Distributed Computing podcasts

Rohan Anil of Core Automation discusses the relationship between pretraining, reinforcement learning and learning after deployment. Drawing on his optimization work at Google and Anthropic, he examines why additional post-training, distillation and longer contexts may not solve continual learning. Other topics include second-order optimizers, distributed training, low-level coding agents and the difficulty of coordinating very large GPU clusters. Hosted by Ravid Shwartz Ziv and Allen Roush. Watch the full conversation on the publisher’s YouTube channel.

Hosted by Ravid Shwartz Ziv, Allen Roush

Charles Frye of Modal examines the infrastructure constraints behind large-scale AI research. With Ravid Shwartz Ziv and Allen Roush, he discusses underused hardware, saturated resources, sharing GPU capacity and deciding when to train a model. The conversation also covers inference and speculative decoding, hardware economics, virtualization and the security boundaries around agents and open models. Hosted by Ravid Shwartz Ziv and Allen Roush. Watch the full conversation on the publisher’s YouTube channel.

Hosted by Ravid Shwartz Ziv, Allen Roush

Berkeley professor Ion Stoica discusses how research software becomes useful infrastructure and, sometimes, a company. The conversation draws on systems projects and companies associated with his laboratory to examine adoption, maintenance and the role of students. It also covers the complexity of the AI software stack, GPU utilization, coding agents in distributed systems, reward design and regulating outcomes. Hosted by Ravid Shwartz Ziv and Allen Roush. Watch the full conversation on the publisher’s YouTube channel.

Hosted by Ravid Shwartz Ziv, Allen Roush

Fireworks AI co-founder Benny Chen explains the infrastructure required to serve machine-learning models at scale, from GPU procurement and kernels to runtimes and request routing. He discusses power constraints, hardware utilization, mixture-of-experts models, quantization, reinforcement-learning efficiency and speculative decoding. The conversation also covers distributed training across data centres, sparse autoencoders, the constraints on open-source interpretability research, and the relationship between open models and compute availability. Hosted by Ravid Shwartz Ziv and Allen Roush; the full conversation is available on YouTube.

Hosted by Ravid Shwartz Ziv, Allen Roush

Philip Kiely surveys inference engineering at the intersection of GPU programming, applied research and distributed systems. Topics include batching, quantisation, speculative execution, cache reuse and workload-specific serving, alongside the choices between hosted APIs, dedicated deployments and internal platforms for multimodal and agent applications.

Hosted by Sam Charrington

Zain Asgar examines running agent inference across heterogeneous hardware. The discussion covers splitting workloads between GPUs and CPUs, compilation for different targets, automated kernel optimisation and hardware-aware scheduling. It also explores networking constraints and trade-offs among numerical precision, application accuracy, performance and cost.

Hosted by Sam Charrington