🔥 Is Deepseek R1 REALLY better than o1? The answer might surprise you! But one thing’s for sure: with 25x more compute, you can SCALE YOUR IMPACT like never before. 🚀
🎥 Featured Links:
Benchy Codebase: https://github.com/disler/benchy
Principled AI Coding: https://agenticengineer.com/principle...
Deepseek R1 Release: https://api-docs.deepseek.com/news/ne...
Flash Thinking: https://ai.google.dev/gemini-api/docs...
o1: https://openai.com/o1/
💡 In this video, we dive deep into the world of reasoning models, comparing Deepseek R1, o1, and Gemini Flash Thinking side by side. Discover how these models think, how they solve problems, and why scaling your compute is the key to scaling your impact in the age of AI.
🔍 What you’ll learn:
How Deepseek R1’s 600B parameter model stacks up against o1 and Gemini Flash Thinking.
Why the chain of thought from reasoning models is a game-changer for improving prompts and workflows.
How to use tools like Benchy to benchmark and compare models in real-time.
Why compute is the ultimate lever for scaling your AI impact.
🚀 Whether you're an AI engineer, developer, or just curious about the future of reasoning models, this video is packed with actionable insights. From AI coding to meta-prompting, we break down the latest advancements and show you how to harness them for maximum impact.
📖 Chapters:
00:00 Reasoning Model Paradigm
01:12 Compare model thoughts side by side
03:19 Thoughts As Feedback
10:15 o1 is still king
20:54 Meta Prompting
29:25 To scale your impact, scale your compute
#promptengineering #llm #aiengineer
