This Day in AI Podcast EP87: Initial Reactions to OpenAI's o1 & ChatGPT Pro
57 snips
Dec 5, 2024 Explore the initial reactions to OpenAI's o1 model, celebrated for its speed and coding capabilities, while skepticism lingers over its real-world applications. Dive into the hefty $200/month subscription for ChatGPT Pro, as hosts debate its value against existing tools. Discover the potential of Anthropic's Model Context Protocol to revolutionize AI efficiency, and grapple with concerns around the pace of AI innovations in comparison to competitors like Microsoft. It's a lively discussion on the future of AI beyond simple web browsing!
AI Snips
Chapters
Transcript
Episode notes
O1 Improvements
- OpenAI's new O1 model, announced in December, is an upgrade fulfilling September's promises.
- It boasts 30% improvement over the preview version, 50% faster speed, and 34% fewer mistakes.
O1 Evaluation Challenges
- Chris and [Speaker 0] discuss the challenge of evaluating O1 without API access, relying on Twitter reactions.
- [Speaker 0] avoids O1 due to thinking time, hoping for better code performance compared to Claude 3.6.
Questionable Demos
- [Speaker 0] questions OpenAI's demo choices, like comparing historical facts or image analysis jokes, finding them unrelatable.
- They feel OpenAI's demos lack real-world context compared to Google's and Microsoft's, despite being simple.