11d • General discussion
AI Developer Accelerator — Coaching Call - March 24
AI Developer Accelerator — Coaching Call - March 24
Meeting Purpose
Key Takeaways
- Self-improving AI systems are the new frontier. Brandon is building systems that eliminate human bottlenecks by using AI to self-evaluate and refine its own prompts against a clear scoring function.
- Agentic DevOps is a force multiplier. Juan is using AI agents with AWS CLI access to build infrastructure, while Scott and Patrick are collaborating on "Ironclaw," a secure, local-first assistant with audited human-in-the-loop governance.
- Biometric auth is solving real-world problems. Ty is developing "FaceGate," a web-based Face ID SDK for shared devices (e.g., kiosks), which eliminates password friction but introduces significant biometric data security risks.
- Community-led AI development is accelerating. Patrick is using community feedback on RecapFlow to train a weekly recap generator, demonstrating a powerful model for rapid, iterative improvement.
Topics
Self-Improving AI Systems
- Brandon's "Eliminate Myself" Approach: Building systems where the human is removed from the evaluation loop, forcing hyper-clarity on the scoring function. Method: Use a tool like Codex to run infinite experiments, refining prompts based on a clear rubric. Application: SOAP Narratives: AI generates narratives and follow-up questions, then grades its own output. SDLC: AI builds features in a sandbox; the goal is to automate the final evaluation step, currently a human bottleneck.
- Patrick's RecapFlow Experiment: Integrating an auto-research loop to improve a weekly recap generator. Two-Part Loop: Mechanical Validation: A fast, internal loop checks for objective criteria (e.g., correct link count, no markdown leakage). Subjective Improvement: Community comments on the recap will be fed back as training data to refine the AI's output over time.
Agentic DevOps & Infrastructure
- Juan's AWS Image-to-Image App: Building a diffusion app with a comprehensive AWS backend. Method: Using AI agents with AWS CLI access to provision infrastructure (EC2, databases, security groups). Recommendation: Use Terraform (Infrastructure as Code) to codify the infrastructure, creating a source of truth that agents can manage.
- Scott & Patrick's "Ironclaw" Project: Collaborating on a secure, local-first AI assistant. Architecture: A smart router directs tasks to the most appropriate model—local (Ollama) for simple queries, frontier (Claude Sonnet) for complex ones. Governance: The system is auditable and requires human approval via Telegram/Discord for all actions, prioritizing security over pure autonomy. Recommendation: Use Discord channels for better memory management than Telegram.
Novel Applications & Tools
- Ty's FaceGate SDK: A web-based Face ID SDK for shared devices. Use Case: Replaces password-based clock-in systems for security guards using shared tablets, eliminating password friction. Security: Stores encrypted face vectors, not images. Accuracy is 99.7%, deemed acceptable for time & attendance. Risk: Biometric data is a high-value target for hackers. Patrick advised extreme caution with hosting it.
- Scott's AI News Brief: An automated daily news digest. Process: RSS feeds → RAG database (Haiku) → Eleven Labs (Tony Stark voice) → Daily audio briefing. Offer: Scott will add anyone to the email list.
Business & Consulting
- Morgan's Cemetery Project: Stalled by unresponsive government contacts. Problem: The primary client is unresponsive, delaying the project. Strategy: Target the list of 687 other Utah cemeteries. Offer: "Almost free" licenses to early adopters to build trust and gather feedback. Marketing: Use "lumpy mail" (physical packages) to cut through digital noise and reach decision-makers directly.
- Elijah's AI Consulting Program: Developing "Embed" and "Equip" programs to help legacy businesses adopt AI. Challenge: Businesses lack understanding of AI's potential and how to structure their data for it. Advice (Brandon): Focus on solving the "context problem" by codifying the business's core outputs into agentic workflows. Resource: Brandon's threebuckets.ai framework provides a structured approach for this.
Next Steps
- Brandon: Share the threebuckets.ai framework. Add interested members to Scott's AI news brief.
- Don: Use the Stripe CLI and rely on Stripe for product/plan data to simplify key management.
- Juan: Investigate using Terraform for Infrastructure as Code.
- Morgan: Implement the "almost free" license strategy and "lumpy mail" for the cemetery project.
- Patrick: Share the "AI Personality Kit" repo. Encourage community comments on RecapFlow to provide training data.
- Scott: Rework the Ironclaw white paper to include Discord for communication. Share the updated paper.
- Ty: Share a test version of the FaceGate SDK next week.
- All: Consider reading the book Accelerando for a perspective on rapid technological change.
Action Items
- Email Brandon the Google multimodal skill link - WATCH (5 secs)
- Update Stripe integration to use CLI and store customer IDs in Stripe - WATCH (5 secs)
- Research FaceGate biometric data privacy/legal; prep test build for group next week - WATCH (5 secs)
- Email Brandon re: Vanta intro - WATCH (5 secs)
- Generate Terraform for current AWS infra; adopt IaC - WATCH (5 secs)
- Collect RecapFlow comments weekly; feed into improvement loop - WATCH (5 secs)
- Email Brandon CMUX link - WATCH (5 secs)
- Scrape LinkedIn for Utah cemetery directors' ages; prioritize outreach - WATCH (5 secs)
- Rework IronClaw guide to use Discord; share w/ group - WATCH (5 secs)
- Email Scott to add to daily AI news brief - WATCH (5 secs)
- Review threebuckets.ai framework - WATCH (5 secs)
1
0 comments
AI Developer Accelerator — Coaching Call - March 24
skool.com/ai-developer-accelerator
Master AI & software development to build apps and unlock new income streams. Transform ideas into profits. 💡➕🤖➕👨💻🟰💰
Leaderboard (30-day)
1
+18
2
+10
3
+9
4
+5
5
+3
Powered by