AI articles and tutorials
In-depth explainers and tutorials on LLMs, AI agents, tools and the latest model releases.
- Building Complex AI Systems with LangChain Deep Agents — Learn how LangChain's Deep Agents harness provides a secure, portable framework for building complex AI agents with sandboxing and human-in-the-loop gating.
- RSIAgent: Unlocking Recursive Self-Improvement with Kimi K3 and GLM-5.3 — Explore how RSIAgent, developed by Aether AI and academic partners, uses autonomous exploration to enable AI models like Kimi K3 to recursively self-improve.
- DeepSeek V4.1 Flash vs GPT-6 Astra: Frontier AI for Less — Explore how DeepSeek V4.1 Flash provides frontier-level AI performance at a fraction of the cost of GPT-6 Astra for business automation and scaling.
- Exa and the Rise of the Agentic Web Search Engine — Learn how Exa is building a programmatic search engine designed for AI agents rather than humans, solving the retrieval problem for the next era of AI.
- How MCP Apps and the Agentic Web are Replacing Browser Tabs — Explore how Liad Yosef and the MCP Apps spec are transforming the web from human-centric dashboards to agent-ready interfaces for the AI era.
- Why Coding Agents Are Not the Blueprint for Knowledge Work — Explore Benjamin Clavié's argument on why coding agents succeed while general knowledge agents struggle, and how 'searcher agents' can bridge the gap.
- The Hidden Cost of Poor AI Performance in Apple's Siri — Explore why bad AI performance leads to user abandonment and how Apple must rebuild trust to keep Siri relevant in the age of LLMs.
- Securing Python Dependencies: Using uv to Block Supply Chain Attacks — Learn how to use uv to protect your Python and JS projects from supply chain attacks via version pinning, age limits, and locked CI/CD syncs.
- How to Build a Profitable AI SaaS Using the Service-to-Software Ladder — Learn Dave Ebbelaar's three-step framework for building a B2B AI SaaS by moving from high-touch services to scalable, self-serve software products.
- Building a Full-Stack GenAI Document Copilot with FastAPI and Supabase — Learn how to build a production-ready RAG application using FastAPI, React, and Supabase to query SEC filings with grounded answers and citations.
- ServiceNow's Midam Kim on the Linguistic Map for Voice Agents — Learn how Midam Kim of ServiceNow uses a linguistic grid of channels and levels to diagnose and fix structured failures in voice AI agents.
- Optimizing AI Behavior with Amit Desai's OUCH Heuristic — Learn how Amit Desai of Roku reduces AI user pain by optimizing behavior thresholds (Act, Confirm, Stop) rather than just chasing model accuracy.
- Lyft Boosts Agent Resolution by 16% with LangGraph and LangSmith — Learn how Lyft used LangGraph's meta-agent architecture and LangSmith's observability to cut agent build time from six months to two weeks.
- Understanding Flow Matching in Generative AI Models — Learn the fundamentals of flow matching, a powerful alternative to diffusion models, and how it uses velocity training to transform noise into images.
- Gradium and the Quest for Full Duplex Voice AI — Learn why current voice AI feels like a walkie-talkie and how Gradium is working to enable 'full duplex' conversations with simultaneous speaking and listening.
- Engineering Voice-First AI: How Tolan Solves Latency and Context — Learn how Tolan, a voice-first AI companion, manages 4 million hours of conversation by optimizing latency, turn detection, and dynamic context reassembly.
- Why Voice AI Reliability is a Massive Scaling Risk — Sumanyu Sharma of Hamming AI explains why voice agent errors are more dangerous than local crime due to their centralized, rapid scaling potential.
- Creating Motion Graphics with Claude Opus 5 and Higgsfield AI — Learn how to combine Claude Opus 5's reasoning with Higgsfield AI's video generation to create professional motion graphics, UI demos, and product ads.
- Google DeepMind's Approach to Native Speech-to-Speech AI — Learn how Google DeepMind is evolving speech AI from fragmented pipelines to natively multimodal models that balance latency, intelligence, and modality.
- Beyond GraphRAG: Exploring Runtime Graph Repair and Agentic RAG — Discover the next evolution of RAG with 'Cognition on Graph,' a research breakthrough focusing on runtime graph repair and bidirectional synergy.
- Building AI Agents with LangChain: Deep Agents and Agent Builder — Learn how to build AI agents using LangChain's Deep Agents for developers and Agent Builder for no-code users, plus how to monitor them with LangSmith.
- Building iPhone Apps with X1: The AI Tool for App Store Success — Discover how X1 uses AI to simplify the entire iPhone app lifecycle, from initial design to navigating the complex Apple App Store submission process.
- Alvin Wang Graylin on why the US-China AI arms race is a myth — Explore why Alvin Wang Graylin argues the US-China AI competition isn't a zero-sum race, how open models prevent monopolies, and what this means for the future.
- Securing AI Agents: PostHog's Battle with Prompt Injection — Learn how PostHog's Sarah Sanders audited the 'Wizard' agent, discovering how prompt injections can turn helpful AI tools into dangerous malware.
- Deploying LangChain Managed Deep Agents to Slack — Learn how to move AI agents from LangSmith Studio to production using Slack as the interaction layer for team collaboration and live monitoring.
- Why Oracle's Kay Malcolm Says Your AI Agent Needs a Unified Database — Learn why fragmented databases kill AI productivity and how Kay Malcolm's anatomical approach to AI memory can turn individual speed into team success.
- Anthropic's Strategy for Giving LLM Tokens Specific Jobs — Learn how Anthropic's Katelyn Lesse and Angela Jiang optimize agent performance by assigning tokens to roles like advising, grading, and dreaming.
- Integrating Clay AI Agents with LangSmith and Vercel AI SDK — Learn how Clay used the Vercel AI SDK and LangSmith to achieve one-line observability, cost tracking, and Snowflake data exports for their AI agents.
- DeepSeek V4.1 Flash: Revolutionizing Long-Context AI Memory — Learn how DeepSeek V4.1 Flash reduces KV-cache memory to 890 bytes per token to enable massive context windows for agents and huge codebases.
- ECDYSIS: Solving Runtime Harness Failures in LLM Agents — Learn how the ECDYSIS method helps LLM agents distinguish between their own logic errors and runtime harness failures to prevent overfitting.