Apr 9 • General discussion
AI Developer Accelerator Weekly Call - April 08 - Recording
Key Takeaways
- Participants are working on diverse AI projects, from vehicle routing optimization to audiobook narration with AI voices
- Several members are exploring advanced RAG techniques and LLM-powered data analysis tools
- There's a focus on leveraging AI for business growth, including thought leadership content creation and lead generation
- The group continues to push boundaries in AI application, tackling complex problems and sharing knowledge
Topics
Vehicle Routing Problem
- Tom Welsh is working on optimizing routes for 49 train stations across three regions
- Challenges include factoring in different transportation modes (car vs. train) and time constraints
- Brandon suggested using the concept of "strong nodes" to cluster locations, potentially reducing complexity from 49 to ~10 nodes
- A high school student solved a similar problem in 57 minutes; Brandon will share the relevant video
AI-Powered Audiobook Narration
- Richard completed a project for audiobook reading with different character voices
- Using OpenAI's voice API, but facing limitations with available voices and sentiment control
- Exploring 11Labs as an alternative, with suggestions to create multiple voice IDs per character for different emotions
- OpenAI's TTS (text-to-speech) hints feature was recommended for more nuanced voice control
RAG and Graph-Based Information Retrieval
- Bastian demonstrated an advanced RAG implementation combining semantic chunking, graph-based retrieval, and multiple rewrite strategies
- The system uses Neo4j for graph storage and implements various query improvement techniques (paraphrase, hypothetical answers, decomposition)
- Brandon suggested creating educational content breaking down different retrieval modes and strategies
AI for Business Growth
- Paul Miller is exploring ways to leverage AI for thought leadership content creation
- Considering using AI-powered deep research tools to provide valuable insights to potential customers
- Brandon recommended focusing on LinkedIn for content distribution and suggested looking into professional ghostwriting services
AI in Consumer Goods Sales
- Maksym provided updates on an AI tool for car salespeople, now onboarding 500 new users
- The system uses a combination of vector and SQL databases for different types of information retrieval
- Manual preprocessing of documents is still required for optimal performance
Next Steps
- Tom to explore the "strong nodes" concept for his vehicle routing problem
- Richard to investigate OpenAI's TTS hints feature and consider creating multiple voice profiles per character in 11Labs
- Bastian to create educational content on different RAG retrieval modes and strategies
- Paul to focus on LinkedIn for thought leadership content, considering professional ghostwriting services
- Brandon to release videos comparing app development platforms and a deep dive into Lovable and Postgres
Action Items
- Implement strong node clustering for train station routing problem; reduce 49 nodes to ~10 entry/exit points - WATCH (5 secs)
- Create LinkedIn post on deep research tools comparison (OpenAI, Perplexity, Google); link to detailed article on company website - WATCH (5 secs)
- Test Manus AI for data analysis/visualization of NYC animal complaint dataset; compare results with current implementation - WATCH (5 secs)
- Investigate OpenAI TTS API for more voice customization options; test voice ID mapping for different character emotions - WATCH (5 secs)
- Create video explaining different retrieval modes (MapReduce, Decompose, etc.) for RAG; use representative queries to demonstrate effectiveness - WATCH (5 secs)
20
3 comments
AI Developer Accelerator Weekly Call - April 08 - Recording
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
+52
2
+30
3
+20
4
+14
5
+12
Powered by