Agentic Systems Podcast Summaries
Agentic Systems on Yedapo: 6 summarized podcast and YouTube episodes. Each includes key takeaways, core concepts and notable quotes with timestamps.
Beyond Prompting: Building Loops That Carry the Load
AI News & Strategy Daily with Nate B. Jones
Jun 24, 2026
Instead of individual prompts, build 'loops'—recurring AI workflows that maintain memory and context. By connecting these into 'loops of loops,' you can automate complex, multi-step life and work tasks without needing to micromanage every AI interaction.
Key insight: The biggest difference between a prompt and a loop is memory; a loop is a recurring job that notices what changed since the last time it ran.
We Need An Ecosystem in AI, And Every Company Can Win A Place In It
No Priors: AI, Machine Learning, Tech, & Startups
Jun 4, 2026
Satya Nadella argues that the true value of AI lies in building 'agentic systems' that allow any enterprise to operate at the frontier. He emphasizes that companies must move beyond generalist models to create private, specialized evaluation frameworks, effectively turning their internal workflows and tacit knowledge into proprietary, balance-sheet-worthy assets.
Key insight: Microsoft built more Azure capacity in the last 15 months than in the first 15 years, shifting their internal focus from managing infrastructure to building agentic systems that manage the infrastructure automatically.
Claude OS is useless without THIS
Simon Scrapes
May 14, 2026
Don't waste time building custom dashboards that AI agents will soon render obsolete. Instead, focus on architectural modularity by chaining small, reusable skills into 'skill systems' managed by an orchestrator.
Key insight: Anthropic is actively building features to solve memory, context, and UI dashboarding, meaning current custom-built dashboards are likely a temporary and redundant investment.
Agentic AI Systems, Clearly Explained
Simon Scrapes
May 9, 2026
Agentic AI is moving beyond simple chatbots into autonomous systems that execute complex goals. This evolution relies on harness engineering—wrapping LLMs with file-based context, memory, and tool access—to transition from reactive prompts to proactive, production-grade operations.
Key insight: The core of advanced agentic systems is surprisingly simple: it is just files in folders. 'Skills,' 'memory,' and 'brand voice' are essentially markdown files that the AI reads and updates, making these systems accessible to non-developers.
How Smart People Are Using Claude Code Skills to Automate Anything
Simon Scrapes
Mar 15, 2026
Most users treat Claude Code as a glorified task-runner, creating fragmented workflows that require constant manual intervention. Simon argues for an "Agentic Operating System" where skills share a unified brand brain and evolve through a continuous learning loop. This architecture transforms AI from a simple helper into a self-maintaining department that runs business operations autonomously.
Key insight: The system utilizes a 'heartbeat' function that automatically scans for new skills and updates its own context matrix, effectively eliminating the need for manual documentation or registry updates.
The Claude Code Skills Trap (Most People Fall For This)
Simon Scrapes
Mar 6, 2026
Installing hundreds of generic marketplace skills actually makes your AI dumber by cluttering its 15,000-character system prompt budget. Simon argues that high-performance agents require a shift from skill hoarding to a curated, multi-file architecture tailored to your specific brand voice.
Key insight: Claude operates on a strict 15,000-character budget for its skills list; overcrowding this limit creates "menu noise" that forces the AI into making generic, ambiguous choices.