Build Wiz AI Show

While Large Language Models excel at creative tasks, they often struggle with the logical precision required for symbolic planning. This episode explores PDDL-INSTRUCT, a novel framework that teaches LLMs to reason through complex plans using a logical "chain-of-thought" approach, verifying each step to ensure its validity. Tune in to learn how this method dramatically improves planning...

In this episode, we explore self-consistency, a novel strategy that significantly improves how large language models perform complex reasoning. The method builds on chain-of-thought prompting by generating multiple diverse reasoning paths for a single problem instead of just one. By simply selecting the most consistent answer from these different lines of thought, this unsupervised technique drama...

AI is being adopted at a record-breaking pace, far exceeding previous technologies like the internet. However, this new report reveals that the AI revolution is starkly uneven, with usage heavily concentrated in high-income countries and among a narrow set of specialized business tasks. We'll unpack the data on who is using AI and how—from collaborative augmentation to full automation—and expl...

This episode unpacks groundbreaking research into how hundreds of millions of people are actually using ChatGPT. Contrary to popular belief, non-work-related messages have surged to over 70% of all use, with the most common topics being "Practical Guidance," "Seeking Information," and "Writing". We explore the surprising user demographics, including a closing gender gap, and what these patterns re...

Welcome to "AI Unpacked," your guide to the fascinating world of Large Language Models! In this episode, we'll break down the  core concepts, techniques, and critical challenges  shaping LLMs, from their internal workings like tokenization and attention mechanisms to real-world deployment considerations. Join us to deepen your understanding of these  revolutionary AI systems

In this insightful episode, Sam Altman and Vinod Khosla delve into the world beyond 2035, discussing the astonishing rate of technological change driven by AI and its profound implications for industries and jobs. They explore the rapid acceleration of AI capabilities, the vision for AI as a "default personal AGI," and its potential to revolutionize enterprise operations and scientific d...

Join us as we explore LLM knowledge distillation , a groundbreaking technique that compresses powerful language models into efficient, task-specific versions for practical deployment. This episode delves into methods like TinyLLM and Distilling Step-by-Step , revealing how they transfer complex reasoning capabilities to smaller models, often outperforming their larger counterparts. We'll discu...

In this episode, we delve into why language models "hallucinate," generating plausible yet incorrect information instead of admitting uncertainty. We'll explore how these overconfident falsehoods arise from the statistical objectives minimized during pretraining and are further reinforced by current evaluation methods that reward guessing over expressing doubt. Join us as we uncover...

Join us as we unpack "Attention Is All You Need," a pivotal paper introducing the Transformer, a novel neural network architecture. This groundbreaking model redefines sequence transduction by relying solely on attention mechanisms, completelydispensing with recurrence and convolutions. Discover how it achieves superior quality, greater parallelizability, and significantly faster trainin...

In this episode, we dive into LoRA, a groundbreaking technique that makes fine-tuning massive language models like GPT-3 more accessible and efficient. Discover how this method drastically reduces the number of trainable parameters and GPU memory needed, all without adding any extra delay during inference. We'll explore how LoRA freezes the original model and injects small, trainable matrices,...

Welcome to a deep dive into Large Language Model fine-tuning , covering everything from foundational concepts to cutting-edge advancements. This episode explores diverse methodologies like supervised and instruction-based fine-tuning, alongside advanced techniques such as Low-Rank Adaptation (LoRA) and Mixture of Experts (MoE). Tune in to understand the comprehensive seven-stage pipeline for fine-...

Large Language Models offer incredible power, but their immense scale creates significant deployment challenges in resource-constrained environments. Join us as we explore the pivotal field of LLM compression, discussing techniques like quantization, pruning, and knowledge distillation to make these models efficient and accessible for real-world applications.

Join us as we dissect the viral MIT NANDA 'GenAI Divide' report , which controversially claims 95% of enterprise generative AI pilots deliver no ROI . We explore why most projects stall due to a critical 'learning gap' and poor integration, while a successful 5% leverage agentic AI and deep workflow customization for measurable value. Tune in to understand the nuances behind the di...

Welcome to our latest episode, where we dive into Elon Musk's XAI project, MacroHard , an ambitious end-to-end neural network operating system, and Meta's exciting new partnership with Midjourney to bring aesthetic AI to billions of users. We also explore groundbreaking developments like Mirage 2 , a real-time generative world engine, and the Figure robot's dextrous towel-folding abili...

Tune into our latest episode where we dive deep into Task-in-Prompt (TIP) adversarial attacks , a novel class of jailbreaks that cleverly embed sequence-to-sequence tasks within prompts to bypass LLM safety safeguards. We'll explore how these attacks successfully generate prohibited content across state-of-the-art models like GPT-4o and LLaMA 3.2, revealing critical weaknesses in current defen...

In this episode, we'll explore why prompt engineering is a critical and essential skill for anyone looking to harness the power of AI. Discover how crafting effective input instructions directly influences the quality and relevance of outputs from models like ChatGPT and Claude, unlocking game-changing productivity . Tune in to master communicating with AI and maximize its potential in your da...

Join us for an insightful episode where Andrew Ng shares lessons from AI Fund on building startups faster with AI . Discover how new AI technologies enable unprecedented execution speed and explore best practices like working on concrete ideas, rapid engineering, and accelerating product feedback. This episode offers critical advice for entrepreneurs aiming for higher odds of success in the AI era...

Tune in to explore LightRAG , an innovative system designed to overcome the limitations of traditional Retrieval-Augmented Generation. Discover how it integrates graph structures into text indexing and uses a dual-level retrieval paradigm to enhance contextual understanding and retrieval efficiency. We'll discuss its comprehensive information retrieval, efficient operation, and rapid adaptatio...

Join us as we explore the dual-edged sword of Large Language Models (LLMs) in cybersecurity . This episode delves into how LLMs are revolutionizing threat detection, malware analysis, and defense operations , while also examining the critical challenges they pose, from prompt injection and data leakage to the automation of sophisticated cyber threats. Discover the latest strategies and best practi...

Tune into our episode on Large Language Models, where we explore the intricate world of fine-tuning techniques like QLoRA, RAFT, and RLHF for specialized tasks. We'll delve into the crucial steps of data preparation, considering quality over quantity , and discuss effective evaluation metrics, emphasizing human involvement for aligning model responses with real-world expectations. Discover how...

Join us as we explore the foundational concepts of Large Language Models (LLMs) , a revolutionary advancement in artificial intelligence. Discover how these models acquire vast knowledge through large-scale pre-training and become universal tools for diverse problems. We'll delve into the core techniques, from generative model architectures to prompting and alignment methods that shape their p...

Welcome to this special episode where we unpack OpenAI's latest leap, GPT-5 , hailed as an AI that feels like conversing with a PhD-level expert on demand. We'll delve into its unprecedented capabilities in coding, writing, learning, and healthcare, showcasing its deep reasoning and enhanced reliability . Join us to discover how GPT-5 is set to transform daily workflows for individuals, bu...

Join us as we explore the booming world of agentic AI and challenge the status quo of Large Language Models (LLMs) . We'll discuss why Small Language Models (SLMs) are argued to be the future of agentic AI, offering sufficient power, inherent operational suitability, and necessary economic advantages for specialized tasks. This episode delves into the compelling case for a shift from LLM-centr...

Welcome to a new episode where we dive into the fascinating world of "deep agents," an advanced form of LLM-based agents capable of planning and acting over longer, more complex tasks . We'll unpack the four key architectural components that enable their remarkable ability to "dive deep": a detailed system prompt , a planning tool , the use of sub agents , and access to a f...

Join us as we explore the captivating journey of Dario Amodei, the outspoken CEO of Anthropic , as he navigates the high-stakes world of artificial intelligence. Discover how a profound personal tragedy fueled his drive to accelerate scientific progress, leading him to found one of the fastest-growing AI companies with a unique focus on both rapid model development based on scaling laws and critic...