"Bullish on claude code"

"I found Soham in our ATS"

"these things have been like RLHF to fuck"

Vibe Tunnel - https://vibetunnel.sh/

Armin Ronacher on Simon Willison’s blog - https://simonwillison.net/tags/armin-ronacher/

Amp by Sourcegraph - https://ampcode.com/

• Matt is finalizing his event for AI Demo Days.
• Juliette completed a challenging marathon with significant elevation.
• The tech news cycle is currently nuanced and interesting.
• Soham's job application saga has sparked widespread discussion.
• Roenacher's insights on Claude Code are valuable for developers.
• Git work trees can be beneficial but have their challenges.
• Environment variables can complicate development processes.
• Context engineering is crucial for effective AI interactions.
• Talent poaching is rampant in the tech industry right now.
• The competitive landscape allows companies to attract top talent easily. Cursor is refining its features to improve user experience.
• The competition between AWS and Cloudflare shapes developer tools.
• Claude Code is designed to work closely with AI models.
• AI can enhance productivity but requires careful oversight.
• Vibe Tunnel offers a novel way to manage terminal sessions.
• MCP servers need to return appropriate data for voice agents.
• Quality checks are essential for AI-generated code.
• Developers should prioritize and start small when building tools.
• The integration of tools should be user-driven, not server-driven.
• The evolution of MCP servers reflects the changing landscape of AI development.

"Context window… kaboom"

"Am I gonna get cancelled for this"

"United is not a good airline"

"Just make more money."

Takeaways

Integrations bring people together in meaningful ways.
Video podcasts are gaining popularity among Gen Z audiences.
Editing podcasts can be time-consuming but rewarding.
Family events can provide a refreshing break from work.
Cloudflare Containers offer new possibilities for developers.
Pricing strategies in cloud services can impact user choices.
AI integrations can simplify complex API interactions.
Work off-sites can enhance team bonding and productivity.
Travel experiences can shape preferences for airlines and destinations.
Effective communication is key in team dynamics.

Summary

In this conversation, Matt Carey and Wilhelm Klopp discuss various topics ranging from the importance of integrations in technology and podcasting trends to personal updates about family events and travel experiences. They delve into the technical aspects of Cloudflare Containers and the Claude Code SDK, exploring their implications for developers and AI integrations. The discussion also touches on the dynamics of work off-sites and the impact of travel on airline preferences, emphasizing the importance of effective communication and team bonding.

Chapters

00:00 Introduction and Admin Chit-Chat
02:55 Exploring Integrations and Podcast Formats
05:50 Cloudflare Containers and Their Use Cases
08:43 Cold Starts and Performance Considerations
11:41 Pricing and Competitive Landscape of Cloud Services
14:41 Integrations and AI Tools Development
17:38 Claude Code SDK and Future Developments
26:10 Navigating API Challenges
27:47 Customizing Tool Calls for Integration
28:56 The Future of API Integration
30:09 Harnessing the Power of Code SDKs
33:49 The Importance of Harnesses in Development
34:54 Building Better Tools with Specifications
37:11 The Role of Type Safety in Code Generation
40:14 Exploring the AISDK and Its Benefits
42:16 Reflections on Work Off-Sites
48:38 Travel Experiences and Airline Preferences

"Cortex podcast is God tier"

"Claude Code is the best devtool this year"
"gotta structure your codebase for AI"

Stuff we talked about:

Cortex podcast - https://www.relay.fm/cortex

Claude Code in Github - https://github.com/anthropics/claude-code-action

Shippie - https://github.com/mattzcarey/shippie

Amp - https://sourcegraph.com/amp
Claude Code Github action system prompt - https://gist.github.com/mattzcarey/dd52a8e1df710c98b44072de46dcc09a

cursor-rules-to-claude - https://github.com/StackOneHQ/cursor-rules-to-claude

Gitlip - https://www.gitlip.com/

Durable Object file system (DOFS) - https://www.npmjs.com/package/dofs

Muppet (MCP) - https://www.muppet.dev/docs

Supermemory MCP blog - https://supermemory.ai/blog/the-ux-and-technicalities-of-awesome-mcps/

Zapier Agents - https://zapier.com/agents

SPARQL - https://www.ontotext.com/knowledgehub/fundamentals/what-is-sparql/

Vibe-rules - https://www.npmjs.com/package/vibe-rules

Simple Poll - https://simplepoll.rocks/
Kolo - https://kolo.app/

TLDR:

• Matt watched Yellowstone late at night and got hyped.
• Meta's acquisition of Scale raises questions about data distribution.
• Acquisitions in tech often lead to speculation and drama.
• Cortex podcast is highly recommended for productivity insights.
• Networking events can introduce you to new circles of people.
• Shippie has introduced new features like subagents.
• Subagents allow for parallel processing in coding tasks.
• Claude Code integration with GitHub simplifies development workflows.
• Documentation should include key concepts for better understanding.
• A product for coding agents should focus on code structure and tooling.

"We're always live."

"I think evals are so hard."

"AI engineering is dead."

• Stag do experiences can lead to unexpected personal updates.
• Consumer behavior in Las Vegas highlights the allure of gambling despite its downsides.
• Podcasting can encompass news, guest interviews, and personal stories.
• Traveling to Peru offers breathtaking experiences like hiking Machu Picchu.
• AI model releases are frequent and can impact development workflows.
• AI agents face challenges in complex coding tasks.
• New tech gadgets like the Garmin Tactics watch enhance outdoor activities.
• Fitness apps often lack user-friendly features for multi-sport activities.
• Conferences provide valuable networking opportunities in the tech industry.
• Deep research can uncover useful programming techniques and tools.

AI Engineering is Dead? - Louis from Bloop.ai

Shippie - autonomous code review and QA
thinkynote.com - a word processor with AI
SimplePoll - the easiest way to do polls on Slack
Kolo - useful things with runtime data in python

Lu, Wilhelm, and Matt discuss the evolution and features of tldraw, "a very good whiteboard". They explore the office culture, community engagement through demos, and the integration of AI technologies. The challenges faced in AI development and the importance of developer relations are also highlighted, emphasizing the need for effective communication and support for users.  

There is chat about the unique challenges faced in SDK development in an era of LLMs, the creative environment of their office, the significance of physical workspaces, and the ongoing rumors surrounding acquisitions in the tech industry. They also discuss innovations in tldraw Computer, the launch of a triathlon brand, and the complexities of navigating the MCP landscape.

• Wilhelm and Matt are starting a podcast together.
• They discuss the differences in humor between the UK and the US.
• Scented trash bags are surprisingly beneficial.
• Matt's Garmin watch has malfunctioned after years of use.
• Wilhelm is adjusting to life in San Francisco, noting its high cost of living.
• They plan to include guests in their podcast to enhance discussions.
• Matt runs AI demo days to showcase startup innovations.
• Cloudflare is praised for its deployment experience.
• Wilhelm shares his challenges with deploying a Python project.
• They reflect on the importance of genuine conversations in podcasting. Cloudflare offers out-of-the-box deployment solutions that simplify the process.
• Using modal for Python applications can significantly speed up deployment.
• Redwood's branding and recent name change reflect its evolution in the tech landscape.
• MCP servers are becoming essential for continuous training and fine-tuning of AI models.
• The popularity of MCP servers can be gauged through download statistics from NPM and PyPI.
• Building autonomous agents requires a clear understanding of feedback loops and context management.
• AI adoption varies widely among developers, with experienced engineers benefiting the most.
• Technical debt now includes the absence of structured rules files in codebases.
• The integration of AI tools into development workflows is still evolving.
• Maintaining a consistent coding pattern is crucial for leveraging AI effectively.