14h • General discussion
AI Developer Accelerator — Coaching Call - November 3rd
AI Developer Accelerator — Coaching Call - November 03
Meeting Purpose
Key Takeaways
- AOE API Meter: Patrick demoed a custom dashboard tracking API spend (Claude, GitHub, etc.) with predictive alerts for budget overages, enabling proactive cost management.
- AI Training Pipeline: Patrick is building a pipeline to auto-generate training videos from changelogs, but corporate liability rules force a switch from flexible resellers (e.g., WaveSpeed) to hyperscalers (Google, Azure).
- Callsitter App: Ty demoed Callsitter, a new app that waits on hold for users, inspired by a 3h 46m call. It charges $0.10/min (min. $2.49, max. $8.99) after a 15-minute wait.
- Community Brain: Patrick rebuilt the meeting recap tool in-house, replacing N8N to gain full control over the multi-model pipeline and reduce processing time to ~15 minutes.
Topics
Corporate AI Adoption: Liability vs. Flexibility
- Problem: Corporate AI adoption is blocked by liability concerns, forcing a trade-off between model flexibility and legal safety.
- Patrick's Use Case: An automated pipeline for generating training videos.
- Goal: Watch product changelogs → auto-update training materials.
- Initial Stack: WaveSpeed, 11Labs, OpenRouter.
- Blocker: Corporate policy prohibits resellers due to liability gaps (e.g., WaveSpeed's $100 liability cap).
- New Stack: Azure OpenAI (image), Google Gemini (video), 11Labs (voice).
- Constraint: Models must be General Availability (GA), not pre-release, further limiting options.
AI-Powered Media Generation
- Core Tech: Minimax H3 model with a 360 Orbit LoRA from HuggingFace.
- Process: Single photo → background removal → 3D avatar video.
- Goal: Build a custom API pipeline to chain models (background removal, LoRA, upscaling) for a robust, production-ready system.
Custom Dev Tools
- Predictive alerts for budget overages.
- Per-key, per-model usage breakdown (where provider API allows).
- YAML-based provider configuration for easy expansion.
- Trigger: A 3h 46m call to a bank.
- Function: Calls a business, waits on hold, and calls the user back when an agent answers.
- Pricing: $0.10/min, with a 15-min minimum ($2.49) and $8.99 max.
- Tech: Rebuilt from N8N to gain full control.
- Process: Fathom transcript → subject-based chunking → summary generation.
- Vector DB: Uses LensDB, a portable file-based format, to simplify sharing the knowledge base.
Next Steps
- Investigate WaveSpeed's billing API for per-model usage data.
- Open-source the Community Brain pipeline.
Action Items
- Investigate Wavespeed per-model billing API for personal accounts; update AOE API Meter if available - WATCH (5 secs)
- Audit/update CallSitter wait-time data sources; fix inaccurate numbers - WATCH (5 secs)
0
0 comments
AI Developer Accelerator — Coaching Call - November 3rd
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
+1
2
+1
3
+1
4
+1
5
+1
Powered by