An expert overview of post-training techniques for language models, covering the entire workflow from data generation and curation to advanced algorithms like Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Reinforcement Learning (RL), along with practical advice on evaluation and iteration.
This talk introduces Anaximander, a system designed to bridge the gap between traditional, GUI-driven Geographic Information System (GIS) workflows and modern, code-heavy machine learning practices. Anaximander integrates geospatial foundation models directly into QGIS, allowing experts to interactively orchestrate, run, and evaluate models for tasks like semantic segmentation and object detection on satellite imagery.
At Applied Compute, efficient Reinforcement Learning is critical for delivering business value. This talk explores the transition from inefficient synchronous RL to a high-throughput asynchronous 'Pipeline RL' system. The core challenge is managing 'staleness'—a side effect of in-flight weight updates that can destabilize training. The speakers detail their first-principles systems model, based on the Roofline model, used to simulate and find the optimal allocation of GPU resources between sampling and training, balancing throughput with algorithmic stability and achieving significant speedups.
An exploration of scientific simplification, questioning the metaphors we use to understand the brain and intelligence. This summary delves into the tension between creating useful models and mistaking them for reality, featuring insights on the mind-as-software debate, the limits of prediction versus understanding, and the philosophical underpinnings of our quest for AGI.
Thomas Wolf, co-founder of Hugging Face, discusses his journey from physics to AI, the power of open-source models to accelerate innovation, the practical challenges of productionalizing AI demos, and why the biggest opportunities for founders now lie in the application layer on top of powerful foundation models.
Daren Fields, Co-Founder & CEO of Virtual Select, discusses the future of manufacturing, emphasizing the role of AI as a tool for human augmentation, not replacement. He explores how to modernize manufacturing by combining a carbon-based workforce with silicon-based systems to prevent defects, reduce costs, and de-risk execution.
Jeff Crume and Patrick Fussell from IBM's X-Force team share a real-world ethical hacking war story, demonstrating an attack from an 'assume breach' perspective. They break down how vulnerabilities in Identity and Access Management (IAM) and legacy systems can lead to a full compromise, starting from an insider threat and escalating to domain administrator privileges through advanced C2 attacks and lateral movement.
Nikesh Arora, CEO of Palo Alto Networks, shares his unconventional journey and leadership philosophy. He provides a masterclass in building a multi-platform company through strategic M&A, explains why founders should sometimes ignore customers, and reveals how to lead with conviction while managing imposter syndrome.
Stewart Butterfield, co-founder of Slack and Flickr, shares the product frameworks and leadership principles that guided his success. He delves into concepts like "utility curves" for feature investment, the "owner's delusion" in product design, and why focusing on "comprehension" is often more important than reducing friction. He also introduces powerful mental models for organizational effectiveness, such as combating "hyper-realistic work-like activities" and applying Parkinson's Law to team growth.
A summary of the talk by Harvey and LanceDB on building a highly optimized retrieval architecture for the legal profession. It covers challenges like query complexity and data scale, the importance of evaluation, and how LanceDB's multimodal lakehouse architecture provides the necessary foundation.
David Karam, formerly of Google Search, presents a pragmatic framework for enhancing RAG systems, advocating a "quality engineering" approach. The talk progresses through a ladder of techniques, from in-memory retrieval and BM25 to custom embeddings, re-ranking, and advanced orchestration, emphasizing that the choice of technique should be driven by empirical analysis of system failures ("loss analysis") and balanced by a "complexity-adjusted impact" mindset.
Jared Kaplan, co-founder of Anthropic, explains how the discovery of predictable, physics-like scaling laws in AI training provides a clear roadmap for progress. He details the two main phases of model training (pre-training and RL), discusses how scaling compute predictably unlocks longer-horizon task capabilities, and outlines the remaining challenges—memory, nuanced oversight, and organizational knowledge—on the path to human-level AI.
OpenSAFELY offers a new paradigm for medical research by bringing analysis code to sensitive patient data, rather than exporting the data itself. This interview explores its secure, privacy-preserving platform, its custom DSL `ehrql`, and how this model enforces reproducible, hypothesis-driven science while navigating the complexities of real-world clinical data.
Will Bryk, CEO of Exa, explains why traditional keyword-based search is insufficient for AI agents and introduces a new paradigm of neural, semantic search. He demonstrates how a hybrid approach, combining neural for discovery and keyword for precision, enables AI agents to perform complex, multi-step information retrieval tasks that were previously impossible.
Researcher Robert Ness discusses the practical implementation of Causal AI, distinguishing it from correlation-based machine learning. He covers the essential role of assumptions about the data-generating process, key Python libraries like DoWhy and Pyro, the intersection with LLMs, and a step-by-step workflow for tackling causal problems.
Get updates straight to your inbox. No spam, just useful content.