Published Aug 8, 2024

📅 ThursdAI - Aug8 - Qwen2-MATH King, tiny OSS VLM beats GPT-4V, everyone slashes prices + 🍓 flavored OAI conspiracy

Alex Volkov and guests delve into cutting-edge AI advancements, showcasing superior open-source models, innovative context caching techniques, and the competitive landscape of AI challenges and conspiracy theories.

Episode Highlights

Topics covered

Episode Highlights

AI Trends

The episode kicks off with discussing the latest buzz in AI tools and trends. He highlights the NeurIPS Hacker Cup 2024, a competition designed to push coding agents to their limits, and the introduction of Cursor's new feature, Composer, which enhances multi-file editing in VS Code 1. joins the conversation to discuss OpenAI's new structured outputs in their API, which improve reliability and reduce costs for developers 2.

Fine-Tuning

and Junyang Lin3. The conversation also touches on the impressive benchmarks achieved by smaller models, like an 8 billion parameter model outperforming GPT-4 Vision on OCR tasks 4.

📅 ThursdAI - Aug8 - Qwen2-MATH King, tiny OSS VLM beats GPT-4V, everyone slashes prices + 🍓 flavored OAI conspiracy

Topics covered

Popular Clips

Episode Highlights

AI Model IntegrationAlex Volkov dives into the latest advancements in AI tools and trends, highlighting new features and performance benchmarks. He also discusses the fine-tuning techniques that are pushing the boundaries of AI capabilities.

AI Model Integration

AI Trends

Fine-Tuning

Open Source Advancements

Podcast HighlightsThe episode explored groundbreaking AI innovations, including Alibaba's Qwen2-Math-72B-Instruct and OpenBMB's MiniCPM-V 2.6, alongside the competitive coding challenges of the NeurIPS Hacker Cup 2024 and intriguing AI conspiracy theories.

Podcast Highlights

Contextual InnovationsNext, Alex Volkov explores the latest advancements in context caching and structured outputs, highlighting their potential to revolutionize AI efficiency and reliability.

Contextual Innovations

Related Episodes