Skip to content
Artwork for Build Wiz AI Show

Build Wiz AI Show

Build Wiz AI

Build Wiz AI Show turns the most interesting AI papers, articles and talks into clear audio for builders.

Each episode takes one source, such as a research paper, an engineering blog or a keynote, and two hosts, Bob and Alice, unpack the core idea, the evidence, the caveats and what you can try right away.

Topics: AI agents, agentic workflows, coding agents, LLMs, AI research and AI safety.

Made with AI from the original source; every episode credits its author.

Want a paper covered?

Email buildwizai@gmail.com or find @buildwizai on X.

Play
  • 35 episodes
  • Avg 18 min
  • English
Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.
  • Yesterday · 15 min

    Self-Improvement Memory: How Agent Harnesses Keep Improving (Google)

    Xin Yu and colleagues at Google introduce Self-Improvement Memory (SIM), which helps an AI agent's harness optimizer remember why each edit was made. Bob and Alice unpack what happens when the benchmark or model changes. In this episode: - Policy model vs harness, and why an evolved harness forgets the why - Edits grouped by recurring problem, with predicted and observed outcomes - Paired coding runs, the limitations and an overfitting counterpoint Sources: Xin Yu et al.; Yoonho Lee et al.; Birgitta Böckeler; Peng Xia et al. Big AI ideas, made clear for builders. luongnv.com/podcast

  • Yesterday · 9 min

    Proactive AI Agents: Memory, UX and the Human Bottleneck (LangChain)

    Sydney Runkle of LangChain shares what her team learned building proactive AI agents, GTM Agent and Robin, that work toward goals instead of waiting for a prompt. Bob and Alice explain why memory and UX decide whether you trust them or mute them. In this episode: - Three triggers: real-world events, developer schedules, agent schedules - Your attention as the bottleneck - Memory, dreaming and check-ins worth answering Sources: Sydney Runkle (@sydneyrunkle); Valerie Chen et al.; Joon Sung Park et al.; Margaret Mitchell et al. Big AI ideas, made clear for builders. luongnv.com/podcast

  • Tuesday · 14 min

    LLMs 101: Tokens, KV Cache and Running Models Locally (Ahmad Osman)

    Ahmad Osman's LLMs 101 guide explains how large language models write one token at a time, and why the KV cache, quantization and VRAM decide what runs on your own machine. Bob and Alice unpack it with everyday metaphors and a test to try at home. In this episode: - Tokens, attention and the KV cache, the model's working memory - Prefill vs decode: the pause before the first word, then the stream - Quantization and VRAM math for local LLMs Sources: Ahmad Osman (@TheAhmadOsman); Jay Alammar; Maarten Grootendorst; Simon Willison Big AI ideas, made clear for builders. luongnv.com/podcast

  • Monday · 11 min

    AI Agent Sandboxes Explained: MicroVMs, gVisor and Containers

    Does every AI agent need its own computer? Bob and Alice unpack Josh Rosen's post on the state of the agent sandbox. In this episode: - MicroVMs, gVisor and containers: how isolation differs. - A restricted interpreter instead of a whole machine. - Snapshots, suspend-resume and shared workspaces. - Credentials, network allowlists and the lethal trifecta. Sources: Josh Rosen; Nicolas Lacasse; Anthropic; Simon Willison Big AI ideas, made clear for builders. luongnv.com/podcast

  • Monday · 13 min

    All Code Is Slop - and how to unslop it

    Can a team ship 11,000 pull requests a month without drowning in slop? Bob and Alice unpack Lauren Tan's AI Engineer talk on trusting AI coding agents. In this episode: - The trust pyramid: code, lint rules and tests before prompts. - Banning what agents misuse; every correction becomes a lint rule. - Runtime and formal verification with TLA+ and Lean. - Guests on AI proofs, code-quality research and AI code security. Sources: Lauren Tan; Prediction: AI will make formal verification go mainstream; Bill Doerrfeld; Jens Wessling Big AI ideas, made clear for builders. luongnv.com/podcast

  • November 12 · 8 min

    When your customer is an AI Agent

    What if your next customer isn't a person? Bob and Alice unpack Welcome to the B2A Economy, Olivia Moore's a16z piece on AI agents as buyers. In this episode: - AI-driven traffic and orders surging, and why agent visitors convert better. - Plug-ins and MCPs as an app store for agents, and who wins in B2A. - Coding agents as the preview, and the Amazon v. Perplexity ruling on shopping agents. Sources: Olivia Moore; Thales Group; Anthropic; PYMNTS Big AI ideas, made clear for builders. luongnv.com/podcast

  • November 11 · 9 min

    The Tech Job Market in Late 2026

    The job AI was supposed to kill is growing fastest. Bob and Alice unpack Lenny Rachitsky's report on the tech job market in late 2026. In this episode: - PM openings up 75% from their low, while engineering openings stall. - Record lows for design, user research and remote roles, and the pull of the Bay Area. - LLM and Claude top the job posts, forward deployed engineers boom, and why openings aren't hires. Sources: Lenny Rachitsky; Guillermo Gallacher; Joe Schmidt; Jeff Cox Big AI ideas, made clear for builders. luongnv.com/podcast

  • November 8 · 14 min

    Your Old Prompts Are Making Claude Worse

    Old prompt tricks like think step by step can now make Claude worse and pricier. In 16 minutes, Bob and Alice unpack Lance Martin's AI Engineer talk on getting more from frontier models. In this episode: - Prompt audits and the prompt cache: up to 72% cheaper - Evals Claude can climb, and hill climbing without overfitting - Days-long agents with Claude Managed Agents Sources: Lance Martin; Thomas Kwa et al.; Hamel Husain; Ethan Mollick Big AI ideas, made clear for builders. luongnv.com/podcast

  • November 6 · 12 min

    $5T opportunity: AI Roll Ups

    Greg Isenberg's X article makes the case for AI roll-ups: buy an established services business and run its back office with AI agents. His playbook: buy the customers and trust, build a rulebook and review checkpoints, and move margins from 5-10% toward 30-40%. Guests add balance: even Thrive needed more than off-the-shelf AI, and AI users get less accurate when they stop checking. Based on $5T opportunity: AI Roll Ups by GREG ISENBERG (@gregisenberg) · x.com $5T opportunity: AI Roll Ups by GREG ISENBERG (@gregisenberg) · x.com The Revolution in Services: General Catalyst's AI Rollup Strategy, with Marc Bhargava by Nathan Labenz · cognitiverevolution.ai OpenAI hopes to automate IT and accounting with Thrive Holdings by Nicole Kobie · itpro.com Centaurs and Cyborgs on the Jagged Frontier by Ethan Mollick · oneusefulthing.org Guests: Theo (presents The Revolution in Services: General Catalyst's AI Rollup Strategy, with Marc Bhargava), Maya (presents Centaurs and Cyborgs on the Jagged Frontier), Sam (presents OpenAI hopes to automate IT and accounting with Thrive Holdings)

  • November 5 · 12 min

    Marketing Engineers: 100x More Experiments with AI Agents

    Greg Isenberg says the marketing engineer, part marketer and part builder, is the new 10x employee. Bob and Alice unpack his playbook: write the business "brain" before any agent, connect your data, add agents one at a time, and keep the guardrails in the tools. Guests Lena, Theo and Sam bring in Sean Ellis on growth hacking, Meta's engineering post on Andromeda and NPR on labelling AI content. Plus Builder's Corner: one thing to try this week. Every claim is checked against its source. Sources are credited in the episode

  • November 5 · 6 min

    Meet the new Build Wiz AI Show

    New here? Meet Bob and Alice, who swap roles every episode, and our four recurring guests: Maya, Lena, Theo and Sam. We explain how an episode is made: each source read in full, tested against outside evidence, and every claim checked word for word. Voices are AI; sources are credited in every episode. Big AI ideas, made clear for builders.

  • November 2 · 6 min

    DeepSeek's super-agent design, explained

    AI agents keep getting better, so the new limit is your attention: most people can follow only three to five agent sessions at once. Nora and Ben explain the DeepSeek Harness team's design for a "super-agent", one agent that is always on and works on many tasks at once. A front-desk router talks to you and hands out tasks to background workers. Cheap code filters out the noise, and the agent can quietly fix bugs before anyone asks. The hosts also go through the team's early test results and the main unsolved problem, security. Source: DeepSeek Harness on X, "Exploring Super-agent Design" (1 Nov 2026).

  • October 24 · 21 min

    Pi durable

    Pi Durable is an experimental TypeScript agent harness designed to create resilient, long-running agent applications that reliably survive crashes and process interruptions. By continuously committing model turns, tool calls, and application states to persistent storage backends like SQLite or JSONL, the library ensures that interrupted tasks resume seamlessly from their exact checkpoints without duplicating side effects. Unlike traditional terminal-bound coding assistants, this foundational building block enables developers to construct proactive agents that initiate tasks from external events, support multi-client interactions, and allow human oversight near the loop. Furthermore, it integrates smoothly with diverse JavaScript runtimes and model providers, offering advanced capabilities such as background compaction, safe replay policies, and multiplayer workflows.

  • October 24 · 24 min

    Decision Model

    Decision models represent a major shift in enterprise artificial intelligence by prioritizing rapid, structured choices over traditional text generation. Instead of relying on slow large language models to produce unstructured essays, these specialized systems process input data to deliver typed outputs like choices, scores, and probabilities in a single forward pass. Platforms such as TypeSafe's Jev and various open-source alternatives introduce unprecedented speed and cost-efficiency while eliminating formatting errors entirely. By providing calibrated confidence scores alongside every decision, these models enable software architectures to gauge their own uncertainty effectively. Consequently, organizations can seamlessly combine decision models with conventional language models to establish a secure control plane for automated agents and high-frequency workflows.

  • September 14 · 24 min

    Pacing the Frontier: A Blueprint for Measured AI Progress

    Dario Amodei advocates for a strategic deceleration of artificial intelligence development to ensure that safety measures can keep pace with rapidly advancing capabilities. He highlights the dual nature of AI, acknowledging its potential to eradicate diseases while warning against catastrophic risks like bioterrorism and autonomous system failures. Amodei introduces a "pacing" framework that begins with embedded third-party evaluators to verify safety protocols within companies like Anthropic. His proposal extends to democratic and global coordination, aiming to establish common standards while maintaining a technological lead over authoritarian regimes. Ultimately, the author argues that prioritizing prudence over speed is essential to preventing misalignment and securing a beneficial future for humanity.

  • August 20 · 21 min

    The AI Agent Harness

    The concept of an AI agent harness, using the analogy of a climbing harness to describe its protective and functional role. Just as a physical harness secures a climber and organizes their gear, a software harness provides a controlled environment where AI models can operate safely and effectively. The source identifies four primary components: a system prompt for guidance, a set of digital tools, an agentic loop for autonomous task completion, and a translation layer for model flexibility. By utilizing an open-source harness, users gain greater agency and the freedom to switch between different AI providers while maintaining ownership of their data. Ultimately, the author argues that these frameworks are essential for ensuring that humans remain in control of artificial intelligence rather than being dominated by it.

  • August 18 · 22 min

    @skills: An Open Protocol for Agent Skill Delivery

    @skills, an open protocol designed to revolutionize how AI agents access and manage procedural knowledge. Current systems rely on a restrictive installation model that forces skill descriptions into a permanent "resident" prompt, leading to token waste, decreased model attention, and a "trigger lottery" where skills often fail to activate. To solve this, @skills separates the content of a skill from its persistence and its ability to auto-trigger, offering a three-tier delivery system that ranges from on-demand references to permanent residency. This protocol utilizes a simple @syntax and a file-tree architecture, allowing users to invoke thousands of skills by name without overtaxing the AI's limited attention budget. Furthermore, it integrates with a central skills hub for discovery while maintaining local-first resolution and compatibility with existing agent workflows. Ultimately, the protocol aims to make the vast "long tail" of specialized AI skills reachable, manageable, and effective for both individual developers and collaborative teams.

  • June 11 · 23 min

    Policy on the AI Exponential

    In this essay, Anthropic CEO Dario Amodei examines the growing friction between the exponential advancement of AI and the comparatively slow pace of government policy. He argues that the emergence of "Powerful AI" presents immediate, significant risks to national security, global biology, and labor stability, necessitating a transition from voluntary transparency to binding regulation. Amodei proposes a five-pillar framework that includes mandatory safety testing, pro-employment economic incentives, and the modernization of regulatory bodies like the FDA to handle AI-driven scientific breakthroughs. He further advocates for protecting civil liberties against automated surveillance and forming a democratic coalition to secure the global AI supply chain. Ultimately, the text serves as a call to action for leaders to proactively manage a technology that could soon match the transformative impact of the Industrial Revolution.

  • June 5 · 20 min

    The Rise of Recursive Self-Improvement at Anthropic

    Anthropic reports that artificial intelligence is rapidly moving toward recursive self-improvement, where models autonomously design and refine their own successors. Current data shows a massive surge in productivity, with AI systems now generating the vast majority of the company's software code and outperforming humans in specialized research tasks. While human judgment and high-level direction remain the current bottlenecks, the gap between machine and human capability is narrowing across complex, open-ended problems. This acceleration suggests a future where the pace of technological progress is dictated by available computing power rather than human labor. Such a shift offers profound benefits for science and medicine but also introduces significant risks regarding the loss of human oversight. Consequently, the organization emphasizes the urgent need for global coordination and verification systems to safely manage the transition to fully autonomous AI development.

  • May 25 · 20 min

    AlphaProof Nexus: Advancing Mathematics Research via AI Formal Proof Search

    What happens when the world’s most powerful AI starts solving math problems that have stumped humans for over 50 years?, This episode explores the debut of AlphaProof Nexus, a groundbreaking tool that cracked legendary open conjectures by pairing advanced reasoning with rigorous computer verification,,. You’ll discover how this new era of human-machine partnership is accelerating the pace of scientific discovery and fundamentally changing how we understand the mathematical universe,.

    • Transcript
Showing 1–20 of 35 episodes