The Software Developer's AI Cheat Sheet Revealed

In this episode of The Programming Podcast, Danny Thompson and Leon Noel are joined by special guest Santiago, a machine learning engineer and teacher of the Machine Learning School cohort, for one of the most actionable and grounded AI discussions in tech today.

We cut through the hype and dive deep into:

- How AI is (and isn’t) replacing developers

- Using AI as a co-pilot vs letting AI drive

- Vector databases, embeddings, and Retrieval-Augmented Generation (RAG)

- Model Context Protocol (MCP), agent-based systems, and why no one agrees on what an agent is

- The underhyped power of edge ML and TinyML

- Real-world agent use cases (like automating DocuSign workflows)

- Getting started with AI as a developer today — beyond building chatbots

Follow our Guest Co-host:

https://x.com/svpino

https://www.ml.school/

https://www.youtube.com/@UCgLxmJ8xER7Y7sywMN5SfWg

This episode is loaded with insights for junior devs, industry veterans, and anyone curious about building real AI-enhanced applications.

🔔 Don’t forget to like, comment, and subscribe for more developer-first conversations!

⏱️ Timestamps & Chapters:

0:00 – Will AI Replace Software Developers?

0:45 – Embracing AI as a Developer Tool

2:00 – Meet the Hosts and Guest (Danny, Leon, Santiago)

3:20 – Leon’s AI Skepticism Journey

4:30 – Why “AI Won’t Replace You, But Someone Using It Might”

6:30 – Becoming a Solo Builder at Scale with AI

8:15 – Real Examples of Using AI to Get More Done

9:45 – AI and the Power of Prototyping

11:00 – AI Co-pilot vs. You Co-piloting AI

13:00 – Future of Tools Built for AI (MCP, A2A)

14:40 – Shifting Between Developer Modes with AI

17:00 – Intro to Vector Databases & Embeddings

20:45 – What is Retrieval-Augmented Generation (RAG)?

23:00 – Large Context Windows vs RAG: Is RAG Still Relevant?

25:30 – Indexing Speed & Clustering Explained with Spice Rack Analogy

27:50 – Why Grounding Matters (Reducing Hallucinations in LLMs)

31:00 – The Importance of the Ingestion Layer

32:10 – Other Crucial AI Trends: MCP, Agent-to-Agent

33:00 – Defining AI Agents (Why No One Agrees)

35:00 – The Browser Wars of Agent Definitions

36:30 – Overhyped vs Underrepresented AI Concepts

38:00 – Specialized Tools for AI Builders

39:45 – Agents Will Be the Next Microservices

41:00 – Real-World Example: AI + DocuSign + Slack + Asana Integration

43:00 – Underhyped: TinyML and AI on Edge Devices

45:00 – Apple’s Adapter Strategy with LoRA Models

47:00 – Companies Aren’t Ready to Train Custom Models Yet

48:30 – Santiago’s Current AI Stack: Cursor, Windsurf, Tracer, and More

50:30 – Specialized Tools for GitHub Issues, Jupyter Notebooks

52:20 – Fine-Tuning Tools & Lightweight Model Training (DPO, RLHF)

54:00 – Ask Danny, Leon & Santiago: How to Get Started in AI

55:00 – First Project Ideas: RAG with YouTube, Building MCP Servers

58:00 – Easy Agent Idea: Email Filter for Promotions

59:00 – Leon’s Learning Stack: FastAI, MCP Docs, Replicate

1:00:00 – Danny’s Favorite Beginner Project: Alt Text Generator

1:00:25 – Wrap-Up and Goodbye