Tokenless

Tokenless

Session on Reasoning

Interactive discovery

Explore the topic map

Follow the connections between themes, people, and ideas across the Tokenless archive in an interactive topic modeling map.

Open the map Browse posts

Machine Learning

View All
Frontier results, on device - RL Nabors, Arize

Frontier results, on device - RL Nabors, Arize

RL Nabors discusses the significant costs associated with using frontier AI models, covering security, latency, and financial implications. She introduces a framework for right-sizing AI solutions by leveraging smaller, task-specific models and Small Language Models (SLMs). The framework details how to prove task feasibility, establish success criteria with golden datasets, conduct capability evaluations (using tools like Phoenix), and select the most appropriate "Small And Good Enough" (SAGE) model. Nabors further demonstrates how prompt engineering, particularly few-shot prompting, and post-processing can close performance gaps with larger models, while advocating for continuous regression evaluations to maintain performance integrity. The overarching message is to "prototype big, deploy small" to optimize AI deployments.

Research to Reality: Bringing Frontier ML Research to Production - Vaidas Razgaitis, Higharc

Research to Reality: Bringing Frontier ML Research to Production - Vaidas Razgaitis, Higharc

Vaidas Razgaitis, Senior Research Engineer at Higharc, shares three tactical tips to accelerate the transition of novel AI/ML research into production-ready features. He emphasizes addressing the critical handoff challenge between ML researchers and software engineers through structured documentation (Research Prototype Taxonomy Document), a well-organized monorepo utilizing decoupled microservices, and a systematic approach to code decomposition and PR review. These strategies aim to improve legibility, maintainability, and delivery speed for ML-driven products.

Uncertainty-Guided Data Augmentation for Engineers | Deep Dive - Yongmin Kwon

Uncertainty-Guided Data Augmentation for Engineers | Deep Dive - Yongmin Kwon

This session details a data-efficient method for training engineering surrogate models by using uncertainty quantification (UQ) to guide geometric data augmentation. Instead of random deformations, the approach lets the deep ensemble model identify its own knowledge gaps (epistemic uncertainty), then uses Free-Form Deformation (FFD) to generate new shapes specifically in those uncertain regions. This ensures every expensive simulation run yields maximally informative data, significantly improving model accuracy for a fixed computational budget across domains like structural mechanics and aerodynamics.

Artificial Intelligence

View All
Session on Reasoning

Session on Reasoning

This session features two talks on optimizing and verifying AI reasoning. Hongxiang Fan discusses cross-stack co-design for efficient AI, focusing on Test-Time Scaling (TTS) challenges, optimal verification granularity, and system-level optimizations for edge deployments. Nagarajan Natarajan introduces 'Advancing Verified Reasoning' with the InterVent platform, aiming to ensure AI agents comply with complex policies through formal verification, dynamic steering, and leveraging verification signals for training. Both emphasize addressing the computational and reliability costs of advanced AI.

Multimodal & Embodied Intelligence (Pt 1), Panel on Multimodal AI: Progress, Pitfalls, Possibilities

Multimodal & Embodied Intelligence (Pt 1), Panel on Multimodal AI: Progress, Pitfalls, Possibilities

This session explored Multimodal and Embodied Intelligence, featuring talks on hybrid AI in robotics (classical vs. end-to-end), AI's role in healthcare (focusing on NCDs, deployment, and uncertainty modeling), and fundamental perception challenges in multimodal reasoning (using educational video QA and visual puzzles). A panel discussed the impact of foundation models, the blurred lines between AGI and human-like AI, critical deployment pitfalls (human factors, efficiency, architectural limits), and future directions, emphasizing task-specific models and the redefinition of 'foundation models.'

Grant Sanderson (@3blue1brown) – AI and the future of math

Grant Sanderson (@3blue1brown) – AI and the future of math

Grant Sanderson and Dwarkesh Patel discuss AI's rapid but uneven progress in mathematics, exploring whether AI can achieve true conceptual breakthroughs, the challenge of measuring creativity, and the long-term implications for human understanding and the future roles of mathematicians. They delve into the unique 'grindability' of math for AI training, the potential of formalization, and why AI currently struggles with 'theory of mind' in writing, offering advice for students navigating an AI-transformed world.

Technology

View All
Plenary Talk 3​: Challenges and research opportunities for global hyperscale services

Plenary Talk 3​: Challenges and research opportunities for global hyperscale services

Jim Kleewein's talk outlines the immense challenges and critical research opportunities in building and operating global hyperscale services like Microsoft 365 and Azure. He emphasizes that at this scale, traditional approaches fail, necessitating a "new golden age of applied research" across areas like continuous availability, data management, security, and sustainability. Kleewein also discusses AI's powerful but limited role, stressing the ongoing need for human expertise, and highlights the ethical imperative to prevent failures that can have life-or-death consequences.

Platforms: Build Abstractions, not Illusions • Gregor Hohpe • GOTO 2025

Platforms: Build Abstractions, not Illusions • Gregor Hohpe • GOTO 2025

Gregor Hohpe explains the critical role of platforms in managing the growing cognitive load on developers due to complex distributed systems. He contrasts platforms, driven by "economies of speed" and fostering innovation through diversity, with traditional IT services and oversimplified abstractions that create dangerous illusions. Hohpe emphasizes building platforms that provide intuitive, domain-specific abstractions to solve real business problems, rather than just repackaging existing cloud services.

Full Stack Greenfield Projects : Are they still relevant?

Full Stack Greenfield Projects : Are they still relevant?

Bharat Goenka, co-founder of Tally, discusses the company's unconventional approach to software development through "Full Stack Greenfield" projects. He explains why building every component from scratch, despite being a high-risk strategy, has been crucial for Tally's success in serving the SMB market, fostering extreme customer loyalty, and aspiring to connect 200 million businesses. The talk delves into the historical context, the philosophy of questioning and choosing constraints, and the distinction between product and custom engineering.

Recent Post

LLM Compression Explained: Build Faster, Efficient AI Models

LLM Compression Explained: Build Faster, Efficient AI Models

Learn how AI model compression and quantization techniques are essential for optimizing Large Language Model (LLM) performance and significantly reducing inference costs in production. This deep dive covers practical examples, benefits like reduced latency and increased throughput, and strategies for different AI use cases, demonstrating how to deploy scalable AI with minimal accuracy degradation.

Agentic Data Management and the Future of Enterprise AI — with Rohit Choudhary

Agentic Data Management and the Future of Enterprise AI — with Rohit Choudhary

Rohit Choudhary, CEO of Acceldata, discusses the imminent 10x annual growth of enterprise data and how most organizations are unprepared. He introduces Acceldata's agentic data management platform, designed to make data self-aware, self-optimizing, and AI-ready. He emphasizes the 1000x cost difference of fixing data early versus late, the need for operational, real-time data governance, and why clear thinking and deep domain expertise, not just programming skills, will be most valuable in the age of AI.

Mistral: Voxtral TTS, Forge, Leanstral, & Mistral 4 — w/ Pavan Kumar Reddy & Guillaume Lample

Mistral: Voxtral TTS, Forge, Leanstral, & Mistral 4 — w/ Pavan Kumar Reddy & Guillaume Lample

Mistral's Pavan (Voxtral lead) and Guillaume (Chief Scientist) discuss the new Voxtral TTS model, its novel architecture using flow matching for efficient, high-quality speech generation. They elaborate on Mistral's strategy of delivering specialized, open-weight models and the Mistral Forge platform, which empowers enterprises to leverage their proprietary data through fine-tuning for privacy, cost-effectiveness, and superior performance. The conversation also covers Mistral Small, the future of AI for science, and the company's commitment to open-source and foundational research, including formal proving as a proxy for long-horizon reasoning.

A Typo Led to the Creation of Spring Cloud Contract • Marcin Grzejszczak & Jakub Pilimon • GOTO 2026

A Typo Led to the Creation of Spring Cloud Contract • Marcin Grzejszczak & Jakub Pilimon • GOTO 2026

Marcin Grzejszczak, a Java Champion and Spring Cloud Contract contributor, discusses how a production-breaking typo led to the creation of Spring Cloud Contract. He outlines an AI-powered future for contract testing, leveraging production traffic and OpenAPI specs to generate and validate contracts automatically. Marcin also emphasizes that 'context' is the most underrated pillar of observability, transforming raw logs, metrics, and traces into actionable insights.

What Tesla and SpaceX Teach Founders About Building Hardware | a16z

What Tesla and SpaceX Teach Founders About Building Hardware | a16z

Ex-Tesla and SpaceX engineers Chandler Luzsicza and Turner Caldwell share the practical lessons they learned about building complex hardware and how they apply them at their own startups. They cover core principles like decision velocity, strategic vertical integration, managing the critical path without creating new bottlenecks, and using aggressive milestones as a forcing function to reveal true priorities. The discussion also delves into the value of a 'factory mindset,' the importance of a rigorous hiring process, and advice for young engineers.

François Chollet: ARC-AGI-3, Beyond Deep Learning & A New Approach To ML

François Chollet: ARC-AGI-3, Beyond Deep Learning & A New Approach To ML

François Chollet discusses his contrarian approach to AI, moving beyond scaling LLMs to understand intelligence from first principles. He explains his work on the ARC benchmark series, including the new ARC-AGI V3, designed to measure 'agentic intelligence' and skill acquisition efficiency. He also introduces his lab, Ndea, which is developing a new ML paradigm based on symbolic models, and shares his perspective on the limits of current systems and the future path to AGI.

← PreviousNext →

Stay In The Loop! Subscribe to Our Newsletter.

Get updates straight to your inbox. No spam, just useful content.