Tech Brew Ride Home Thu. 02/15 – We Already Have Gemini 1.5!
Feb 15, 2024
Google has swiftly launched Gemini 1.5, a powerful AI model ready to revolutionize video analysis. Meanwhile, a pivotal EU court ruling emphasizes that breaking encryption could violate human rights. Users are returning their Apple Vision Pros in droves, raising questions about comfort and practicality. OpenAI is developing a web search tool and adding 'memory' features to ChatGPT, while Stability AI seeks to elevate text-to-image generation. Notably, Nvidia's rise in market power showcases the impact of AI on tech giants.
AI Snips
Chapters
Transcript
Episode notes
Gemini 1.5 Launch and Performance
- Google launched Gemini 1.5 with support for 1 million tokens.
- Gemini 1.5 Pro performs similarly to Gemini Ultra and outperforms Gemini 1.0 Pro in most benchmarks.
Gemini 1.5's Large Context Window
- Gemini 1.5's context window handles larger queries and more information (1 million tokens).
- CEO Sundar Pichai is excited about this breakthrough, enabling analysis of extensive content like videos or code.
Lord of the Rings Example
- Sundar Pichai mentioned testing a 10-million-token context window, fitting the entire Lord of the Rings trilogy.
- Brian McCullough jokingly questioned if Google used it to check for continuity errors in the trilogy.