newsfilter.io

Francois Chaubard

Showing 1–4 of 4 transcripts.

  1. Y Combinator53 min

    Going In Deep On Data | YC Paper Club

    Francois Chaubard, Vincent Sunn Chen, Volo Kuleshov, Shayne Longpre

    Industry experts from Snorkel, Inception Labs, and MIT articulate a paradigm shift where data has evolved from a passive resource into a $100 billion sector requiring complex product craftsmanship and continuous expert supervision. Presentations detailed breakthroughs such as Snorkel's Senior SWE-bench with validation agents, Inception's Mercury 2 diffusion model achieving 1,000+ tokens per second, and novel cross-lingual scaling laws that optimize low-resource language performance. These developments collectively redefine the production frontier by demonstrating that architectural advances alone are insufficient without scalable data programming and domain-specific synthetic environments to resolve real-world distribution gaps.

  2. Y Combinator1h 24m

    Why Robotics Still Isn't Solved - But Could Be Soon | YC Paper Club

    Francois Chaubard, Marcel Torne, Milan Ganai, Tyler Ga Wei Lum, Niko West, Bill Jiao, Guanming Wang, Shyam Gollakota

    Recent industry presentations address critical scaling barriers in robotics, including Sim-to-Real gaps, deformable object handling, and the lack of sensory feedback by introducing specialized architectures like Multi-Scale Embodied Memory and self-supervised Chain of Thought frameworks. Concurrently, a strategic shift toward "Robotics Application Companies" emphasizes proprietary data collection and teleoperation over generalist models, while new World Action Model optimizations reduce inference latency to enable real-time physical AI deployment. These technical and business advancements collectively aim to transition robotics from isolated demos to mass-market, long-horizon task execution.

  3. Y Combinator1h 16m

    Kernels and Chips: Cursor, NVIDIA, and Meta Researchers on GPU Performance | YC Paper Club

    Stuart Sul, John, Francois Chaubard, Jon Saad-Falcon, Mark Saroufim, Misha Smelyanskiy, Brennan Shacklett

    The event synthesizes critical industry shifts toward specialized ASICs and disaggregated architectures that separate training and inference workloads to maximize intelligence per watt. Technical deep dives highlight breakthroughs in multi-GPU kernel optimization and local inference, demonstrating that consumer accelerators can now deliver nearly 90% of frontier model utility while reducing energy costs by 70%. These innovations collectively address systemic bottlenecks in network communication and latency, establishing a roadmap for distributed, cost-efficient AI ecosystems that minimize reliance on centralized cloud resources.

  4. Y Combinator38 min

    Recursion Is The Next Scaling Law In AI

    Ankit Gupta, Francois Chaubard

    Two 2025 research initiatives, Hierarchical Reasoning Models (HRM) and Tiny Recursive Models (TRM), challenge standard scaling laws by utilizing inference-time recursion to achieve state-of-the-art reasoning with drastically fewer parameters. HRM reaches 27 million parameters using a three-level weighted hierarchy, while the simplified TRM distills the architecture to just 7 million parameters yet achieves 87% accuracy on ARC-Prize benchmarks by treating recurrence as a dynamic latent memory tape. These systems overcome historical RNN limitations through Deep Equilibrium Models and latent recursion, offering a pathway to efficient, deep reasoning that diverges from traditional Chain-of-Thought constraints.