New Claude Opus 4.8: 15 Things You May’ve Missed

AI Explained

Summary

This video provides a deep dive into the 244-page Anthropic release report for Claude Opus 4.8. The analyst breaks down the model's performance improvements in reasoning and coding, while highlighting a paradoxical decline in business-related skills due to alignment training. Key discussions include the model's new 'org chart' generation abilities within Claude Code, the complexities of honesty and self-awareness in AI, and why Anthropic is effectively managing massive compute resources from multiple global providers to scale their models.

Chapters

  1. 0:00Introduction to Claude Opus 4.8
  2. 2:27Web Thinking and Redacted Blocks
  3. 3:48Honesty and Uncertainty Flagging
  4. 8:39Coding Benchmarks and Performance
  5. 14:13Chart Reasoning and Training Data
  6. 17:10Business Skills and Alignment
  7. 20:15Cyber Capabilities and Safeguards
  8. 22:25Conclusion and Final Thoughts

Topics covered

  • Honesty and Uncertainty
  • Alignment and Business Skills

+5 more topics (full content when you follow the channel)

Transcript excerpt

I've read the 244 page report that Anthropic have put out on the new Claude Opus 4.8. I've also read many of the papers cited therein and tested the model myself in real code bases and on a private benchmark. And the 15 highlights that I'll bring you span from the humorous where Anthropic dropped their plans to take Opus 4.8 ate to business school because they found that focusing on business skills led to greater dishonesty to some more safety oriented highlights where for example the new Opus is noticing it's being tested but not letting you know that it knows. The 15 points will include where Opus 4.8's performance matches Mythos to Claude's eye opening new ability within Claude code to spawn its own org charts. But fresh off the headline that Anthropic is now valued at almost $1 trillion. What's the first point to touch on? Well, actually it's what they said about Mythos, which is tha…