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.
Anna Goldie and Azalia Mirhoseini of Ricursive Intelligence discuss how their work on Google's AlphaChip, which used AI to design TPUs, is now being extended to automate the entire chip design process. They explain their vision for a 'designless' industry and a recursive self-improvement loop where AI designs better chips, which in turn accelerates AI development.
This session from Okta and Auth0 introduces a comprehensive framework for securing AI agents, covering identity establishment, delegated API access via Token Vault, user consent for risky operations using Asynchronous Authorization (CIBA), and integration with MCP servers.
A discussion on the evolving cybersecurity landscape, covering the persistent threat of ransomware gangs adapting with AI, the critical failures in identity security highlighted by the Zestix case, the emergence of AI agents as a new class of insider threats, and the physical-world risks demonstrated by hacking humanoid robots.
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.
Intuit CEO Sasan Goodarzi discusses the operational playbook for reinventing a 40-year-old company, from its slow transition to SaaS to its early adoption of AI. He shares insights on winning the SMB market by treating small businesses like consumers, building effective channel partnerships, and developing a platform strategy. Goodarzi also details his leadership philosophy, emphasizing that grit and curiosity are more critical than raw talent.
A detailed overview of the formal verification of Apple's iMessage PQ3 protocol using the Tamarin prover. The talk covers PQ3's hybrid cryptographic design, its post-quantum security goals like forward secrecy and post-compromise security, the powerful adversary model it resists, and the successful formal analysis of its unbounded double ratchet structure.
Joe Reis and Matt Housley, authors of "Fundamentals of Data Engineering," reflect on the book's principles three years after its publication. They discuss how the rise of AI has created both powerful tools and dangerous "bear traps" for engineers, the critical role of expertise in a world of AI-generated content, and why foundational knowledge is more important than ever.
SMTP smuggling is a critical vulnerability allowing email spoofing that bypasses standard authentication protocols like SPF and DMARC. This research presents a large-scale study on its prevalence, uncovering widespread issues due to shared infrastructure and incomplete patches, and introduces a novel, non-intrusive methodology for ethically testing private email services.
This talk introduces an "abusability analysis framework" to evaluate technologies like passkeys under interpersonal threat models, such as intimate partner violence. An analysis of 19 services revealed critical implementation flaws, including irrevocable cloned passkeys and the failure of password resets to secure accounts, which can create persistent backdoors for abusers.
An empirical study of the 'hi mum and dad' SMS scam, a new interaction-based fraud technique. The research dissects the scam's lifecycle, from initial contact to financial exploitation, by analyzing scam messages, mobile network data, and mule accounts in collaboration with a UK mobile operator. Key findings reveal the psychological principles used, the abuse of mobile and financial infrastructure, and the underlying structure of scammer networks.
Kwangjun Ahn from Microsoft Research introduces Dion, a next-generation optimizer that improves upon Muon by using amortized power iteration. Dion enables efficient, scalable training for massive models by orthonormalizing a low-rank subspace, reducing compute and communication overhead in distributed settings.
Get updates straight to your inbox. No spam, just useful content.