AI VIDEOS
One-sentence takeaways and concise summaries of important AI videos.
How to Build Quality Gates into Agentic Coding Workflows - Nnenna Ndukwe, Qodo AI
Nnenna Ndukwe demonstrates how explicit plans, repository rules, deterministic checks and contextual AI review can expose unsafe assumptions in agent-generated payment code.
The AI Paradox: 5 Human Skills That Are More Valuable Than Ever
Nate B Jones argues that better AI increases the leverage of taste, agency, focus, resourcefulness and follow-through when people apply them to real problems.
Microsoft's New AI Is Secretly Qwen
Nick Saraev and Jack Roberts examine AI's pressure on software businesses, a reported Nicolas Cage likeness dispute and Microsoft's use of Qwen for a specialized decision model.
Building Self-Learning Loops for Your Agent - Fuad Ali, Arize AI
Fuad Ali shows how production traces, repository context, adaptive evaluators and failure-dataset experiments can form a detect, investigate, fix and verify loop for agents.
Can Your Agent Hear You Now? Building Live Voice Agents with Gemini - Thor Schaeff
Thor Schaeff demonstrates how Gemini's native audio, stateful sessions and tool calls support multilingual agents that combine voice, vision and generated media.
Agent Speedrun - Elizabeth Fuentes Leone & Sandhya Subramani, AWS
Elizabeth Fuentes Leone and Sandhya Subramani build a customer-service agent, add explicit controls and contextual skills, and deploy the same workflow with observable AgentCore sessions.
The SDLC You Build Yourself - Andy Wong, Andrei Bocan & Shane Wolf, Atlassian
Atlassian's workshop connects coding agents to organizational context, human-verified standards and bounded backlog automation, while separating demonstrated controls from future capabilities.
Agents That Own Their Inference - Du'an Lightfoot & Khaja Omer, Akamai Technologies
Du'an Lightfoot and Khaja Omer show how memory budgets, workload evaluations and latency-throughput measurements guide self-hosted inference, including optimizations that fail in practice.
AI Security Engineer Foundations + Certificate - Micah Silverman, Snyk
Micah Silverman connects familiar application-security controls to AI risks, emphasizing that prompts cannot enforce permissions and that inventories, threat models and restricted agent access must work together.
Vector Isn't Enough: Hybrid Search & Retrieval - Jeff Vestal & James Williams, Elastic
Jeff Vestal and James Williams explain how hybrid search, filtering and measured reranking improve the evidence supplied to an agent.
Filip Makraduli Builds Private Document AI with MCP
Filip Makraduli demonstrates how an MCP layer can preprocess and redact documents with self-hosted open models before returning useful context to a coding agent.
Palak Agarwal and Omar Alhait Build Structured Document Pipelines
Palak Agarwal and Omar Alhait show why usable document extraction depends on layout, schemas, source citations and task-specific evaluation, not just OCR.
Yohan Raju Builds AI Agents with Real-Time Web Data
Yohan Raju demonstrates an MCP-connected agent using live product and review data, while explaining freshness, structured datasets and collection reliability.
Nathaniel Whittemore Examines AI Mathematics and Proof Verification
Nathaniel Whittemore examines reported AI-generated mathematics and argues that proof checking, interpretation and the limits of cross-field automation matter as much as output volume.
Nyah Macklin and Jeremy Adams - Casañas Build GraphRAG for Connected Context
Nyah Macklin and Jeremy Adams - Casañas demonstrate graph-based retrieval, while a failed counting example shows why extraction, deduplication and query validation still matter.
The Pretrained Pod Examines Whether AI Proof Translation Preserves the Claim
Pierce Freeman and Richard Diehl Martinez discuss why a Lean proof compiling does not establish that its formal statement faithfully represents the original mathematical argument.
Peter Werry Builds a Relational Context Engine Beyond Similarity Search
Peter Werry demonstrates why engineering context needs structured queries and trusted validation alongside semantic retrieval, using pull requests as a working example.
Testing Claude Code Workflow Instructions with AutoResearch
AICodeKing outlines a controlled experiment for improving Claude Code workflow instructions while keeping the model, grading rules and evaluation limits fixed.
AI Science, Mathematics and Open Models - AI Search
AI Search surveys AI-assisted science and mathematical proof claims alongside new generative, embedding and simulation models.
AI Agents in Mortgage Origination - Mike Yu
Mike Yu explains why atomic tasks, existing verification processes and measurable costs make mortgage origination a promising use case for AI agents.