Post-reInvent Quiet Period: CloudFlare Code Orange, Apple-Google GenAI Deal, and OpenAI's Cerberus Partnership
A post-reInvent roundup covering CloudFlare's Code Orange response, Apple's strategic GenAI partnership with Google, and OpenAI's move toward chip independence with Cerberus. The hosts explore the transparency lessons from CloudFlare's outage response and the implications of tech giants partnering rather than building in-house AI capabilities.
The post-reInvent period is typically quiet for AWS news, but the broader cloud and AI ecosystem keeps moving. In this episode, Brett and Travers tackle three significant developments that reveal deeper trends shaping enterprise cloud strategy.
First, CloudFlare’s “Code Orange” response to their recent outages demonstrates how transparency can actually build customer confidence. Rather than corporate deflection, they detailed exactly what went wrong and how they’re simplifying their architecture to prevent future failures.
The Apple-Google GenAI partnership represents a seismic shift in AI strategy - even Apple admits they can’t build frontier models in-house. This validates the “partnership over internal development” approach for enterprises wrestling with similar build-vs-buy decisions.
Finally, OpenAI’s Cerberus chip partnership highlights the infrastructure wars behind AI services, with purpose-built chips promising 10x throughput improvements over NVIDIA GPUs.
Who should listen: Cloud architects, DevOps leaders, and technology executives making strategic decisions about AI infrastructure and vendor relationships.
Show Video
Links from today’s episode
General News
- Code Orange: Fail Small - our resilience plan… - CloudFlare details their architectural improvements to scope down failure areas and reduce outage impact after their 2024 incidents.
GenAI News
- Google and Apple sign a multi-year GenAI agreement - Apple will integrate Google’s Gemini AI into Apple Intelligence and Siri while running instances on Apple hardware in Apple-managed data centers for privacy.
- From the Verge: Apple-Google AI Deal Analysis - Additional commentary and analysis on the strategic implications of the Apple-Google GenAI partnership.
- Anthropic releases Cowork - Claude’s desktop execution capabilities extend to local operation with subagents and human-in-the-loop workflows for Max plan subscribers.
- Superhuman AI Exfiltrates Emails - Security researchers demonstrated how hidden prompts in emails could extract data from Superhuman’s AI summarization feature (issue now resolved).
- Measuring the performance of our models on real-world tasks - OpenAI’s new GDPval benchmark claims frontier models complete real-world knowledge work 100x faster and cheaper than human experts, with Claude Opus performing best.
- Pulumi Cloud updates just got 20x faster - Pulumi implemented journaling to dramatically improve cloud infrastructure update performance.
- Gemini introduces Personal Intelligence - Google’s Gemini now integrates data from Gmail, Photos, YouTube, and Search for US-based Google AI Pro and Ultra subscribers on personal accounts.
- Clawdbot - A service claiming to build AI-powered “second brain” capabilities for personal knowledge management.
- OpenAI partners with Cerebras - OpenAI partners with Cerebras to access purpose-built AI inference chips, reducing dependence on NVIDIA GPUs for better performance and cost efficiency.## GenAI News