In this episode, the hosts discuss Opus V, Cursor subscriptions and usage, and the concept of graph engineering in AI workflows. They provide feedback on Opus V, compare it to 4.8, and share their experiences with Cursor subscriptions. They also delve into the principles of graph engineering and its practical application in AI workflows, highlighting the benefits of parallelized work and the optimization of AI usage. The conversation delves into the concept of graph engineering and its application in creating connections between siloed information. It then explores the idea of extremely long-term autonomous agents and the need for continuous supervision and adaptation. Finally, it introduces Grokbot, a new tool that operates as a messaging app with each contact being a different bot, allowing for chaining of tasks and interactions between bots.

Takeaways

• Opus V feedback
• Cursor subscriptions and usage
• Graph engineering and its application in AI workflows Graph engineering is about using a graph as a foreign key database of connections between siloed information.
• Graph engineering enables extremely long-term autonomous agents, allowing for continuous supervision and adaptation.
• Grokbot is a new tool that operates as a messaging app with each contact being a different bot, allowing for chaining of tasks and interactions between bots.

Chapters

• 00:00 Graph Engineering in Practice
• 25:32 Graph Engineering and Graph Rag
• 35:11 Extremely Long-Term Autonomous Agents
• 45:52 Introduction to Grokbot

The conversation covers the release of Grok 4.5, the comparison between Cursor and Claude code, and the evolution of AI coding tools. It also delves into the impact of these tools on coding workflows and the changing landscape of AI-powered development. The conversation delves into the impact of AI on job roles, the evolution of software development, and the need for market integration for software products. It explores the changing landscape of job roles due to AI, the rapid evolution of software development processes, and the necessity for a new approach to market integration for software products.

Takeaways

• Grok 4.5 release
• Cursor vs. Claude code
• Evolution of AI coding tools Impact of AI on Job Roles
• Evolution of Software Development
• Market Integration for Software Products

Chapters

• 00:00 Grok 4.5 Release and SpaceX IPO
• 02:32 Comparison: Cursor vs. Claude Code
• 03:02 Evolution of AI Coding Tools
• 39:41 The Impact of AI on Job Roles
• 41:25 Evolution of Software Development
• 01:20:11 Market Integration for Software Products

The episode delves into the concept of forward deployed engineering (FDE) and its significance, highlighting the challenges faced by established companies in adopting AI. It also explores the shift from SaaS to utilities and the impact on SaaS companies. The tool of the week, 'Loops,' is discussed in detail, emphasizing its power in automating recurring tasks and enabling self-improvement. The conversation delves into advanced prompts and loops, the concept of goals and verifiable end states, financial optimization in AI spend, the shift to agent-based workloads, the future of AI models and providers, quality assurance and AI tools, token usage and model optimization, misconceptions about AI training, government work and AI training, and upcoming episode announcements.

Takeaways

• The premise of forward deployed engineering and its role in integrating and ushering people into AI usage
• The shift from SaaS to utilities and the impact on SaaS companies, as well as the value of platforms AI models and providers are evolving rapidly, impacting market dynamics and competition.
• Balancing token optimization with model quality is a critical challenge in AI implementation.

Chapters

• 00:00 Introduction to Forward Deployed Engineering
• 08:00 Challenges for Established Companies
• 28:26 Exploring the Tool of the Week: Loops
• 38:22 Advanced Prompts and Loops
• 44:13 Financial Optimization and AI Spend
• 51:30 Future of AI Models and Providers
• 01:00:29 Token Usage and Model Optimization
• 01:06:48 Government Work and AI Training