video editing

🤑 Fundraising

The (AI) Intelligent Investor

• AIDash raises $50M to spot wildfires with AI

• ElevenLabs Releases New Voice AI Products and Raises $80M Series B

• Bret Taylor's AI Firm to Get Funds at Near-$1 Billion Value

🤖 Nerd Out

Technical and Business Readings

• Artificial intelligence deep dive by Chamath

• OpenAI, Nvidia, NASA, NSF and more partner on federal program to increase access to AI resources

• The FTC is reviewing investments in OpenAI and Anthropic by Microsoft, Amazon and Google

😜 Fast but Not Furious!

Meet the AI Boss

If you were forwarded this newsletter, you can access more of our content by subscribing here.

Best,

Robin and Saswat