Sunday, May 10, 2026
On Monday, Rally hosted Erika Anderson and Yaoli Mao for a walkthrough of HumaneBench, their open benchmark scoring 800 scenarios against eight care ethics principles. Erika described model cards as nutrition facts written in a language most people can't read, placing us at the same choice point food was in before nutrition labels became standard. The room drew a crowd who agree that we should measure the impact of AI models and systems on well-being.
Two days later at Anthropic's Code with Claude developer conference, the conversation was about keeping up with a staggering change rate. The developers who win are the ones optimizing their architectures for the next intelligence change. The favorite features: doubled rate limits, Managed Agents adding orchestration, outcomes, and a "dreaming" capability that writes lessons from past sessions back into memory.
On a personal note, Friday was my last day at my corporate day job. I'm building full-time, working on Rally and the consulting. Excited to see what the future holds and I'll keep you in the loop.
See you at Builders Night tomorrow, The Commons, 550 Laguna St.
- Lee
Coming up at Rally
Builders Night
2026-05-11 · 6:30 PM - 9:00 PM
Monday co-working for AI builders. Two hours of focused build time with a share-back round to close.
AI Research Circle: Physical AI
2026-05-25 · 7:00 PM - 8:45 PM
The first wave of AI rebuilt the digital 15% of the economy. The next wave runs on the other 85%: manufacturing, energy, infrastructure. Atharva Atre (Niantic Spatial) walks through the five technical layers it runs on, from simulated training environments to spatial intelligence.
Around SF this week
Wednesday, May 13, 3 PM - 5:30 PM
**Reading Group (+🧋): Code Synthesis for Agentic Decision-Making** We're digging into two papers on how agents use code generation to reason and act: Code World Models (learning environment dynamics through executable code) and Autoharness (automated benchmark creation for LLM evaluation). Come ready to discuss how code-as-reasoning changes agent architectures, plus the practical challenges of benchmarking these systems. Bubble tea provided.
Wednesday, May 13, 3 PM - 8 PM
**Agent Building Day @ Corgi Cafe** Spend the day building AI agents alongside other SF builders — bring your laptop, your current project, and questions you're stuck on. You'll get unstructured time to code, troubleshoot architecture decisions with people who've shipped agents in production, and leave with real progress. Plus: corgis.
Wednesday, May 13, 5:30 PM - 8:30 PM
Demo nights are where you see what's actually being built right now — not what's in the roadmap. We're teaming up with The AI Collective to showcase founders shipping real products, followed by open feedback and networking with other builders. Come to watch, or bring something to show.
Wednesday, May 13, 6 PM - 9 PM
Fin (Intercom's AI agent) and Metronome (usage-based billing infrastructure) are teaming up to tackle the thorniest problem in AI products: how to price agent work when costs and value are both hard to predict. You'll hear real pricing models from companies shipping agents in production, plus the billing architecture that makes complex usage pricing actually work. If you're figuring out whether to charge per conversation, per resolution, or per API call, this is your crash course.
Wednesday, May 13, 6:30 PM - 9 PM
**SPC Demo Night** brings together founders from South Park Commons' latest batch to show what they've been building. You'll see live product demos across the AI stack — from infra tools to end-user apps — and get direct access to builders shipping in real-time. Best for engineers and founders who want to see what's actually working before it hits Product Hunt.
Friday, May 15, 5 PM - 9 PM
Sat, May 16, 9 AM - 8 PM
Two-day in-person hackathon on Notion's dev platform with Anthropic, Vercel, OpenAI, Planetscale, and Minimax sponsoring. Teams of up to four, application required.