Opus 5 vs The Hype Cycle | Unbiased Breakdown
Summary
Hosts Danny and Leon review Anthropic’s Opus 5, criticizing its high friction and poor handling of existing workflows despite lower costs. They discuss the shift in AI strategy toward machine-verifiable outputs over user experience, sharing personal testing methods and tool setups. The episode concludes with career advice for a listener struggling to find software engineering jobs, emphasizing strategic goal setting and community support.
Topics discussed
Sponsors: Accenture and LinkedIn Hiring Pro
Opus 5 release and initial mixed reviews
Opus 5 benchmarks: Knowledge work and coding
Critique of AI benchmarks and code review habits
Opus 5 verbosity and workflow friction issues
Breaking existing skills with model updates
Sponsors: Perfect Bistro and Starbucks
Anthropic's lack of communication on skill changes
AI becoming less pleasant and more jargon-heavy
Preference for legacy models in production workflows
Model toggling strategy and cost efficiency
Sponsors: USAA and Identity monitoring
Value maxing vs. token maxing in AI usage
Testing new models with chessboards and UI tasks
Hardware setup for voice-controlled coding
Listener Q&A: Career strategy and open source positioning
Sponsors: The Bear and Silicon Valley Bank
Listen ad-free on Castria